[tests] Skip issues test on Windows for now.

Add system:legacy_stdio_definitions.lib to .odin for Windows
Merge pull request #2102 from yay/darwin-read-write
2026-07-05 11:11:37 -07:00 · 2022-10-02 20:56:30 +02:00 · 2022-10-02 11:58:17 +01:00 · 2022-10-02 11:51:00 +01:00 · 2022-10-02 07:47:22 +01:00 · 2022-10-02 07:40:28 +01:00
733 changed files with 143957 additions and 39604 deletions
@@ -1,3 +1,4 @@
 # These are supported funding model platforms

+github: odin-lang
 patreon: gingerbill
@@ -1,5 +1,5 @@
 name: CI
-on: [push, pull_request]
+on: [push, pull_request, workflow_dispatch]

 jobs:
  build_linux:
@@ -7,9 +7,9 @@ jobs:
    steps:
      - uses: actions/checkout@v1
      - name: Download LLVM, botan
-        run: sudo apt-get install llvm-11 clang-11 llvm libbotan-2-dev botan
+        run: sudo apt-get install llvm-11 clang-11 libbotan-2-dev botan
      - name: build odin
-        run: make release
+        run: ./build_odin.sh release
      - name: Odin version
        run: ./odin version
        timeout-minutes: 1
@@ -17,13 +17,16 @@ jobs:
        run: ./odin report
        timeout-minutes: 1
      - name: Odin check
-        run: ./odin check examples/demo/demo.odin -vet
+        run: ./odin check examples/demo -vet
        timeout-minutes: 10
      - name: Odin run
-        run: ./odin run examples/demo/demo.odin
+        run: ./odin run examples/demo
        timeout-minutes: 10
      - name: Odin run -debug
-        run: ./odin run examples/demo/demo.odin -debug
+        run: ./odin run examples/demo -debug
+        timeout-minutes: 10
+      - name: Odin check examples/all
+        run: ./odin check examples/all -strict-style
        timeout-minutes: 10
      - name: Core library tests
        run: |
@@ -35,6 +38,15 @@ jobs:
          cd tests/vendor
          make
        timeout-minutes: 10
+      - name: Odin check examples/all for Linux i386
+        run: ./odin check examples/all -vet -strict-style -target:linux_i386
+        timeout-minutes: 10
+      - name: Odin check examples/all for FreeBSD amd64
+        run: ./odin check examples/all -vet -strict-style -target:freebsd_amd64
+        timeout-minutes: 10
+      - name: Odin check examples/all for OpenBSD amd64
+        run: ./odin check examples/all -vet -strict-style -target:openbsd_amd64
+        timeout-minutes: 10
  build_macOS:
    runs-on: macos-latest
    steps:
@@ -46,7 +58,7 @@ jobs:
          TMP_PATH=$(xcrun --show-sdk-path)/user/include
          echo "CPATH=$TMP_PATH" >> $GITHUB_ENV
      - name: build odin
-        run: make release
+        run: ./build_odin.sh release
      - name: Odin version
        run: ./odin version
        timeout-minutes: 1
@@ -54,13 +66,16 @@ jobs:
        run: ./odin report
        timeout-minutes: 1
      - name: Odin check
-        run: ./odin check examples/demo/demo.odin -vet
+        run: ./odin check examples/demo -vet
        timeout-minutes: 10
      - name: Odin run
-        run: ./odin run examples/demo/demo.odin
+        run: ./odin run examples/demo
        timeout-minutes: 10
      - name: Odin run -debug
-        run: ./odin run examples/demo/demo.odin -debug
+        run: ./odin run examples/demo -debug
+        timeout-minutes: 10
+      - name: Odin check examples/all
+        run: ./odin check examples/all -strict-style
        timeout-minutes: 10
      - name: Core library tests
        run: |
@@ -72,8 +87,19 @@ jobs:
          cd tests/vendor
          make
        timeout-minutes: 10
+      - name: Odin issues tests
+        run: |
+          cd tests/issues
+          ./run.sh
+        timeout-minutes: 10
+      - name: Odin check examples/all for Darwin arm64
+        run: ./odin check examples/all -vet -strict-style -target:darwin_arm64
+        timeout-minutes: 10
+      - name: Odin check examples/all for Linux arm64
+        run: ./odin check examples/all -vet -strict-style -target:linux_arm64
+        timeout-minutes: 10
  build_windows:
-    runs-on: windows-latest
+    runs-on: windows-2019
    steps:
      - uses: actions/checkout@v1
      - name: build Odin
@@ -91,19 +117,25 @@ jobs:
        shell: cmd
        run: |
          call "C:\Program Files (x86)\Microsoft Visual Studio\2019\Enterprise\VC\Auxiliary\Build\vcvars64.bat
-          odin check examples/demo/demo.odin -vet
+          odin check examples/demo -vet
        timeout-minutes: 10
      - name: Odin run
        shell: cmd
        run: |
          call "C:\Program Files (x86)\Microsoft Visual Studio\2019\Enterprise\VC\Auxiliary\Build\vcvars64.bat
-          odin run examples/demo/demo.odin
+          odin run examples/demo
        timeout-minutes: 10
      - name: Odin run -debug
        shell: cmd
        run: |
          call "C:\Program Files (x86)\Microsoft Visual Studio\2019\Enterprise\VC\Auxiliary\Build\vcvars64.bat
-          odin run examples/demo/demo.odin -debug
+          odin run examples/demo -debug
+        timeout-minutes: 10
+      - name: Odin check examples/all
+        shell: cmd
+        run: |
+          call "C:\Program Files (x86)\Microsoft Visual Studio\2019\Enterprise\VC\Auxiliary\Build\vcvars64.bat
+          odin check examples/all -strict-style
        timeout-minutes: 10
      - name: Core library tests
        shell: cmd
@@ -126,3 +158,16 @@ jobs:
          cd tests\core\math\big
          call build.bat
        timeout-minutes: 10
+      - name: Odin issues tests
+        shell: cmd
+        run: |
+          call "C:\Program Files (x86)\Microsoft Visual Studio\2019\Enterprise\VC\Auxiliary\Build\vcvars64.bat
+          cd tests\issues
+          call run.bat
+        timeout-minutes: 10
+      - name: Odin check examples/all for Windows 32bits
+        shell: cmd
+        run: |
+          call "C:\Program Files (x86)\Microsoft Visual Studio\2019\Enterprise\VC\Auxiliary\Build\vcvars64.bat
+          odin check examples/all -strict-style -target:windows_i386
+        timeout-minutes: 10
@@ -7,7 +7,7 @@ on:

 jobs:
  build_windows:
-    runs-on: windows-latest
+    runs-on: windows-2019
    steps:
      - uses: actions/checkout@v1
      - name: build Odin
@@ -19,7 +19,7 @@ jobs:
        shell: cmd
        run: |
          call "C:\Program Files (x86)\Microsoft Visual Studio\2019\Enterprise\VC\Auxiliary\Build\vcvars64.bat
-          odin run examples/demo/demo.odin
+          odin run examples/demo
      - name: Copy artifacts
        run: |
          rm bin/llvm/windows/LLVM-C.lib
@@ -41,15 +41,16 @@ jobs:
    steps:
      - uses: actions/checkout@v1
      - name: (Linux) Download LLVM
-        run: sudo apt-get install llvm-11 clang-11 llvm
+        run: sudo apt-get install llvm-11 clang-11
      - name: build odin
        run: make nightly
      - name: Odin run
-        run: ./odin run examples/demo/demo.odin
+        run: ./odin run examples/demo
      - name: Copy artifacts
        run: |
          mkdir dist
          cp odin dist
+          cp libLLVM* dist
          cp -r shared dist
          cp -r core dist
          cp -r vendor dist
@@ -72,7 +73,7 @@ jobs:
      - name: build odin
        run: make nightly
      - name: Odin run
-        run: ./odin run examples/demo/demo.odin
+        run: ./odin run examples/demo
      - name: Copy artifacts
        run: |
          mkdir dist
@@ -129,7 +130,7 @@ jobs:
        run: |
          echo Authorizing B2 account
          b2 authorize-account "$APPID" "$APPKEY"
-          
+
          echo Uploading artifcates to B2
          chmod +x ./ci/upload_create_nightly.sh
          ./ci/upload_create_nightly.sh "$BUCKET" windows-amd64 windows_artifacts/
@@ -141,7 +142,7 @@ jobs:

          echo Creating nightly.json
          python3 ci/create_nightly_json.py "$BUCKET" > nightly.json
-          
+
          echo Uploading nightly.json
          b2 upload-file "$BUCKET" nightly.json nightly.json

@@ -0,0 +1,43 @@
+name: "Close Stale Issues & PRs"
+on:
+  workflow_dispatch:
+  schedule:
+    - cron: "0 21 * * *"
+    
+permissions:
+  issues: write
+  pull-requests: write
+
+jobs:
+  stale:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Close Stale Issues
+        uses: actions/stale@v4.1.0
+        with:
+#          stale-issue-message: |
+#            Hello!
+#            
+#            I am marking this issue as stale as it has not received any engagement from the community or maintainers 120 days. That does not imply that the issue has no merit! If you feel strongly about this issue
+#            - open a PR referencing and resolving the issue;
+#            - leave a comment on it and discuss ideas how you could contribute towards resolving it;
+#            - leave a comment and describe in detail why this issue is critical for your use case;
+#            - open a new issue with updated details and a plan on resolving the issue.
+#
+#            The motivation for this automation is to help prioritize issues in the backlog and not ignore, reject, or belittle anyone..
+#
+#          stale-pr-message: |
+#            Hello!
+#            
+#            I am marking this PR as stale as it has not received any engagement from the community or maintainers 120 days. That does not imply that the issue has no merit! If you feel strongly about this issue
+#            - leave a comment on it and discuss ideas how you could contribute towards resolving it;
+#            - leave a comment and describe in detail why this issue is critical for your use case;
+#
+#           The motivation for this automation is to help prioritize issues in the backlog and not ignore, reject, or belittle anyone..
+
+          days-before-stale: 120
+          days-before-close: 30
+          exempt-draft-pr: true
+          ascending: true
+          operations-per-run: 1000
+          exempt-issue-labels: "ignore"
@@ -7,6 +7,9 @@
 # User-specific files (MonoDevelop/Xamarin Studio)
 *.userprefs

+# For macOS
+.DS_Store
+
 # Build results
 [Dd]ebug/
 [Dd]ebugPublic/
@@ -266,6 +269,9 @@ bin/
 # - Linux/MacOS
 odin
 odin.dSYM
+*.bin
+demo.bin
+libLLVM*.so*

 # shared collection
 shared/
@@ -276,3 +282,6 @@ shared/
 *.ll

 *.sublime-workspace
+examples/bug/
+build.sh
+!core/debug/
@@ -1,4 +1,4 @@
-Copyright (c) 2016-2021 Ginger Bill. All rights reserved.
+Copyright (c) 2016-2022 Ginger Bill. All rights reserved.

 Redistribution and use in source and binary forms, with or without
 modification, are permitted provided that the following conditions are met:
@@ -1,58 +1,19 @@
-GIT_SHA=$(shell git rev-parse --short HEAD)
-DISABLED_WARNINGS=-Wno-switch -Wno-macro-redefined -Wno-unused-value
-LDFLAGS=-pthread -ldl -lm -lstdc++
-CFLAGS=-std=c++14 -DGIT_SHA=\"$(GIT_SHA)\"
-CFLAGS:=$(CFLAGS) -DODIN_VERSION_RAW=\"dev-$(shell date +"%Y-%m")\"
-CC=clang
-
-OS=$(shell uname)
-
-ifeq ($(OS), Darwin)
-    LLVM_CONFIG=llvm-config
-    ifneq ($(shell llvm-config --version | grep '^11\.'),)
-        LLVM_CONFIG=llvm-config
-    else
-        $(error "Requirement: llvm-config must be version 11")
-    endif
-
-    LDFLAGS:=$(LDFLAGS) -liconv
-    CFLAGS:=$(CFLAGS) $(shell $(LLVM_CONFIG) --cxxflags --ldflags)
-    LDFLAGS:=$(LDFLAGS) -lLLVM-C
-endif
-ifeq ($(OS), Linux)
-    LLVM_CONFIG=llvm-config-11
-    ifneq ($(shell which llvm-config-11 2>/dev/null),)
-        LLVM_CONFIG=llvm-config-11
-    else ifneq ($(shell which llvm-config-11-64 2>/dev/null),)
-        LLVM_CONFIG=llvm-config-11-64
-    else
-        ifneq ($(shell llvm-config --version | grep '^11\.'),)
-            LLVM_CONFIG=llvm-config
-        else
-            $(error "Requirement: llvm-config must be version 11")
-        endif
-    endif
-
-    CFLAGS:=$(CFLAGS) $(shell $(LLVM_CONFIG) --cxxflags --ldflags)
-    LDFLAGS:=$(LDFLAGS) $(shell $(LLVM_CONFIG) --libs core native --system-libs)
-endif
-
-all: debug demo
+all: debug

 demo:
-	./odin run examples/demo/demo.odin
+	./odin run examples/demo/demo.odin -file

 report:
 	./odin report

 debug:
-	$(CC) src/main.cpp src/libtommath.cpp $(DISABLED_WARNINGS) $(CFLAGS) -g $(LDFLAGS) -o odin
+	./build_odin.sh debug

 release:
-	$(CC) src/main.cpp src/libtommath.cpp $(DISABLED_WARNINGS) $(CFLAGS) -O3 $(LDFLAGS) -o odin
+	./build_odin.sh release

 release_native:
-	$(CC) src/main.cpp src/libtommath.cpp $(DISABLED_WARNINGS) $(CFLAGS) -O3 -march=native $(LDFLAGS) -o odin
+	./build_odin.sh release-native

 nightly:
-	$(CC) src/main.cpp src/libtommath.cpp $(DISABLED_WARNINGS) $(CFLAGS) -DNIGHTLY -O3 $(LDFLAGS) -o odin
+	./build_odin.sh nightly
@@ -11,7 +11,7 @@
        <img src="https://img.shields.io/badge/platforms-Windows%20|%20Linux%20|%20macOS-green.svg">
    </a>
    <br>
-    <a href="https://discord.gg/hnwN2Rj">
+    <a href="https://discord.gg/odinlang">
        <img src="https://img.shields.io/discord/568138951836172421?logo=discord">
    </a>
    <a href="https://github.com/odin-lang/odin/actions">
@@ -58,6 +58,10 @@ main :: proc() {

 Instructions for downloading and installing the Odin compiler and libraries.

+#### [Nightly Builds](https://odin-lang.org/docs/nightly/)
+
+Get the latest nightly builds of Odin.
+
 ### Learning Odin

 #### [Overview of Odin](https://odin-lang.org/docs/overview)
@@ -68,6 +72,10 @@ An overview of the Odin programming language.

 Answers to common questions about Odin.

+#### [Packages](https://pkg.odin-lang.org/)
+
+Documentation for all the official packages part of the [core](https://pkg.odin-lang.org/core/) and [vendor](https://pkg.odin-lang.org/vendor/) library collections.
+
 #### [The Odin Wiki](https://github.com/odin-lang/Odin/wiki)

 A wiki maintained by the Odin community.
@@ -76,15 +84,9 @@ A wiki maintained by the Odin community.

 Get live support and talk with other odiners on the Odin Discord.

-### References
-
-#### [Language Specification](https://odin-lang.org/docs/spec/)
-
-The official Odin Language specification.
-
 ### Articles

-#### [The Odin Blog](https://odin-lang.org/blog)
+#### [The Odin Blog](https://odin-lang.org/news/)

 The official blog of the Odin programming language, featuring announcements, news, and in-depth articles by the Odin team and guests.

@@ -58,7 +58,7 @@ set libs= ^
 set linker_flags= -incremental:no -opt:ref -subsystem:console

 if %release_mode% EQU 0 ( rem Debug
-	set linker_flags=%linker_flags% -debug
+	set linker_flags=%linker_flags% -debug /NATVIS:src\odin_compiler.natvis
 ) else ( rem Release
 	set linker_flags=%linker_flags% -debug
 )
@@ -0,0 +1,179 @@
+#!/usr/bin/env bash
+set -eu
+
+: ${CXX=clang++}
+: ${CPPFLAGS=}
+: ${CXXFLAGS=}
+: ${LDFLAGS=}
+: ${ODIN_VERSION=dev-$(date +"%Y-%m")}
+
+CPPFLAGS="$CPPFLAGS -DODIN_VERSION_RAW=\"$ODIN_VERSION\""
+CXXFLAGS="$CXXFLAGS -std=c++14"
+LDFLAGS="$LDFLAGS -pthread -lm -lstdc++"
+
+GIT_SHA=$(git rev-parse --short HEAD || :)
+if [ "$GIT_SHA" ]; then CPPFLAGS="$CPPFLAGS -DGIT_SHA=\"$GIT_SHA\""; fi
+
+DISABLED_WARNINGS="-Wno-switch -Wno-macro-redefined -Wno-unused-value"
+OS=$(uname)
+
+panic() {
+	printf "%s\n" "$1"
+	exit 1
+}
+
+version() { echo "$@" | awk -F. '{ printf("%d%03d%03d%03d\n", $1,$2,$3,$4); }'; }
+
+config_darwin() {
+	ARCH=$(uname -m)
+	: ${LLVM_CONFIG=llvm-config}
+
+	# allow for arm only llvm's with version 13
+	if [ ARCH == arm64 ]; then
+		MIN_LLVM_VERSION=("13.0.0")
+	else
+		# allow for x86 / amd64 all llvm versions beginning from 11
+		MIN_LLVM_VERSION=("11.1.0")
+	fi
+
+	if [ $(version $($LLVM_CONFIG --version)) -lt $(version $MIN_LLVM_VERSION) ]; then
+		if [ ARCH == arm64 ]; then
+			panic "Requirement: llvm-config must be base version 13 for arm64"
+		else
+			panic "Requirement: llvm-config must be base version greater than 11 for amd64/x86"
+		fi
+	fi
+
+	LDFLAGS="$LDFLAGS -liconv -ldl"
+	CXXFLAGS="$CXXFLAGS $($LLVM_CONFIG --cxxflags --ldflags)"
+	LDFLAGS="$LDFLAGS -lLLVM-C"
+}
+
+config_freebsd() {
+	: ${LLVM_CONFIG=}
+
+	if [ ! "$LLVM_CONFIG" ]; then
+		if which llvm-config11 > /dev/null 2>&1; then
+			LLVM_CONFIG=llvm-config11
+		elif which llvm-config12 > /dev/null 2>&1; then
+			LLVM_CONFIG=llvm-config12
+		elif which llvm-config13 > /dev/null 2>&1; then
+			LLVM_CONFIG=llvm-config13
+		else
+			panic "Unable to find LLVM-config"
+		fi
+	fi
+
+	CXXFLAGS="$CXXFLAGS $($LLVM_CONFIG --cxxflags --ldflags)"
+	LDFLAGS="$LDFLAGS $($LLVM_CONFIG --libs core native --system-libs)"
+}
+
+config_openbsd() {
+	: ${LLVM_CONFIG=/usr/local/bin/llvm-config}
+
+	LDFLAGS="$LDFLAGS -liconv"
+	CXXFLAGS="$CXXFLAGS $($LLVM_CONFIG --cxxflags --ldflags)"
+	LDFLAGS="$LDFLAGS $($LLVM_CONFIG --libs core native --system-libs)"
+}
+
+config_linux() {
+	: ${LLVM_CONFIG=}
+
+	if [ ! "$LLVM_CONFIG" ]; then
+		if which llvm-config > /dev/null 2>&1; then
+			LLVM_CONFIG=llvm-config
+		elif which llvm-config-11 > /dev/null 2>&1; then
+			LLVM_CONFIG=llvm-config-11
+		elif which llvm-config-11-64 > /dev/null 2>&1; then
+			LLVM_CONFIG=llvm-config-11-64
+		else
+			panic "Unable to find LLVM-config"
+		fi
+	fi
+
+	MIN_LLVM_VERSION=("11.0.0")
+	if [ $(version $($LLVM_CONFIG --version)) -lt $(version $MIN_LLVM_VERSION) ]; then
+		echo "Tried to use " $(which $LLVM_CONFIG) "version" $($LLVM_CONFIG --version)
+		panic "Requirement: llvm-config must be base version greater than 11"
+	fi
+
+	LDFLAGS="$LDFLAGS -ldl"
+	CXXFLAGS="$CXXFLAGS $($LLVM_CONFIG --cxxflags --ldflags)"
+	LDFLAGS="$LDFLAGS $($LLVM_CONFIG  --libs core native --system-libs --libfiles) -Wl,-rpath=\$ORIGIN"
+
+	# Creates a copy of the llvm library in the build dir, this is meant to support compiler explorer.
+	# The annoyance is that this copy can be cluttering the development folder. TODO: split staging folders
+	# for development and compiler explorer builds
+	cp $(readlink -f $($LLVM_CONFIG --libfiles)) ./
+}
+
+build_odin() {
+	case $1 in
+	debug)
+		EXTRAFLAGS="-g"
+		;;
+	release)
+		EXTRAFLAGS="-O3"
+		;;
+	release-native)
+		EXTRAFLAGS="-O3 -march=native"
+		;;
+	nightly)
+		EXTRAFLAGS="-DNIGHTLY -O3"
+		;;
+	*)
+		panic "Build mode unsupported!"
+	esac
+
+	set -x
+	$CXX src/main.cpp src/libtommath.cpp $DISABLED_WARNINGS $CPPFLAGS $CXXFLAGS $EXTRAFLAGS $LDFLAGS -o odin
+	set +x
+}
+
+run_demo() {
+	./odin run examples/demo/demo.odin -file
+}
+
+case $OS in
+Linux)
+	config_linux
+	;;
+Darwin)
+	config_darwin
+	;;
+OpenBSD)
+	config_openbsd
+	;;
+FreeBSD)
+	config_freebsd
+	;;
+*)
+	panic "Platform unsupported!"
+esac
+
+if [[ $# -eq 0 ]]; then
+	build_odin debug
+	run_demo
+	exit 0
+fi
+
+if [[ $# -eq 1 ]]; then
+	case $1 in
+	report)
+		if [[ ! -f "./odin" ]]; then
+			build_odin debug
+		fi
+
+		./odin report
+		exit 0
+		;;
+	*)
+		build_odin $1
+		;;
+	esac
+
+	run_demo
+	exit 0
+else
+	panic "Too many arguments!"
+fi
@@ -15,20 +15,16 @@ read_writer_init :: proc(rw: ^Read_Writer, r: ^Reader, w: ^Writer) {

 read_writer_to_stream :: proc(rw: ^Read_Writer) -> (s: io.Stream) {
 	s.stream_data = rw
-	s.stream_vtable = _read_writer_vtable
+	s.stream_vtable = &_read_writer_vtable
 	return
 }

@(private)
-_read_writer_vtable := &io.Stream_VTable{
+_read_writer_vtable := io.Stream_VTable{
 	impl_read = proc(s: io.Stream, p: []byte) -> (n: int, err: io.Error) {
 		b := (^Read_Writer)(s.stream_data).r
 		return reader_read(b, p)
 	},
-	impl_read_byte = proc(s: io.Stream) -> (c: byte, err: io.Error) {
-		b := (^Read_Writer)(s.stream_data).r
-		return reader_read_byte(b)
-	},
 	impl_unread_byte = proc(s: io.Stream) -> io.Error {
 		b := (^Read_Writer)(s.stream_data).r
 		return reader_unread_byte(b)
@@ -353,14 +353,14 @@ reader_write_to :: proc(b: ^Reader, w: io.Writer) -> (n: i64, err: io.Error) {
 // reader_to_stream converts a Reader into an io.Stream
 reader_to_stream :: proc(b: ^Reader) -> (s: io.Stream) {
 	s.stream_data = b
-	s.stream_vtable = _reader_vtable
+	s.stream_vtable = &_reader_vtable
 	return
 }



@(private)
-_reader_vtable := &io.Stream_VTable{
+_reader_vtable := io.Stream_VTable{
 	impl_destroy = proc(s: io.Stream) -> io.Error {
 		b := (^Reader)(s.stream_data)
 		reader_destroy(b)
@@ -8,6 +8,7 @@ import "core:intrinsics"

 // Extra errors returns by scanning procedures
 Scanner_Extra_Error :: enum i32 {
+	None,
 	Negative_Advance,
 	Advanced_Too_Far,
 	Bad_Read_Count,
@@ -15,7 +16,7 @@ Scanner_Extra_Error :: enum i32 {
 	Too_Short,
 }

-Scanner_Error :: union {
+Scanner_Error :: union #shared_nil {
 	io.Error,
 	Scanner_Extra_Error,
 }
@@ -65,10 +66,10 @@ scanner_destroy :: proc(s: ^Scanner) {
 }


-// Returns the first non-EOF error that was encounted by the scanner
+// Returns the first non-EOF error that was encountered by the scanner
 scanner_error :: proc(s: ^Scanner) -> Scanner_Error {
 	switch s._err {
-	case .EOF, .None:
+	case .EOF, nil:
 		return nil
 	}
 	return s._err
@@ -93,10 +94,6 @@ scanner_text :: proc(s: ^Scanner) -> string {
 // scanner_scan advances the scanner
 scanner_scan :: proc(s: ^Scanner) -> bool {
 	set_err :: proc(s: ^Scanner, err: Scanner_Error) {
-		err := err
-		if err == .None {
-			err = nil
-		}
 		switch s._err {
 		case nil, .EOF:
 			s._err = err
@@ -223,14 +223,14 @@ writer_read_from :: proc(b: ^Writer, r: io.Reader) -> (n: i64, err: io.Error) {
 // writer_to_stream converts a Writer into an io.Stream
 writer_to_stream :: proc(b: ^Writer) -> (s: io.Stream) {
 	s.stream_data = b
-	s.stream_vtable = _writer_vtable
+	s.stream_vtable = &_writer_vtable
 	return
 }



@(private)
-_writer_vtable := &io.Stream_VTable{
+_writer_vtable := io.Stream_VTable{
 	impl_destroy = proc(s: io.Stream) -> io.Error {
 		b := (^Writer)(s.stream_data)
 		writer_destroy(b)
@@ -1,90 +1,90 @@
 // This is purely for documentation
 package builtin

-nil   :: nil;
-false :: 0!==0;
-true  :: 0==0;
+nil   :: nil
+false :: 0!=0
+true  :: 0==0

-ODIN_OS      :: ODIN_OS;
-ODIN_ARCH    :: ODIN_ARCH;
-ODIN_ENDIAN  :: ODIN_ENDIAN;
-ODIN_VENDOR  :: ODIN_VENDOR;
-ODIN_VERSION :: ODIN_VERSION;
-ODIN_ROOT    :: ODIN_ROOT;
-ODIN_DEBUG   :: ODIN_DEBUG;
+ODIN_OS      :: ODIN_OS
+ODIN_ARCH    :: ODIN_ARCH
+ODIN_ENDIAN  :: ODIN_ENDIAN
+ODIN_VENDOR  :: ODIN_VENDOR
+ODIN_VERSION :: ODIN_VERSION
+ODIN_ROOT    :: ODIN_ROOT
+ODIN_DEBUG   :: ODIN_DEBUG

-byte :: u8; // alias
+byte :: u8 // alias

-bool          :: bool;
-b8            :: b8;
-b16           :: b16;
-b32           :: b32;
-b64           :: b64;
+bool          :: bool
+b8            :: b8
+b16           :: b16
+b32           :: b32
+b64           :: b64

-i8            :: i8;
-u8            :: u8;
-i16           :: i16;
-u16           :: u16;
-i32           :: i32;
-u32           :: u32;
-i64           :: i64;
-u64           :: u64;
+i8            :: i8
+u8            :: u8
+i16           :: i16
+u16           :: u16
+i32           :: i32
+u32           :: u32
+i64           :: i64
+u64           :: u64

-i128          :: i128;
-u128          :: u128;
+i128          :: i128
+u128          :: u128

-rune          :: rune;
+rune          :: rune

-f16           :: f16;
-f32           :: f32;
-f64           :: f64;
+f16           :: f16
+f32           :: f32
+f64           :: f64

-complex32     :: complex32;
-complex64     :: complex64;
-complex128    :: complex128;
+complex32     :: complex32
+complex64     :: complex64
+complex128    :: complex128

-quaternion64  :: quaternion64;
-quaternion128 :: quaternion128;
-quaternion256 :: quaternion256;
+quaternion64  :: quaternion64
+quaternion128 :: quaternion128
+quaternion256 :: quaternion256

-int           :: int;
-uint          :: uint;
-uintptr       :: uintptr;
+int           :: int
+uint          :: uint
+uintptr       :: uintptr

-rawptr        :: rawptr;
-string        :: string;
-cstring       :: cstring;
-any           :: any;
+rawptr        :: rawptr
+string        :: string
+cstring       :: cstring
+any           :: any

-typeid        :: typeid;
+typeid        :: typeid

 // Endian Specific Types
-i16le         :: i16le;
-u16le         :: u16le;
-i32le         :: i32le;
-u32le         :: u32le;
-i64le         :: i64le;
-u64le         :: u64le;
-i128le        :: i128le;
-u128le        :: u128le;
+i16le         :: i16le
+u16le         :: u16le
+i32le         :: i32le
+u32le         :: u32le
+i64le         :: i64le
+u64le         :: u64le
+i128le        :: i128le
+u128le        :: u128le

-i16be         :: i16be;
-u16be         :: u16be;
-i32be         :: i32be;
-u32be         :: u32be;
-i64be         :: i64be;
-u64be         :: u64be;
-i128be        :: i128be;
-u128be        :: u128be;
+i16be         :: i16be
+u16be         :: u16be
+i32be         :: i32be
+u32be         :: u32be
+i64be         :: i64be
+u64be         :: u64be
+i128be        :: i128be
+u128be        :: u128be


-f16le         :: f16le;
-f32le         :: f32le;
-f64le         :: f64le;
+f16le         :: f16le
+f32le         :: f32le
+f64le         :: f64le

-f16be         :: f16be;
-f32be         :: f32be;
-f64be         :: f64be;
+f16be         :: f16be
+f32be         :: f32be
+f64be         :: f64be



@@ -161,6 +161,10 @@ buffer_write :: proc(b: ^Buffer, p: []byte) -> (n: int, err: io.Error) {
 	return copy(b.buf[m:], p), nil
 }

+buffer_write_ptr :: proc(b: ^Buffer, ptr: rawptr, size: int) -> (n: int, err: io.Error) {
+	return buffer_write(b, ([^]byte)(ptr)[:size])
+}
+
 buffer_write_string :: proc(b: ^Buffer, s: string) -> (n: int, err: io.Error) {
 	b.last_read = .Invalid
 	m, ok := _buffer_try_grow(b, len(s))
@@ -229,17 +233,18 @@ buffer_read :: proc(b: ^Buffer, p: []byte) -> (n: int, err: io.Error) {
 	return
 }

+buffer_read_ptr :: proc(b: ^Buffer, ptr: rawptr, size: int) -> (n: int, err: io.Error) {
+	return buffer_read(b, ([^]byte)(ptr)[:size])
+}
+
 buffer_read_at :: proc(b: ^Buffer, p: []byte, offset: int) -> (n: int, err: io.Error) {
 	b.last_read = .Invalid

-	if offset < 0 || offset >= len(b.buf) {
+	if uint(offset) >= len(b.buf) {
 		err = .Invalid_Offset
 		return
 	}
-
-	if 0 <= offset && offset < len(b.buf) {
-		n = copy(p, b.buf[offset:])
-	}
+	n = copy(p, b.buf[offset:])
 	if n > 0 {
 		b.last_read = .Read
 	}
@@ -366,12 +371,12 @@ buffer_read_from :: proc(b: ^Buffer, r: io.Reader) -> (n: i64, err: io.Error) #n

 buffer_to_stream :: proc(b: ^Buffer) -> (s: io.Stream) {
 	s.stream_data = b
-	s.stream_vtable = _buffer_vtable
+	s.stream_vtable = &_buffer_vtable
 	return
 }

@(private)
-_buffer_vtable := &io.Stream_VTable{
+_buffer_vtable := io.Stream_VTable{
 	impl_size = proc(s: io.Stream) -> i64 {
 		b := (^Buffer)(s.stream_data)
 		return i64(buffer_capacity(b))
@@ -5,13 +5,19 @@ import "core:unicode"
 import "core:unicode/utf8"

 clone :: proc(s: []byte, allocator := context.allocator, loc := #caller_location) -> []byte {
-	c := make([]byte, len(s)+1, allocator, loc)
+	c := make([]byte, len(s), allocator, loc)
 	copy(c, s)
-	c[len(s)] = 0
 	return c[:len(s)]
 }

-ptr_from_slice :: proc(str: []byte) -> ^byte {
+clone_safe :: proc(s: []byte, allocator := context.allocator, loc := #caller_location) -> (data: []byte, err: mem.Allocator_Error) {
+	c := make([]byte, len(s), allocator, loc) or_return
+	copy(c, s)
+	return c[:len(s)], nil
+}
+
+ptr_from_slice :: ptr_from_bytes
+ptr_from_bytes :: proc(str: []byte) -> ^byte {
 	d := transmute(mem.Raw_String)str
 	return d.data
 }
@@ -135,6 +141,25 @@ join :: proc(a: [][]byte, sep: []byte, allocator := context.allocator) -> []byte
 	return b
 }

+join_safe :: proc(a: [][]byte, sep: []byte, allocator := context.allocator) -> (data: []byte, err: mem.Allocator_Error) {
+	if len(a) == 0 {
+		return nil, nil
+	}
+
+	n := len(sep) * (len(a) - 1)
+	for s in a {
+		n += len(s)
+	}
+
+	b := make([]byte, n, allocator) or_return
+	i := copy(b, a[0])
+	for s in a[1:] {
+		i += copy(b[i:], sep)
+		i += copy(b[i:], s)
+	}
+	return b, nil
+}
+
 concatenate :: proc(a: [][]byte, allocator := context.allocator) -> []byte {
 	if len(a) == 0 {
 		return nil
@@ -152,6 +177,24 @@ concatenate :: proc(a: [][]byte, allocator := context.allocator) -> []byte {
 	return b
 }

+concatenate_safe :: proc(a: [][]byte, allocator := context.allocator) -> (data: []byte, err: mem.Allocator_Error) {
+	if len(a) == 0 {
+		return nil, nil
+	}
+
+	n := 0
+	for s in a {
+		n += len(s)
+	}
+	b := make([]byte, n, allocator) or_return
+	i := 0
+	for s in a {
+		i += copy(b[i:], s)
+	}
+	return b, nil
+}
+
+
@private
 _split :: proc(s, sep: []byte, sep_save, n: int, allocator := context.allocator) -> [][]byte {
 	s, n := s, n
@@ -219,61 +262,37 @@ split_after_n :: proc(s, sep: []byte, n: int, allocator := context.allocator) ->


@private
-_split_iterator :: proc(s: ^[]byte, sep: []byte, sep_save, n: int) -> (res: []byte, ok: bool) {
-	s, n := s, n
-
-	if n == 0 {
-		return
-	}
-
-	if sep == nil {
+_split_iterator :: proc(s: ^[]byte, sep: []byte, sep_save: int) -> (res: []byte, ok: bool) {
+	if len(sep) == 0 {
 		res = s[:]
 		ok = true
 		s^ = s[len(s):]
 		return
 	}

-	if n < 0 {
-		n = count(s^, sep) + 1
-	}
-
-	n -= 1
-
-	i := 0
-	for ; i < n; i += 1 {
-		m := index(s^, sep)
-		if m < 0 {
-			break
-		}
+	m := index(s^, sep)
+	if m < 0 {
+		// not found
+		res = s[:]
+		ok = len(res) != 0
+		s^ = s[len(s):]
+	} else {
 		res = s[:m+sep_save]
 		ok = true
 		s^ = s[m+len(sep):]
-		return
 	}
-	res = s[:]
-	ok = res != nil
-	s^ = s[len(s):]
 	return
 }


 split_iterator :: proc(s: ^[]byte, sep: []byte) -> ([]byte, bool) {
-	return _split_iterator(s, sep, 0, -1)
-}
-
-split_n_iterator :: proc(s: ^[]byte, sep: []byte, n: int) -> ([]byte, bool) {
-	return _split_iterator(s, sep, 0, n)
+	return _split_iterator(s, sep, 0)
 }

 split_after_iterator :: proc(s: ^[]byte, sep: []byte) -> ([]byte, bool) {
-	return _split_iterator(s, sep, len(sep), -1)
+	return _split_iterator(s, sep, len(sep))
 }

-split_after_n_iterator :: proc(s: ^[]byte, sep: []byte, n: int) -> ([]byte, bool) {
-	return _split_iterator(s, sep, len(sep), n)
-}
-
-

 index_byte :: proc(s: []byte, c: byte) -> int {
 	for i := 0; i < len(s); i += 1 {
@@ -619,7 +638,7 @@ trim_left_proc :: proc(s: []byte, p: proc(rune) -> bool) -> []byte {

 index_rune :: proc(s: []byte, r: rune) -> int {
 	switch {
-	case 0 <= r && r < utf8.RUNE_SELF:
+	case u32(r) < utf8.RUNE_SELF:
 		return index_byte(s, byte(r))

 	case r == utf8.RUNE_ERROR:
@@ -1143,7 +1162,7 @@ fields_proc :: proc(s: []byte, f: proc(rune) -> bool, allocator := context.alloc
 	}

 	if start >= 0 {
-		append(&subslices, s[start : end])
+		append(&subslices, s[start : len(s)])
 	}

 	return subslices[:]
@@ -17,7 +17,7 @@ reader_init :: proc(r: ^Reader, s: []byte) {

 reader_to_stream :: proc(r: ^Reader) -> (s: io.Stream) {
 	s.stream_data = r
-	s.stream_vtable = _reader_vtable
+	s.stream_vtable = &_reader_vtable
 	return
 }

@@ -137,7 +137,7 @@ reader_write_to :: proc(r: ^Reader, w: io.Writer) -> (n: i64, err: io.Error) {


@(private)
-_reader_vtable := &io.Stream_VTable{
+_reader_vtable := io.Stream_VTable{
 	impl_size = proc(s: io.Stream) -> i64 {
 		r := (^Reader)(s.stream_data)
 		return reader_size(r)
@@ -3,22 +3,24 @@ package c
 import builtin "core:builtin"

 char           :: builtin.u8  // assuming -funsigned-char
+
+schar          :: builtin.i8
 short          :: builtin.i16
 int            :: builtin.i32
-long           :: builtin.i32 when (ODIN_OS == "windows" || size_of(builtin.rawptr) == 4) else builtin.i64
+long           :: builtin.i32 when (ODIN_OS == .Windows || size_of(builtin.rawptr) == 4) else builtin.i64
 longlong       :: builtin.i64

 uchar          :: builtin.u8
 ushort         :: builtin.u16
 uint           :: builtin.u32
-ulong          :: builtin.u32 when (ODIN_OS == "windows" || size_of(builtin.rawptr) == 4) else builtin.u64
+ulong          :: builtin.u32 when (ODIN_OS == .Windows || size_of(builtin.rawptr) == 4) else builtin.u64
 ulonglong      :: builtin.u64

 bool           :: builtin.bool

 size_t         :: builtin.uint
 ssize_t        :: builtin.int
-wchar_t        :: builtin.u16 when (ODIN_OS == "windows") else builtin.u32
+wchar_t        :: builtin.u16 when (ODIN_OS == .Windows) else builtin.u32

 float          :: builtin.f32
 double         :: builtin.f64
@@ -46,7 +48,7 @@ int_least64_t  :: builtin.i64
 uint_least64_t :: builtin.u64

 // Same on Windows, Linux, and FreeBSD
-when ODIN_ARCH == "386" || ODIN_ARCH == "amd64" {
+when ODIN_ARCH == .i386 || ODIN_ARCH == .amd64 {
 	int_fast8_t    :: builtin.i8
 	uint_fast8_t   :: builtin.u8
 	int_fast16_t   :: builtin.i32
@@ -519,7 +519,7 @@ join_adjacent_string_literals :: proc(cpp: ^Preprocessor, initial_tok: ^Token) {


 quote_string :: proc(s: string) -> []byte {
-	b := strings.make_builder(0, len(s)+2)
+	b := strings.builder_make(0, len(s)+2)
 	io.write_quoted_string(strings.to_writer(&b), s, '"')
 	return b.buf[:]
 }
@@ -956,7 +956,7 @@ substitute_token :: proc(cpp: ^Preprocessor, tok: ^Token, args: ^Macro_Arg) -> ^
 			continue
 		}

-		if tok.lit == "__VA__OPT__" && tok.next.lit == "(" {
+		if tok.lit == "__VA_OPT__" && tok.next.lit == "(" {
 			opt_arg := read_macro_arg_one(cpp, &tok, tok.next.next, true)
 			if has_varargs(args) {
 				for t := opt_arg.tok; t.kind != .EOF; t = t.next {
@@ -1276,7 +1276,7 @@ preprocess_internal :: proc(cpp: ^Preprocessor, tok: ^Token) -> ^Token {
 				if start.file != nil {
 					dir = filepath.dir(start.file.name)
 				}
-				path := filepath.join(dir, filename)
+				path := filepath.join({dir, filename})
 				if os.exists(path) {
 					tok = include_file(cpp, tok, path, start.next.next)
 					continue
@@ -2,9 +2,9 @@ package libc

 // 7.3 Complex arithmetic

-when ODIN_OS == "windows" {
+when ODIN_OS == .Windows {
 	foreign import libc "system:libucrt.lib"
-} else when ODIN_OS == "darwin" {
+} else when ODIN_OS == .Darwin {
 	foreign import libc "system:System.framework"
 } else {
 	foreign import libc "system:c"
@@ -49,8 +49,8 @@ foreign libc {
 	// 7.3.8 Power and absolute-value functions
 	cabs    :: proc(z: complex_double) -> complex_double ---
 	cabsf   :: proc(z: complex_float) -> complex_float ---
-	cpow    :: proc(z: complex_double) -> complex_double ---
-	cpowf   :: proc(z: complex_float) -> complex_float ---
+	cpow    :: proc(x, y: complex_double) -> complex_double ---
+	cpowf   :: proc(x, y: complex_float) -> complex_float ---
 	csqrt   :: proc(z: complex_double) -> complex_double ---
 	csqrtf  :: proc(z: complex_float) -> complex_float ---

@@ -1,8 +1,8 @@
 package libc

-when ODIN_OS == "windows" {
+when ODIN_OS == .Windows {
 	foreign import libc "system:libucrt.lib"
-} else when ODIN_OS == "darwin" {
+} else when ODIN_OS == .Darwin {
 	foreign import libc "system:System.framework"
 } else {
 	foreign import libc "system:c"
@@ -2,9 +2,9 @@ package libc

 // 7.5 Errors

-when ODIN_OS == "windows" {
+when ODIN_OS == .Windows {
 	foreign import libc "system:libucrt.lib"
-} else when ODIN_OS == "darwin" {
+} else when ODIN_OS == .Darwin {
 	foreign import libc "system:System.framework"
 } else {
 	foreign import libc "system:c"
@@ -14,7 +14,7 @@ when ODIN_OS == "windows" {
 //	EDOM,
 //	EILSEQ
 //	ERANGE
-when ODIN_OS == "linux" || ODIN_OS == "freebsd" {
+when ODIN_OS == .Linux || ODIN_OS == .FreeBSD {
 	@(private="file")
 	@(default_calling_convention="c")
 	foreign libc {
@@ -27,7 +27,20 @@ when ODIN_OS == "linux" || ODIN_OS == "freebsd" {
 	ERANGE :: 34
 }

-when ODIN_OS == "windows" {
+when ODIN_OS == .OpenBSD {
+	@(private="file")
+	@(default_calling_convention="c")
+	foreign libc {
+		@(link_name="__errno")
+		_get_errno :: proc() -> ^int ---
+	}
+
+	EDOM   :: 33
+	EILSEQ :: 84
+	ERANGE :: 34
+}
+
+when ODIN_OS == .Windows {
 	@(private="file")
 	@(default_calling_convention="c")
 	foreign libc {
@@ -40,7 +53,7 @@ when ODIN_OS == "windows" {
 	ERANGE :: 34
 }

-when ODIN_OS == "darwin" {
+when ODIN_OS == .Darwin {
 	@(private="file")
 	@(default_calling_convention="c")
 	foreign libc {
@@ -4,9 +4,9 @@ package libc

 import "core:intrinsics"

-when ODIN_OS == "windows" {
+when ODIN_OS == .Windows {
 	foreign import libc "system:libucrt.lib"
-} else when ODIN_OS == "darwin" {
+} else when ODIN_OS == .Darwin {
 	foreign import libc "system:System.framework"
 } else {
 	foreign import libc "system:c"
@@ -211,19 +211,19 @@ _signbitf :: #force_inline proc(x: float) -> int {
 	return int(transmute(uint32_t)x >> 31)
 }

-isfinite :: #force_inline proc(x: $T) where intrinsics.type_is_float(T) {
+isfinite :: #force_inline proc(x: $T) -> bool where intrinsics.type_is_float(T) {
 	return fpclassify(x) == FP_INFINITE
 }

-isinf :: #force_inline proc(x: $T) where intrinsics.type_is_float(T) {
+isinf :: #force_inline proc(x: $T) -> bool where intrinsics.type_is_float(T) {
 	return fpclassify(x) > FP_INFINITE
 }

-isnan :: #force_inline proc(x: $T) where intrinsics.type_is_float(T) {
+isnan :: #force_inline proc(x: $T) -> bool where intrinsics.type_is_float(T) {
 	return fpclassify(x) == FP_NAN
 }

-isnormal :: #force_inline proc(x: $T) where intrinsics.type_is_float(T) {
+isnormal :: #force_inline proc(x: $T) -> bool where intrinsics.type_is_float(T) {
 	return fpclassify(x) == FP_NORMAL
 }

@@ -231,27 +231,27 @@ isnormal :: #force_inline proc(x: $T) where intrinsics.type_is_float(T) {
 // implemented as the relational comparisons, as that would produce an invalid
 // "sticky" state that propagates and affects maths results. These need
 // to be implemented natively in Odin assuming isunordered to prevent that.
-isgreater :: #force_inline proc(x, y: $T) where intrinsics.type_is_float(T) {
+isgreater :: #force_inline proc(x, y: $T) -> bool where intrinsics.type_is_float(T) {
 	return !isunordered(x, y) && x > y
 }

-isgreaterequal :: #force_inline proc(x, y: $T) where intrinsics.type_is_float(T) {
+isgreaterequal :: #force_inline proc(x, y: $T) -> bool where intrinsics.type_is_float(T) {
 	return !isunordered(x, y) && x >= y
 }

-isless :: #force_inline proc(x, y: $T) where intrinsics.type_is_float(T) {
+isless :: #force_inline proc(x, y: $T) -> bool where intrinsics.type_is_float(T) {
 	return !isunordered(x, y) && x < y
 }

-islessequal :: #force_inline proc(x, y: $T) where intrinsics.type_is_float(T) {
+islessequal :: #force_inline proc(x, y: $T) -> bool where intrinsics.type_is_float(T) {
 	return !isunordered(x, y) && x <= y
 }

-islessgreater :: #force_inline proc(x, y: $T) where intrinsics.type_is_float(T) {
+islessgreater :: #force_inline proc(x, y: $T) -> bool where intrinsics.type_is_float(T) {
 	return !isunordered(x, y) && x <= y
 }

-isunordered :: #force_inline proc(x, y: $T) where intrinsics.type_is_float(T) {
+isunordered :: #force_inline proc(x, y: $T) -> bool where intrinsics.type_is_float(T) {
 	if isnan(x) {
 		// Force evaluation of y to propagate exceptions for ordering semantics.
 		// To ensure correct semantics of IEEE 754 this cannot be compiled away.
@@ -331,7 +331,7 @@ fmin       :: proc{libc_fmin, libc_fminf}
 fma        :: proc{libc_fma, libc_fmaf}

 // But retain the 'f' suffix-variant functions as well so they can be used,
-// a trick is used here where we use explicit procedrual overloading of one
+// a trick is used here where we use explicit procedural overloading of one
 // procedure. This is done because the foreign block is marked @(private) and
 // aliasing functions does not remove privateness from the entity.
 acosf      :: proc{libc_acosf}
@@ -2,14 +2,14 @@ package libc

 // 7.13 Nonlocal jumps

-when ODIN_OS == "windows" {
+when ODIN_OS == .Windows {
 	foreign import libc "system:libucrt.lib"
-} else when ODIN_OS == "darwin" {
+} else when ODIN_OS == .Darwin {
 	foreign import libc "system:System.framework"
 } else {
 	foreign import libc "system:c"
 }
-when ODIN_OS == "windows" {
+when ODIN_OS == .Windows {
 	@(default_calling_convention="c")
 	foreign libc {
 		// 7.13.1 Save calling environment
@@ -2,9 +2,9 @@ package libc

 // 7.14 Signal handling

-when ODIN_OS == "windows" {
+when ODIN_OS == .Windows {
 	foreign import libc "system:libucrt.lib"
-} else when ODIN_OS == "darwin" {
+} else when ODIN_OS == .Darwin {
 	foreign import libc "system:System.framework"
 } else {
 	foreign import libc "system:c"
@@ -21,7 +21,7 @@ foreign libc {
 	raise  :: proc(sig: int) -> int ---
 }

-when ODIN_OS == "windows" {
+when ODIN_OS == .Windows {
 	SIG_ERR :: rawptr(~uintptr(0)) 
 	SIG_DFL :: rawptr(uintptr(0))
 	SIG_IGN :: rawptr(uintptr(1))
@@ -34,7 +34,7 @@ when ODIN_OS == "windows" {
 	SIGTERM :: 15
 }

-when ODIN_OS == "linux" || ODIN_OS == "freebsd" {
+when ODIN_OS == .Linux || ODIN_OS == .FreeBSD {
 	SIG_ERR  :: rawptr(~uintptr(0))
 	SIG_DFL  :: rawptr(uintptr(0))
 	SIG_IGN  :: rawptr(uintptr(1)) 
@@ -47,7 +47,7 @@ when ODIN_OS == "linux" || ODIN_OS == "freebsd" {
 	SIGTERM  :: 15
 }

-when ODIN_OS == "darwin" {
+when ODIN_OS == .Darwin {
 	SIG_ERR  :: rawptr(~uintptr(0))
 	SIG_DFL  :: rawptr(uintptr(0))
 	SIG_IGN  :: rawptr(uintptr(1)) 
@@ -47,29 +47,30 @@ kill_dependency :: #force_inline proc(value: $T) -> T {

 // 7.17.4 Fences
 atomic_thread_fence :: #force_inline proc(order: memory_order) {
-	switch (order) {
-	case .relaxed:
-		return
-	case .consume:
-		intrinsics.atomic_fence_acq()
-	case .acquire:
-		intrinsics.atomic_fence_acq()
-	case .release:
-		intrinsics.atomic_fence_rel()
-	case .acq_rel:
-		intrinsics.atomic_fence_acqrel()
-	case .seq_cst:
-		intrinsics.atomic_fence_acqrel()
+	assert(order != .relaxed)
+	assert(order != .consume)
+	#partial switch order {
+	case .acquire: intrinsics.atomic_thread_fence(.Acquire)
+	case .release: intrinsics.atomic_thread_fence(.Release)
+	case .acq_rel: intrinsics.atomic_thread_fence(.Acq_Rel)
+	case .seq_cst: intrinsics.atomic_thread_fence(.Seq_Cst)
 	}
 }

 atomic_signal_fence :: #force_inline proc(order: memory_order) {
-	atomic_thread_fence(order)
+	assert(order != .relaxed)
+	assert(order != .consume)
+	#partial switch order {
+	case .acquire: intrinsics.atomic_signal_fence(.Acquire)
+	case .release: intrinsics.atomic_signal_fence(.Release)
+	case .acq_rel: intrinsics.atomic_signal_fence(.Acq_Rel)
+	case .seq_cst: intrinsics.atomic_signal_fence(.Seq_Cst)
+	}
 }

 // 7.17.5 Lock-free property
 atomic_is_lock_free :: #force_inline proc(obj: ^$T) -> bool {
-	return size_of(T) <= 8 && (intrinsics.type_is_integer(T) || intrinsics.type_is_pointer(T))
+	return intrinsics.atomic_type_is_lock_free(T)
 }

 // 7.17.6 Atomic integer types
@@ -121,13 +122,10 @@ atomic_store_explicit :: #force_inline proc(object: ^$T, desired: T, order: memo
 	assert(order != .acquire)
 	assert(order != .acq_rel)

-	#partial switch (order) {
-	case .relaxed:
-		intrinsics.atomic_store_relaxed(object, desired)
-	case .release:
-		intrinsics.atomic_store_rel(object, desired)
-	case .seq_cst:
-		intrinsics.atomic_store(object, desired)
+	#partial switch order {
+	case .relaxed: intrinsics.atomic_store_explicit(object, desired, .Relaxed)
+	case .release: intrinsics.atomic_store_explicit(object, desired, .Release)
+	case .seq_cst: intrinsics.atomic_store_explicit(object, desired, .Seq_Cst)
 	}
 }

@@ -139,36 +137,26 @@ atomic_load_explicit :: #force_inline proc(object: ^$T, order: memory_order) {
 	assert(order != .release)
 	assert(order != .acq_rel)

-	#partial switch (order) {
-	case .relaxed:
-		return intrinsics.atomic_load_relaxed(object)
-	case .consume:
-		return intrinsics.atomic_load_acq(object)
-	case .acquire:
-		return intrinsics.atomic_load_acq(object)
-	case .seq_cst:
-		return intrinsics.atomic_load(object)
+	#partial switch order {
+	case .relaxed: return intrinsics.atomic_load_explicit(object, .Relaxed)
+	case .consume: return intrinsics.atomic_load_explicit(object, .Consume)
+	case .acquire: return intrinsics.atomic_load_explicit(object, .Acquire)
+	case .seq_cst: return intrinsics.atomic_load_explicit(object, .Seq_Cst)
 	}
 }

 atomic_exchange :: #force_inline proc(object: ^$T, desired: T) -> T {
-	return intrinsics.atomic_xchg(object, desired)
+	return intrinsics.atomic_exchange(object, desired)
 }

 atomic_exchange_explicit :: #force_inline proc(object: ^$T, desired: T, order: memory_order) -> T {
-	switch (order) {
-	case .relaxed:
-		return intrinsics.atomic_xchg_relaxed(object, desired)
-	case .consume:
-		return intrinsics.atomic_xchg_acq(object, desired)
-	case .acquire:
-		return intrinsics.atomic_xchg_acq(object, desired)
-	case .release:
-		return intrinsics.atomic_xchg_rel(object, desired)
-	case .acq_rel:
-		return intrinsics.atomic_xchg_acqrel(object, desired)
-	case .seq_cst:
-		return intrinsics.atomic_xchg(object, desired)
+	switch order {
+	case .relaxed: return intrinsics.atomic_exchange_explicit(object, desired, .Relaxed)
+	case .consume: return intrinsics.atomic_exchange_explicit(object, desired, .Consume)
+	case .acquire: return intrinsics.atomic_exchange_explicit(object, desired, .Acquire)
+	case .release: return intrinsics.atomic_exchange_explicit(object, desired, .Release)
+	case .acq_rel: return intrinsics.atomic_exchange_explicit(object, desired, .Acq_Rel)
+	case .seq_cst: return intrinsics.atomic_exchange_explicit(object, desired, .Seq_Cst)
 	}
 	return false
 }
@@ -189,102 +177,104 @@ atomic_exchange_explicit :: #force_inline proc(object: ^$T, desired: T, order: m
 // 	[success = seq_cst, failure = acquire] => failacq
 // 	[success = acquire, failure = relaxed] => acq_failrelaxed
 // 	[success = acq_rel, failure = relaxed] => acqrel_failrelaxed
-atomic_compare_exchange_strong :: #force_inline proc(object, expected: ^$T, desired: T) {
-	value, ok := intrinsics.atomic_cxchg(object, expected^, desired)
+atomic_compare_exchange_strong :: #force_inline proc(object, expected: ^$T, desired: T) -> bool {
+	value, ok := intrinsics.atomic_compare_exchange_strong(object, expected^, desired)
 	if !ok { expected^ = value } 
 	return ok
 }

-atomic_compare_exchange_strong_explicit :: #force_inline proc(object, expected: ^$T, desired: T, success, failure: memory_order) {
+atomic_compare_exchange_strong_explicit :: #force_inline proc(object, expected: ^$T, desired: T, success, failure: memory_order) -> bool {
 	assert(failure != .release)
 	assert(failure != .acq_rel)

 	value: T; ok: bool
-	#partial switch (failure) {
+	#partial switch failure {
 	case .seq_cst:
 		assert(success != .relaxed)
-		#partial switch (success) {
+		#partial switch success {
 		case .seq_cst:
-			value, ok := intrinsics.atomic_cxchg(object, expected^, desired)
+			value, ok = intrinsics.atomic_compare_exchange_strong_explicit(object, expected^, desired, .Seq_Cst, .Seq_Cst)
 		case .acquire:
-			value, ok := intrinsics.atomic_cxchg_acq(object, expected^, desired)
+			value, ok = intrinsics.atomic_compare_exchange_strong_explicit(object, expected^, desired, .Acquire, .Seq_Cst)
 		case .consume:
-			value, ok := intrinsics.atomic_cxchg_acq(object, expected^, desired)
+			value, ok = intrinsics.atomic_compare_exchange_strong_explicit(object, expected^, desired, .Consume, .Seq_Cst)
 		case .release:
-			value, ok := intrinsics.atomic_cxchg_rel(object, expected^, desired)
+			value, ok = intrinsics.atomic_compare_exchange_strong_explicit(object, expected^, desired, .Release, .Seq_Cst)
 		case .acq_rel:
-			value, ok := intrinsics.atomic_cxchg_acqrel(object, expected^, desired)
+			value, ok = intrinsics.atomic_compare_exchange_strong_explicit(object, expected^, desired, .Acq_Rel, .Seq_Cst)
 		}
 	case .relaxed:
 		assert(success != .release)
-		#partial switch (success) {
+		#partial switch success {
 		case .relaxed:
-			value, ok := intrinsics.atomic_cxchg_relaxed(object, expected^, desired)
+			value, ok = intrinsics.atomic_compare_exchange_strong_explicit(object, expected^, desired, .Relaxed, .Relaxed)
 		case .seq_cst:
-			value, ok := intrinsics.atomic_cxchg_failrelaxed(object, expected^, desired)
+			value, ok = intrinsics.atomic_compare_exchange_strong_explicit(object, expected^, desired, .Seq_Cst, .Relaxed)
 		case .acquire:
-			value, ok := intrinsics.atomic_cxchg_acq_failrelaxed(object, expected^, desired)
+			value, ok = intrinsics.atomic_compare_exchange_strong_explicit(object, expected^, desired, .Acquire, .Relaxed)
 		case .consume:
-			value, ok := intrinsics.atomic_cxchg_acq_failrelaxed(object, expected^, desired)
+			value, ok = intrinsics.atomic_compare_exchange_strong_explicit(object, expected^, desired, .Consume, .Relaxed)
 		case .acq_rel:
-			value, ok := intrinsics.atomic_cxchg_acqrel_failrelaxed(object, expected^, desired)
+			value, ok = intrinsics.atomic_compare_exchange_strong_explicit(object, expected^, desired, .Acq_Rel, .Relaxed)
 		}
 	case .consume:
-		fallthrough
+		assert(success == .seq_cst)
+		value, ok = intrinsics.atomic_compare_exchange_strong_explicit(object, expected^, desired, .Seq_Cst, .Consume)
 	case .acquire:
 		assert(success == .seq_cst)
-		value, ok := intrinsics.atomic_cxchg_failacq(object, expected^, desired)
+		value, ok = intrinsics.atomic_compare_exchange_strong_explicit(object, expected^, desired, .Seq_Cst, .Acquire)

 	}
 	if !ok { expected^ = value }
 	return ok
 }

-atomic_compare_exchange_weak :: #force_inline proc(object, expected: ^$T, desired: T) {
-	value, ok := intrinsics.atomic_cxchgweak(object, expected^, desired)
+atomic_compare_exchange_weak :: #force_inline proc(object, expected: ^$T, desired: T) -> bool {
+	value, ok := intrinsics.atomic_compare_exchange_weak(object, expected^, desired)
 	if !ok { expected^ = value }
 	return ok
 }

-atomic_compare_exchange_weak_explicit :: #force_inline proc(object, expected: ^$T, desited: T, success, failure: memory_order) {
+atomic_compare_exchange_weak_explicit :: #force_inline proc(object, expected: ^$T, desited: T, success, failure: memory_order) -> bool {
 	assert(failure != .release)
 	assert(failure != .acq_rel)

 	value: T; ok: bool
-	#partial switch (failure) {
+	#partial switch failure {
 	case .seq_cst:
 		assert(success != .relaxed)
-		#partial switch (success) {
+		#partial switch success {
 		case .seq_cst:
-			value, ok := intrinsics.atomic_cxchgweak(object, expected^, desired)
+			value, ok = intrinsics.atomic_compare_exchange_weak_explicit(object, expected^, desired, .Seq_Cst, .Seq_Cst)
 		case .acquire:
-			value, ok := intrinsics.atomic_cxchgweak_acq(object, expected^, desired)
+			value, ok = intrinsics.atomic_compare_exchange_weak_explicit(object, expected^, desired, .Acquire, .Seq_Cst)
 		case .consume:
-			value, ok := intrinsics.atomic_cxchgweak_acq(object, expected^, desired)
+			value, ok = intrinsics.atomic_compare_exchange_weak_explicit(object, expected^, desired, .Consume, .Seq_Cst)
 		case .release:
-			value, ok := intrinsics.atomic_cxchgweak_rel(object, expected^, desired)
+			value, ok = intrinsics.atomic_compare_exchange_weak_explicit(object, expected^, desired, .Release, .Seq_Cst)
 		case .acq_rel:
-			value, ok := intrinsics.atomic_cxchgweak_acqrel(object, expected^, desired)
+			value, ok = intrinsics.atomic_compare_exchange_weak_explicit(object, expected^, desired, .Acq_Rel, .Seq_Cst)
 		}
 	case .relaxed:
 		assert(success != .release)
-		#partial switch (success) {
+		#partial switch success {
 		case .relaxed:
-			value, ok := intrinsics.atomic_cxchgweak_relaxed(object, expected^, desired)
+			value, ok = intrinsics.atomic_compare_exchange_weak_explicit(object, expected^, desired, .Relaxed, .Relaxed)
 		case .seq_cst:
-			value, ok := intrinsics.atomic_cxchgweak_failrelaxed(object, expected^, desired)
+			value, ok = intrinsics.atomic_compare_exchange_weak_explicit(object, expected^, desired, .Seq_Cst, .Relaxed)
 		case .acquire:
-			value, ok := intrinsics.atomic_cxchgweak_acq_failrelaxed(object, expected^, desired)
+			value, ok = intrinsics.atomic_compare_exchange_weak_explicit(object, expected^, desired, .Acquire, .Relaxed)
 		case .consume:
-			value, ok := intrinsics.atomic_cxchgweak_acq_failrelaxed(object, expected^, desired)
+			value, ok = intrinsics.atomic_compare_exchange_weak_explicit(object, expected^, desired, .Consume, .Relaxed)
 		case .acq_rel:
-			value, ok := intrinsics.atomic_cxchgweak_acqrel_failrelaxed(object, expected^, desired)
+			value, ok = intrinsics.atomic_compare_exchange_weak_explicit(object, expected^, desired, .Acq_Rel, .Relaxed)
 		}
 	case .consume:
-		fallthrough
+		assert(success == .seq_cst)
+		value, ok = intrinsics.atomic_compare_exchange_weak_explicit(object, expected^, desired, .Seq_Cst, .Consume)
 	case .acquire:
 		assert(success == .seq_cst)
-		value, ok := intrinsics.atomic_cxchgweak_failacq(object, expected^, desired)
+		value, ok = intrinsics.atomic_compare_exchange_weak_explicit(object, expected^, desired, .Seq_Cst, .Acquire)

 	}
 	if !ok { expected^ = value }
@@ -297,19 +287,14 @@ atomic_fetch_add :: #force_inline proc(object: ^$T, operand: T) -> T {
 }

 atomic_fetch_add_explicit :: #force_inline proc(object: ^$T, operand: T, order: memory_order) -> T {
-	switch (order) {
-	case .relaxed:
-		return intrinsics.atomic_add_relaxed(object, operand)
-	case .consume:
-		return intrinsics.atomic_add_acq(object, operand)
-	case .acquire:
-		return intrinsics.atomic_add_acq(object, operand)
-	case .release:
-		return intrinsics.atomic_add_rel(object, operand)
-	case .acq_rel:
-		return intrinsics.atomic_add_acqrel(object, operand)
-	case .seq_cst:
-		return intrinsics.atomic_add(object, operand)
+	switch order {
+	case .relaxed: return intrinsics.atomic_add_explicit(object, operand, .Relaxed)
+	case .consume: return intrinsics.atomic_add_explicit(object, operand, .Consume)
+	case .acquire: return intrinsics.atomic_add_explicit(object, operand, .Acquire)
+	case .release: return intrinsics.atomic_add_explicit(object, operand, .Release)
+	case .acq_rel: return intrinsics.atomic_add_explicit(object, operand, .Acq_Rel)
+	case: fallthrough
+	case .seq_cst: return intrinsics.atomic_add_explicit(object, operand, .Seq_Cst)
 	}
 }

@@ -318,19 +303,14 @@ atomic_fetch_sub :: #force_inline proc(object: ^$T, operand: T) -> T {
 }

 atomic_fetch_sub_explicit :: #force_inline proc(object: ^$T, operand: T, order: memory_order) -> T {
-	switch (order) {
-	case .relaxed:
-		return intrinsics.atomic_sub_relaxed(object, operand)
-	case .consume:
-		return intrinsics.atomic_sub_acq(object, operand)
-	case .acquire:
-		return intrinsics.atomic_sub_acq(object, operand)
-	case .release:
-		return intrinsics.atomic_sub_rel(object, operand)
-	case .acq_rel:
-		return intrinsics.atomic_sub_acqrel(object, operand)
-	case .seq_cst:
-		return intrinsics.atomic_sub(object, operand)
+	switch order {
+	case .relaxed: return intrinsics.atomic_sub_explicit(object, operand, .Relaxed)
+	case .consume: return intrinsics.atomic_sub_explicit(object, operand, .Consume)
+	case .acquire: return intrinsics.atomic_sub_explicit(object, operand, .Acquire)
+	case .release: return intrinsics.atomic_sub_explicit(object, operand, .Release)
+	case .acq_rel: return intrinsics.atomic_sub_explicit(object, operand, .Acq_Rel)
+	case: fallthrough
+	case .seq_cst: return intrinsics.atomic_sub_explicit(object, operand, .Seq_Cst)
 	}
 }

@@ -339,19 +319,14 @@ atomic_fetch_or :: #force_inline proc(object: ^$T, operand: T) -> T {
 }

 atomic_fetch_or_explicit :: #force_inline proc(object: ^$T, operand: T, order: memory_order) -> T {
-	switch (order) {
-	case .relaxed:
-		return intrinsics.atomic_or_relaxed(object, operand)
-	case .consume:
-		return intrinsics.atomic_or_acq(object, operand)
-	case .acquire:
-		return intrinsics.atomic_or_acq(object, operand)
-	case .release:
-		return intrinsics.atomic_or_rel(object, operand)
-	case .acq_rel:
-		return intrinsics.atomic_or_acqrel(object, operand)
-	case .seq_cst:
-		return intrinsics.atomic_or(object, operand)
+	switch order {
+	case .relaxed: return intrinsics.atomic_or_explicit(object, operand, .Relaxed)
+	case .consume: return intrinsics.atomic_or_explicit(object, operand, .Consume)
+	case .acquire: return intrinsics.atomic_or_explicit(object, operand, .Acquire)
+	case .release: return intrinsics.atomic_or_explicit(object, operand, .Release)
+	case .acq_rel: return intrinsics.atomic_or_explicit(object, operand, .Acq_Rel)
+	case: fallthrough
+	case .seq_cst: return intrinsics.atomic_or_explicit(object, operand, .Seq_Cst)
 	}
 }

@@ -360,19 +335,14 @@ atomic_fetch_xor :: #force_inline proc(object: ^$T, operand: T) -> T {
 }

 atomic_fetch_xor_explicit :: #force_inline proc(object: ^$T, operand: T, order: memory_order) -> T {
-	switch (order) {
-	case .relaxed:
-		return intrinsics.atomic_xor_relaxed(object, operand)
-	case .consume:
-		return intrinsics.atomic_xor_acq(object, operand)
-	case .acquire:
-		return intrinsics.atomic_xor_acq(object, operand)
-	case .release:
-		return intrinsics.atomic_xor_rel(object, operand)
-	case .acq_rel:
-		return intrinsics.atomic_xor_acqrel(object, operand)
-	case .seq_cst:
-		return intrinsics.atomic_xor(object, operand)
+	switch order {
+	case .relaxed: return intrinsics.atomic_xor_explicit(object, operand, .Relaxed)
+	case .consume: return intrinsics.atomic_xor_explicit(object, operand, .Consume)
+	case .acquire: return intrinsics.atomic_xor_explicit(object, operand, .Acquire)
+	case .release: return intrinsics.atomic_xor_explicit(object, operand, .Release)
+	case .acq_rel: return intrinsics.atomic_xor_explicit(object, operand, .Acq_Rel)
+	case: fallthrough
+	case .seq_cst: return intrinsics.atomic_xor_explicit(object, operand, .Seq_Cst)
 	}
 }

@@ -380,19 +350,14 @@ atomic_fetch_and :: #force_inline proc(object: ^$T, operand: T) -> T {
 	return intrinsics.atomic_and(object, operand)
 }
 atomic_fetch_and_explicit :: #force_inline proc(object: ^$T, operand: T, order: memory_order) -> T {
-	switch (order) {
-	case .relaxed:
-		return intrinsics.atomic_and_relaxed(object, operand)
-	case .consume:
-		return intrinsics.atomic_and_acq(object, operand)
-	case .acquire:
-		return intrinsics.atomic_and_acq(object, operand)
-	case .release:
-		return intrinsics.atomic_and_rel(object, operand)
-	case .acq_rel:
-		return intrinsics.atomic_and_acqrel(object, operand)
-	case .seq_cst:
-		return intrinsics.atomic_and(object, operand)
+	switch order {
+	case .relaxed: return intrinsics.atomic_and_explicit(object, operand, .Relaxed)
+	case .consume: return intrinsics.atomic_and_explicit(object, operand, .Consume)
+	case .acquire: return intrinsics.atomic_and_explicit(object, operand, .Acquire)
+	case .release: return intrinsics.atomic_and_explicit(object, operand, .Release)
+	case .acq_rel: return intrinsics.atomic_and_explicit(object, operand, .Acq_Rel)
+	case: fallthrough
+	case .seq_cst: return intrinsics.atomic_and_explicit(object, operand, .Seq_Cst)
 	}
 }

@@ -1,8 +1,11 @@
 package libc

-when ODIN_OS == "windows" {
-	foreign import libc "system:libucrt.lib"
-} else when ODIN_OS == "darwin" {
+when ODIN_OS == .Windows {
+	foreign import libc {
+		"system:libucrt.lib",
+		"system:legacy_stdio_definitions.lib",
+	}
+} else when ODIN_OS == .Darwin {
 	foreign import libc "system:System.framework"
 } else {
 	foreign import libc "system:c"
@@ -13,7 +16,7 @@ when ODIN_OS == "windows" {
 FILE :: struct {}

 // MSVCRT compatible.
-when ODIN_OS == "windows" {
+when ODIN_OS == .Windows {
 	_IOFBF       :: 0x0000
 	_IONBF       :: 0x0004
 	_IOLBF       :: 0x0040
@@ -48,7 +51,7 @@ when ODIN_OS == "windows" {
 }

 // GLIBC and MUSL compatible.
-when ODIN_OS == "linux" {
+when ODIN_OS == .Linux {
 	fpos_t        :: struct #raw_union { _: [16]char, _: longlong, _: double, }

 	_IOFBF        :: 0
@@ -78,7 +81,57 @@ when ODIN_OS == "linux" {
 	}
 }

-when ODIN_OS == "darwin" {
+when ODIN_OS == .OpenBSD {
+	fpos_t :: distinct i64
+
+	_IOFBF :: 0
+	_IOLBF :: 1
+	_IONBF :: 1
+
+	BUFSIZ :: 1024
+
+	EOF :: int(-1)
+
+	FOPEN_MAX	:: 20
+	FILENAME_MAX	:: 1024
+
+	SEEK_SET :: 0
+	SEEK_CUR :: 1
+	SEEK_END :: 2
+
+	foreign libc {
+		stderr: ^FILE
+		stdin:  ^FILE
+		stdout: ^FILE
+	}
+}
+
+when ODIN_OS == .FreeBSD {
+	fpos_t :: distinct i64
+
+	_IOFBF :: 0
+	_IOLBF :: 1
+	_IONBF :: 1
+
+	BUFSIZ :: 1024
+
+	EOF :: int(-1)
+
+	FOPEN_MAX	:: 20
+	FILENAME_MAX	:: 1024
+
+	SEEK_SET :: 0
+	SEEK_CUR :: 1
+	SEEK_END :: 2
+
+	foreign libc {
+		stderr: ^FILE
+		stdin:  ^FILE
+		stdout: ^FILE
+	}
+}
+
+when ODIN_OS == .Darwin {
 	fpos_t :: distinct i64
 	
 	_IOFBF        :: 0
@@ -146,10 +199,10 @@ foreign libc {
 	getc      :: proc(stream: ^FILE) -> int ---
 	getchar   :: proc() -> int ---
 	putc      :: proc(c: int, stream: ^FILE) -> int ---
-	putchar   :: proc() -> int ---
+	putchar   :: proc(c: int) -> int ---
 	puts      :: proc(s: cstring) -> int ---
 	ungetc    :: proc(c: int, stream: ^FILE) -> int ---
-	fread     :: proc(ptr: rawptr, size: size_t, stream: ^FILE) -> size_t ---
+	fread     :: proc(ptr: rawptr, size: size_t, nmemb: size_t, stream: ^FILE) -> size_t ---
 	fwrite    :: proc(ptr: rawptr, size: size_t, nmemb: size_t, stream: ^FILE) -> size_t ---

 	// 7.21.9 File positioning functions
@@ -2,15 +2,15 @@ package libc

 // 7.22 General utilities

-when ODIN_OS == "windows" {
+when ODIN_OS == .Windows {
 	foreign import libc "system:libucrt.lib"
-} else when ODIN_OS == "darwin" {
+} else when ODIN_OS == .Darwin {
 	foreign import libc "system:System.framework"
 } else {
 	foreign import libc "system:c"
 }

-when ODIN_OS == "windows" {
+when ODIN_OS == .Windows {
 	RAND_MAX :: 0x7fff

 	@(private="file")
@@ -24,7 +24,7 @@ when ODIN_OS == "windows" {
 	}
 }

-when ODIN_OS == "linux" {
+when ODIN_OS == .Linux {
 	RAND_MAX :: 0x7fffffff

 	// GLIBC and MUSL only
@@ -40,7 +40,7 @@ when ODIN_OS == "linux" {
 }


-when ODIN_OS == "darwin" {
+when ODIN_OS == .Darwin {
 	RAND_MAX :: 0x7fffffff

 	// GLIBC and MUSL only
@@ -88,7 +88,6 @@ foreign libc {
 	srand         :: proc(seed: uint) ---

 	// 7.22.3 Memory management functions
-	aligned_alloc :: proc(aligment, size: size_t) -> rawptr ---
 	calloc        :: proc(nmemb, size: size_t) -> rawptr ---
 	free          :: proc(ptr: rawptr) ---
 	malloc        :: proc(size: size_t) -> rawptr ---
@@ -125,3 +124,30 @@ foreign libc {
 	mbstowcs      :: proc(pwcs: ^wchar_t, s: cstring, n: size_t) -> size_t ---
 	wcstombs      :: proc(s: [^]char, pwcs: ^wchar_t, n: size_t) -> size_t ---
 }
+
+
+aligned_alloc :: #force_inline proc "c" (alignment, size: size_t) -> rawptr {
+	when ODIN_OS == .Windows {
+		foreign libc {
+			_aligned_malloc :: proc(size, alignment: size_t) -> rawptr ---
+		}
+		return _aligned_malloc(size=size, alignment=alignment)
+	} else {
+		foreign libc {
+			aligned_alloc :: proc(alignment, size: size_t) -> rawptr ---
+		}
+		return aligned_alloc(alignment=alignment, size=size)
+	}
+}
+
+
+aligned_free :: #force_inline proc "c" (ptr: rawptr) {
+	when ODIN_OS == .Windows {
+		foreign libc {
+			_aligned_free :: proc(ptr: rawptr) ---
+		}
+		_aligned_free(ptr)
+	} else {
+		free(ptr)
+	}
+}
@@ -4,9 +4,9 @@ import "core:runtime"

 // 7.24 String handling

-when ODIN_OS == "windows" {
+when ODIN_OS == .Windows {
 	foreign import libc "system:libucrt.lib"
-} else when ODIN_OS == "darwin" {
+} else when ODIN_OS == .Darwin {
 	foreign import libc "system:System.framework"
 } else {
 	foreign import libc "system:c"
@@ -5,10 +5,10 @@ package libc
 thrd_start_t :: proc "c" (rawptr) -> int
 tss_dtor_t   :: proc "c" (rawptr)

-when ODIN_OS == "windows" {
+when ODIN_OS == .Windows {
 	foreign import libc {
 		"system:libucrt.lib", 
-		"system:msvcprt.lib"
+		"system:msvcprt.lib",
 	}

 	thrd_success        :: 0                             // _Thrd_success
@@ -74,10 +74,10 @@ when ODIN_OS == "windows" {
 }

 // GLIBC and MUSL compatible constants and types.
-when ODIN_OS == "linux" {
+when ODIN_OS == .Linux {
 	foreign import libc {
 		"system:c",
-		"system:pthread"
+		"system:pthread",
 	}

 	thrd_success        :: 0
@@ -138,6 +138,6 @@ when ODIN_OS == "linux" {
 }


-when ODIN_OS == "darwin" {
+when ODIN_OS == .Darwin {
 	// TODO: find out what this is meant to be!
 }
@@ -2,9 +2,9 @@ package libc

 // 7.27 Date and time

-when ODIN_OS == "windows" {
+when ODIN_OS == .Windows {
 	foreign import libc "system:libucrt.lib"
-} else when ODIN_OS == "darwin" {
+} else when ODIN_OS == .Darwin {
 	foreign import libc "system:System.framework"
 } else {
 	foreign import libc "system:c"
@@ -12,7 +12,7 @@ when ODIN_OS == "windows" {

 // We enforce 64-bit time_t and timespec as there is no reason to use 32-bit as
 // we approach the 2038 problem. Windows has defaulted to this since VC8 (2005).
-when ODIN_OS == "windows" {
+when ODIN_OS == .Windows {
 	foreign libc {
 		// 7.27.2 Time manipulation functions
 		                               clock        :: proc() -> clock_t ---
@@ -45,7 +45,7 @@ when ODIN_OS == "windows" {
 	}
 }

-when ODIN_OS == "linux" || ODIN_OS == "freebsd" || ODIN_OS == "darwin" {
+when ODIN_OS == .Linux || ODIN_OS == .FreeBSD || ODIN_OS == .Darwin || ODIN_OS == .OpenBSD {
 	@(default_calling_convention="c")
 	foreign libc {
 		// 7.27.2 Time manipulation functions
@@ -63,7 +63,12 @@ when ODIN_OS == "linux" || ODIN_OS == "freebsd" || ODIN_OS == "darwin" {
 		strftime     :: proc(s: [^]char, maxsize: size_t, format: cstring, timeptr: ^tm) -> size_t ---
 	}

-	CLOCKS_PER_SEC :: 1000000
+	when ODIN_OS == .OpenBSD {
+		CLOCKS_PER_SEC :: 100
+	} else {
+		CLOCKS_PER_SEC :: 1000000
+	}
+
 	TIME_UTC       :: 1

 	time_t         :: distinct i64
@@ -3,6 +3,8 @@ package libc
 import "core:c"

 char           :: c.char // assuming -funsigned-char
+
+schar          :: c.schar
 short          :: c.short
 int            :: c.int
 long           :: c.long
@@ -2,9 +2,9 @@ package libc

 // 7.28 Unicode utilities

-when ODIN_OS == "windows" {
+when ODIN_OS == .Windows {
 	foreign import libc "system:libucrt.lib"
-} else when ODIN_OS == "darwin" {
+} else when ODIN_OS == .Darwin {
 	foreign import libc "system:System.framework"
 } else {
 	foreign import libc "system:c"
@@ -2,9 +2,9 @@ package libc

 // 7.29 Extended multibyte and wide character utilities

-when ODIN_OS == "windows" {
+when ODIN_OS == .Windows {
 	foreign import libc "system:libucrt.lib"
-} else when ODIN_OS == "darwin" {
+} else when ODIN_OS == .Darwin {
 	foreign import libc "system:System.framework"
 } else {
 	foreign import libc "system:c"
@@ -2,27 +2,34 @@ package libc

 // 7.30 Wide character classification and mapping utilities

-when ODIN_OS == "windows" {
+when ODIN_OS == .Windows {
 	foreign import libc "system:libucrt.lib"
-} else when ODIN_OS == "darwin" {
+} else when ODIN_OS == .Darwin {
 	foreign import libc "system:System.framework"
 } else {
 	foreign import libc "system:c"
 }

-when ODIN_OS == "windows" {
+when ODIN_OS == .Windows {
 	wctrans_t :: distinct wchar_t
 	wctype_t  :: distinct ushort
-}

-when ODIN_OS == "linux" {
+} else when ODIN_OS == .Linux {
 	wctrans_t :: distinct intptr_t
 	wctype_t  :: distinct ulong
-}

-when ODIN_OS == "darwin" {
+} else when ODIN_OS == .Darwin {
 	wctrans_t :: distinct int
 	wctype_t  :: distinct u32
+
+} else when ODIN_OS == .OpenBSD {
+	wctrans_t :: distinct rawptr
+	wctype_t  :: distinct rawptr
+
+} else when ODIN_OS == .FreeBSD {
+	wctrans_t :: distinct int
+	wctype_t  :: distinct ulong
+	
 }

@(default_calling_convention="c")
@@ -5,6 +5,9 @@
 	List of contributors:
 		Jeroen van Rijn: Initial implementation, optimization.
 */
+
+
+// package compress is a collection of utilities to aid with other compression packages
 package compress

 import "core:io"
@@ -44,7 +47,7 @@ when size_of(uintptr) == 8 {
 }


-Error :: union {
+Error :: union #shared_nil {
 	General_Error,
 	Deflate_Error,
 	ZLIB_Error,
@@ -55,6 +58,7 @@ Error :: union {
 }

 General_Error :: enum {
+	None = 0,
 	File_Not_Found,
 	Cannot_Open_File,
 	File_Too_Short,
@@ -73,6 +77,7 @@ General_Error :: enum {
 }

 GZIP_Error :: enum {
+	None = 0,
 	Invalid_GZIP_Signature,
 	Reserved_Flag_Set,
 	Invalid_Extra_Data,
@@ -97,6 +102,7 @@ GZIP_Error :: enum {
 }

 ZIP_Error :: enum {
+	None = 0,
 	Invalid_ZIP_File_Signature,
 	Unexpected_Signature,
 	Insert_Next_Disk,
@@ -104,6 +110,7 @@ ZIP_Error :: enum {
 }

 ZLIB_Error :: enum {
+	None = 0,
 	Unsupported_Window_Size,
 	FDICT_Unsupported,
 	Unsupported_Compression_Level,
@@ -111,6 +118,7 @@ ZLIB_Error :: enum {
 }

 Deflate_Error :: enum {
+	None = 0,
 	Huffman_Bad_Sizes,
 	Huffman_Bad_Code_Lengths,
 	Inflate_Error,
@@ -120,7 +128,6 @@ Deflate_Error :: enum {
 	BType_3,
 }

-
 // General I/O context for ZLIB, LZW, etc.
 Context_Memory_Input :: struct #packed {
 	input_data:        []u8,
@@ -136,7 +143,12 @@ Context_Memory_Input :: struct #packed {
 	size_packed:       i64,
 	size_unpacked:     i64,
 }
-#assert(size_of(Context_Memory_Input) == 64)
+when size_of(rawptr) == 8 {
+	#assert(size_of(Context_Memory_Input) == 64)
+} else {
+	// e.g. `-target:windows_i386`
+	#assert(size_of(Context_Memory_Input) == 52)
+}

 Context_Stream_Input :: struct #packed {
 	input_data:        []u8,
@@ -171,8 +183,6 @@ Context_Stream_Input :: struct #packed {
 	This simplifies end-of-stream handling where bits may be left in the bit buffer.
 */

-// TODO: Make these return compress.Error errors.
-
 input_size_from_memory :: proc(z: ^Context_Memory_Input) -> (res: i64, err: Error) {
 	return i64(len(z.input_data)), nil
 }
@@ -284,6 +294,24 @@ peek_data_from_memory :: #force_inline proc(z: ^Context_Memory_Input, $T: typeid
 	}
 }

+@(optimization_mode="speed")
+peek_data_at_offset_from_memory :: #force_inline proc(z: ^Context_Memory_Input, $T: typeid, #any_int offset: int) -> (res: T, err: io.Error) {
+	size :: size_of(T)
+
+	#no_bounds_check {
+		if len(z.input_data) >= size + offset {
+			buf := z.input_data[offset:][:size]
+			return (^T)(&buf[0])^, .None
+		}
+	}
+
+	if len(z.input_data) == 0 {
+		return T{}, .EOF
+	} else {
+		return T{}, .Short_Buffer
+	}
+}
+
@(optimization_mode="speed")
 peek_data_from_stream :: #force_inline proc(z: ^Context_Stream_Input, $T: typeid) -> (res: T, err: io.Error) {
 	size :: size_of(T)
@@ -311,7 +339,44 @@ peek_data_from_stream :: #force_inline proc(z: ^Context_Stream_Input, $T: typeid
 	return res, .None
 }

-peek_data :: proc{peek_data_from_memory, peek_data_from_stream}
+@(optimization_mode="speed")
+peek_data_at_offset_from_stream :: #force_inline proc(z: ^Context_Stream_Input, $T: typeid, #any_int offset: int) -> (res: T, err: io.Error) {
+	size :: size_of(T)
+
+	// Get current position to return to.
+	cur_pos, e1 := z.input->impl_seek(0, .Current)
+	if e1 != .None {
+		return T{}, e1
+	}
+
+	// Seek to offset.
+	pos, e2 := z.input->impl_seek(offset, .Start)
+	if e2 != .None {
+		return T{}, e2
+	}
+
+	r, e3 := io.to_reader_at(z.input)
+	if !e3 {
+		return T{}, .Empty
+	}
+	when size <= 128 {
+		b: [size]u8
+	} else {
+		b := make([]u8, size, context.temp_allocator)
+	}
+	_, e4 := io.read_at(r, b[:], pos)
+	if e4 != .None {
+		return T{}, .Empty
+	}
+
+	// Return read head to original position.
+	z.input->impl_seek(cur_pos, .Start)
+
+	res = (^T)(&b[0])^
+	return res, .None
+}
+
+peek_data :: proc{peek_data_from_memory, peek_data_from_stream, peek_data_at_offset_from_memory, peek_data_at_offset_from_stream}



@@ -470,4 +535,4 @@ discard_to_next_byte_lsb_from_stream :: proc(z: ^Context_Stream_Input) {
 	consume_bits_lsb(z, discard)
 }

-discard_to_next_byte_lsb :: proc{discard_to_next_byte_lsb_from_memory, discard_to_next_byte_lsb_from_stream};
+discard_to_next_byte_lsb :: proc{discard_to_next_byte_lsb_from_memory, discard_to_next_byte_lsb_from_stream}
@@ -45,7 +45,7 @@ main :: proc() {

 	if len(args) < 2 {
 		stderr("No input file specified.\n")
-		err := load(slice=TEST, buf=&buf, known_gzip_size=len(TEST))
+		err := load(data=TEST, buf=&buf, known_gzip_size=len(TEST))
 		if err == nil {
 			stdout("Displaying test vector: ")
 			stdout(bytes.buffer_to_string(&buf))
@@ -66,7 +66,8 @@ OS :: enum u8 {
 	_Unknown     = 14,
 	Unknown      = 255,
 }
-OS_Name :: #partial [OS]string{
+OS_Name :: #sparse[OS]string{
+	._Unknown     = "",
 	.FAT          = "FAT",
 	.Amiga        = "Amiga",
 	.VMS          = "VMS/OpenVMS",
@@ -99,9 +100,9 @@ E_GZIP    :: compress.GZIP_Error
 E_ZLIB    :: compress.ZLIB_Error
 E_Deflate :: compress.Deflate_Error

-GZIP_MAX_PAYLOAD_SIZE :: int(max(u32le))
+GZIP_MAX_PAYLOAD_SIZE :: i64(max(u32le))

-load :: proc{load_from_slice, load_from_file, load_from_context}
+load :: proc{load_from_bytes, load_from_file, load_from_context}

 load_from_file :: proc(filename: string, buf: ^bytes.Buffer, expected_output_size := -1, allocator := context.allocator) -> (err: Error) {
 	context.allocator = allocator
@@ -111,16 +112,16 @@ load_from_file :: proc(filename: string, buf: ^bytes.Buffer, expected_output_siz

 	err = E_General.File_Not_Found
 	if ok {
-		err = load_from_slice(data, buf, len(data), expected_output_size)
+		err = load_from_bytes(data, buf, len(data), expected_output_size)
 	}
 	return
 }

-load_from_slice :: proc(slice: []u8, buf: ^bytes.Buffer, known_gzip_size := -1, expected_output_size := -1, allocator := context.allocator) -> (err: Error) {
+load_from_bytes :: proc(data: []byte, buf: ^bytes.Buffer, known_gzip_size := -1, expected_output_size := -1, allocator := context.allocator) -> (err: Error) {
 	buf := buf

 	z := &compress.Context_Memory_Input{
-		input_data = slice,
+		input_data = data,
 		output = buf,
 	}
 	return load_from_context(z, buf, known_gzip_size, expected_output_size, allocator)
@@ -135,7 +136,7 @@ load_from_context :: proc(z: ^$C, buf: ^bytes.Buffer, known_gzip_size := -1, exp

 	z.output = buf

-	if expected_output_size > GZIP_MAX_PAYLOAD_SIZE {
+	if i64(expected_output_size) > i64(GZIP_MAX_PAYLOAD_SIZE) {
 		return E_GZIP.Payload_Size_Exceeds_Max_Payload
 	}

@@ -0,0 +1,148 @@
+/*
+	This file was generated, so don't edit this by hand.
+	Transliterated from https://github.com/Ed-von-Schleck/shoco/blob/master/shoco_model.h,
+	which is an English word model.
+*/
+
+// package shoco is an implementation of the shoco short string compressor
+package shoco
+
+DEFAULT_MODEL :: Shoco_Model {
+	min_char = 39,
+	max_char = 122,
+	characters_by_id = {
+		'e', 'a', 'i', 'o', 't', 'h', 'n', 'r', 's', 'l', 'u', 'c', 'w', 'm', 'd', 'b', 'p', 'f', 'g', 'v', 'y', 'k', '-', 'H', 'M', 'T', '\'', 'B', 'x', 'I', 'W', 'L',
+	},
+	ids_by_character = {
+		-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 26, -1, -1, -1, -1, -1, 22, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 27, -1, -1, -1, -1, -1, 23, 29, -1, -1, 31, 24, -1, -1, -1, -1, -1, -1, 25, -1, -1, 30, -1, -1, -1, -1, -1, -1, -1, -1, -1, 1, 15, 11, 14, 0, 17, 18, 5, 2, -1, 21, 9, 13, 6, 3, 16, -1, 7, 8, 4, 10, 19, 12, 28, 20, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
+	},
+	successors_by_bigram = {
+		7, 4, 12, -1, 6, -1, 1, 0, 3, 5, -1, 9, -1, 8, 2, -1, 15, 14, -1, 10, 11, -1, -1, -1, -1, -1, -1, -1, 13, -1, -1, -1,
+		1, -1, 6, -1, 1, -1, 0, 3, 2, 4, 15, 11, -1, 9, 5, 10, 13, -1, 12, 8, 7, 14, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
+		9, 11, -1, 4, 2, -1, 0, 8, 1, 5, -1, 6, -1, 3, 7, 15, -1, 12, 10, 13, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
+		-1, -1, 14, 7, 5, -1, 1, 2, 8, 9, 0, 15, 6, 4, 11, -1, 12, 3, -1, 10, -1, 13, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
+		2, 4, 3, 1, 5, 0, -1, 6, 10, 9, 7, 12, 11, -1, -1, -1, -1, 13, -1, -1, 8, -1, 15, -1, -1, -1, 14, -1, -1, -1, -1, -1,
+		0, 1, 2, 3, 4, -1, -1, 5, 9, 10, 6, -1, -1, 8, 15, 11, -1, 14, -1, -1, 7, -1, 13, -1, -1, -1, 12, -1, -1, -1, -1, -1,
+		2, 8, 7, 4, 3, -1, 9, -1, 6, 11, -1, 5, -1, -1, 0, -1, -1, 14, 1, 15, 10, 12, -1, -1, -1, -1, 13, -1, -1, -1, -1, -1,
+		0, 3, 1, 2, 6, -1, 9, 8, 4, 12, 13, 10, -1, 11, 7, -1, -1, 15, 14, -1, 5, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
+		0, 6, 3, 4, 1, 2, -1, -1, 5, 10, 7, 9, 11, 12, -1, -1, 8, 14, -1, -1, 15, 13, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
+		0, 6, 2, 5, 9, -1, -1, -1, 10, 1, 8, -1, 12, 14, 4, -1, 15, 7, -1, 13, 3, 11, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
+		8, 10, 9, 15, 1, -1, 4, 0, 3, 2, -1, 6, -1, 12, 11, 13, 7, 14, 5, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
+		1, 3, 6, 0, 4, 2, -1, 7, 13, 8, 9, 11, -1, -1, 15, -1, -1, -1, -1, -1, 10, 5, 14, -1, -1, -1, -1, -1, -1, -1, -1, -1,
+		3, 0, 1, 4, -1, 2, 5, 6, 7, 8, -1, 14, -1, -1, 9, 15, -1, 12, -1, -1, -1, 10, 11, -1, -1, -1, 13, -1, -1, -1, -1, -1,
+		0, 1, 3, 2, 15, -1, 12, -1, 7, 14, 4, -1, -1, 9, -1, 8, 5, 10, -1, -1, 6, -1, 13, -1, -1, -1, 11, -1, -1, -1, -1, -1,
+		0, 3, 1, 2, -1, -1, 12, 6, 4, 9, 7, -1, -1, 14, 8, -1, -1, 15, 11, 13, 5, -1, 10, -1, -1, -1, -1, -1, -1, -1, -1, -1,
+		0, 5, 7, 2, 10, 13, -1, 6, 8, 1, 3, -1, -1, 14, 15, 11, -1, -1, -1, 12, 4, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
+		0, 2, 6, 3, 7, 10, -1, 1, 9, 4, 8, -1, -1, 15, -1, 12, 5, -1, -1, -1, 11, -1, 13, -1, -1, -1, 14, -1, -1, -1, -1, -1,
+		1, 3, 4, 0, 7, -1, 12, 2, 11, 8, 6, 13, -1, -1, -1, -1, -1, 5, -1, -1, 10, 15, 9, -1, -1, -1, 14, -1, -1, -1, -1, -1,
+		1, 3, 5, 2, 13, 0, 9, 4, 7, 6, 8, -1, -1, 15, -1, 11, -1, -1, 10, -1, 14, -1, 12, -1, -1, -1, -1, -1, -1, -1, -1, -1,
+		0, 2, 1, 3, -1, -1, -1, 6, -1, -1, 5, -1, -1, -1, -1, -1, -1, -1, -1, -1, 4, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
+		1, 11, 4, 0, 3, -1, 13, 12, 2, 7, -1, -1, 15, 10, 5, 8, 14, -1, -1, -1, -1, -1, 9, -1, -1, -1, 6, -1, -1, -1, -1, -1,
+		0, 9, 2, 14, 15, 4, 1, 13, 3, 5, -1, -1, 10, -1, -1, -1, -1, 6, 12, -1, 7, -1, 8, -1, -1, -1, 11, -1, -1, -1, -1, -1,
+		-1, 2, 14, -1, 1, 5, 8, 7, 4, 12, -1, 6, 9, 11, 13, 3, 10, 15, -1, -1, -1, -1, 0, -1, -1, -1, -1, -1, -1, -1, -1, -1,
+		0, 1, 3, 2, -1, -1, -1, -1, -1, -1, 4, -1, -1, -1, -1, -1, -1, -1, -1, -1, 6, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
+		4, 3, 1, 5, -1, -1, -1, 0, -1, -1, 6, -1, -1, -1, -1, -1, -1, -1, -1, -1, 2, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
+		2, 8, 4, 1, -1, 0, -1, 6, -1, -1, 5, -1, 7, -1, -1, -1, -1, -1, -1, -1, 10, -1, -1, 9, -1, -1, -1, -1, -1, -1, -1, -1,
+		12, 5, -1, -1, 1, -1, -1, 7, 0, 3, -1, 2, -1, 4, 6, -1, -1, -1, -1, 8, -1, -1, 15, -1, 13, 9, -1, -1, -1, -1, -1, 11,
+		1, 3, 2, 4, -1, -1, -1, 5, -1, 7, 0, -1, -1, -1, -1, -1, -1, -1, -1, -1, 6, -1, -1, -1, -1, -1, -1, -1, -1, 8, -1, -1,
+		5, 3, 4, 12, 1, 6, -1, -1, -1, -1, 8, 2, -1, -1, -1, -1, 0, 9, -1, -1, 11, -1, 10, -1, -1, -1, -1, -1, -1, -1, -1, -1,
+		-1, -1, -1, -1, 0, -1, 1, 12, 3, -1, -1, -1, -1, 5, -1, -1, -1, 2, -1, -1, -1, -1, -1, -1, -1, -1, 4, -1, -1, 6, -1, 10,
+		2, 3, 1, 4, -1, 0, -1, 5, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 7, -1, -1, -1, -1, -1, -1, -1, -1, 6, -1, -1,
+		5, 1, 3, 0, -1, -1, -1, -1, -1, -1, 4, -1, -1, -1, -1, -1, -1, -1, -1, -1, 2, -1, -1, -1, -1, -1, 9, -1, -1, 6, -1, 7,
+	},
+	successors_reversed = {
+		's', 't', 'c', 'l', 'm', 'a', 'd', 'r', 'v', 'T', 'A', 'L', 'e', 'M', 'Y', '-',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'-', 't', 'a', 'b', 's', 'h', 'c', 'r', 'n', 'w', 'p', 'm', 'l', 'd', 'i', 'f',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'u', 'e', 'i', 'a', 'o', 'r', 'y', 'l', 'I', 'E', 'R', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'e', 'a', 'o', 'i', 'u', 'A', 'y', 'E', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		't', 'n', 'f', 's', '\'', 'm', 'I', 'N', 'A', 'E', 'L', 'Z', 'r', 'V', 'R', 'C',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'o', 'a', 'y', 'i', 'u', 'e', 'I', 'L', 'D', '\'', 'E', 'Y', '\x00', '\x00', '\x00', '\x00',
+		'r', 'i', 'y', 'a', 'e', 'o', 'u', 'Y', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'h', 'o', 'e', 'E', 'i', 'u', 'r', 'w', 'a', 'H', 'y', 'R', 'Z', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'h', 'i', 'e', 'a', 'o', 'r', 'I', 'y', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'n', 't', 's', 'r', 'l', 'd', 'i', 'y', 'v', 'm', 'b', 'c', 'g', 'p', 'k', 'u',
+		'e', 'l', 'o', 'u', 'y', 'a', 'r', 'i', 's', 'j', 't', 'b', 'v', 'h', 'm', 'd',
+		'o', 'e', 'h', 'a', 't', 'k', 'i', 'r', 'l', 'u', 'y', 'c', 'q', 's', '-', 'd',
+		'e', 'i', 'o', 'a', 's', 'y', 'r', 'u', 'd', 'l', '-', 'g', 'n', 'v', 'm', 'f',
+		'r', 'n', 'd', 's', 'a', 'l', 't', 'e', 'm', 'c', 'v', 'y', 'i', 'x', 'f', 'p',
+		'o', 'e', 'r', 'a', 'i', 'f', 'u', 't', 'l', '-', 'y', 's', 'n', 'c', '\'', 'k',
+		'h', 'e', 'o', 'a', 'r', 'i', 'l', 's', 'u', 'n', 'g', 'b', '-', 't', 'y', 'm',
+		'e', 'a', 'i', 'o', 't', 'r', 'u', 'y', 'm', 's', 'l', 'b', '\'', '-', 'f', 'd',
+		'n', 's', 't', 'm', 'o', 'l', 'c', 'd', 'r', 'e', 'g', 'a', 'f', 'v', 'z', 'b',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'e', 'n', 'i', 's', 'h', 'l', 'f', 'y', '-', 'a', 'w', '\'', 'g', 'r', 'o', 't',
+		'e', 'l', 'i', 'y', 'd', 'o', 'a', 'f', 'u', 't', 's', 'k', 'w', 'v', 'm', 'p',
+		'e', 'a', 'o', 'i', 'u', 'p', 'y', 's', 'b', 'm', 'f', '\'', 'n', '-', 'l', 't',
+		'd', 'g', 'e', 't', 'o', 'c', 's', 'i', 'a', 'n', 'y', 'l', 'k', '\'', 'f', 'v',
+		'u', 'n', 'r', 'f', 'm', 't', 'w', 'o', 's', 'l', 'v', 'd', 'p', 'k', 'i', 'c',
+		'e', 'r', 'a', 'o', 'l', 'p', 'i', 't', 'u', 's', 'h', 'y', 'b', '-', '\'', 'm',
+		'\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'e', 'i', 'o', 'a', 's', 'y', 't', 'd', 'r', 'n', 'c', 'm', 'l', 'u', 'g', 'f',
+		'e', 't', 'h', 'i', 'o', 's', 'a', 'u', 'p', 'c', 'l', 'w', 'm', 'k', 'f', 'y',
+		'h', 'o', 'e', 'i', 'a', 't', 'r', 'u', 'y', 'l', 's', 'w', 'c', 'f', '\'', '-',
+		'r', 't', 'l', 's', 'n', 'g', 'c', 'p', 'e', 'i', 'a', 'd', 'm', 'b', 'f', 'o',
+		'e', 'i', 'a', 'o', 'y', 'u', 'r', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+		'a', 'i', 'h', 'e', 'o', 'n', 'r', 's', 'l', 'd', 'k', '-', 'f', '\'', 'c', 'b',
+		'p', 't', 'c', 'a', 'i', 'e', 'h', 'q', 'u', 'f', '-', 'y', 'o', '\x00', '\x00', '\x00',
+		'o', 'e', 's', 't', 'i', 'd', '\'', 'l', 'b', '-', 'm', 'a', 'r', 'n', 'p', 'w',
+	},
+
+	character_count = 32,
+	successor_count = 16,
+
+	max_successor_n = 7,
+	packs = {
+		{ 0x80000000, 1, 2, { 26, 24, 24, 24, 24, 24, 24, 24 }, { 15,  3,  0,  0, 0, 0, 0, 0 }, 0xc0, 0x80 },
+		{ 0xc0000000, 2, 4, { 25, 22, 19, 16, 16, 16, 16, 16 }, { 15,  7,  7,  7, 0, 0, 0, 0 }, 0xe0, 0xc0 },
+		{ 0xe0000000, 4, 8, { 23, 19, 15, 11,  8,  5,  2,  0 }, { 31, 15, 15, 15, 7, 7, 7, 3 }, 0xf0, 0xe0 },
+	},
+}
@@ -0,0 +1,318 @@
+/*
+	Copyright 2022 Jeroen van Rijn <nom@duclavier.com>.
+	Made available under Odin's BSD-3 license.
+
+	List of contributors:
+		Jeroen van Rijn: Initial implementation.
+
+	An implementation of [shoco](https://github.com/Ed-von-Schleck/shoco) by Christian Schramm.
+*/
+
+// package shoco is an implementation of the shoco short string compressor
+package shoco
+
+import "core:intrinsics"
+import "core:compress"
+
+Shoco_Pack :: struct {
+	word:           u32,
+	bytes_packed:   i8,
+	bytes_unpacked: i8,
+	offsets:        [8]u16,
+	masks:          [8]i16,
+	header_mask:    u8,
+	header:         u8,
+}
+
+Shoco_Model :: struct {
+	min_char:             u8,
+	max_char:             u8,
+	characters_by_id:     []u8,
+	ids_by_character:     [256]i16,
+	successors_by_bigram: []i8,
+	successors_reversed:  []u8,
+
+	character_count:      u8,
+	successor_count:      u8,
+	max_successor_n:      i8,
+	packs:                []Shoco_Pack,
+}
+
+compress_bound :: proc(uncompressed_size: int) -> (worst_case_compressed_size: int) {
+	// Worst case compression happens when input is non-ASCII (128-255)
+	// Encoded as 0x00 + the byte in question.
+	return uncompressed_size * 2
+}
+
+decompress_bound :: proc(compressed_size: int, model := DEFAULT_MODEL) -> (maximum_decompressed_size: int) {
+	// Best case compression is 2:1
+	most: f64
+	for pack in model.packs {
+		val := f64(compressed_size) / f64(pack.bytes_packed) * f64(pack.bytes_unpacked)
+		most = max(most, val)
+	}
+	return int(most)
+}
+
+find_best_encoding :: proc(indices: []i16, n_consecutive: i8, model := DEFAULT_MODEL) -> (res: int) {
+	for p := len(model.packs); p > 0; p -= 1 {
+		pack := model.packs[p - 1]
+		if n_consecutive >= pack.bytes_unpacked {
+			have_index := true
+			for i := 0; i < int(pack.bytes_unpacked); i += 1 {
+				if indices[i] > pack.masks[i] {
+					have_index = false
+					break
+				}
+			}
+			if have_index {
+				return p - 1
+			}
+		}
+	}
+	return -1
+}
+
+validate_model :: proc(model: Shoco_Model) -> (int, compress.Error) {
+	if len(model.characters_by_id) != int(model.character_count) {
+		return 0, .Unknown_Compression_Method
+	}
+
+	if len(model.successors_by_bigram) != int(model.character_count) * int(model.character_count) {
+		return 0, .Unknown_Compression_Method
+	}
+
+	if len(model.successors_reversed) != int(model.successor_count) * int(model.max_char - model.min_char) {
+		return 0, .Unknown_Compression_Method
+	}
+
+	// Model seems legit.
+	return 0, nil
+}
+
+// Decompresses into provided buffer.
+decompress_slice_to_output_buffer :: proc(input: []u8, output: []u8, model := DEFAULT_MODEL) -> (size: int, err: compress.Error) {
+	inp, inp_end := 0, len(input)
+	out, out_end := 0, len(output)
+
+	validate_model(model) or_return
+
+	for inp < inp_end {
+		val  := transmute(i8)input[inp]
+		mark := int(-1)
+
+		for val < 0 {
+			val <<= 1
+			mark += 1
+		}
+
+		if mark > len(model.packs) {
+			return out, .Unknown_Compression_Method
+		}
+
+		if mark < 0 {
+			if out >= out_end {
+				return out, .Output_Too_Short
+			}
+
+			// Ignore the sentinel value for non-ASCII chars
+			if input[inp] == 0x00 {
+				inp += 1
+				if inp >= inp_end {
+					return out, .Stream_Too_Short
+				}
+			}
+			output[out] = input[inp]
+			inp, out = inp + 1, out + 1
+
+		} else {
+			pack := model.packs[mark]
+
+			if out + int(pack.bytes_unpacked) > out_end {
+				return out, .Output_Too_Short
+			} else if inp + int(pack.bytes_packed) > inp_end {
+				return out, .Stream_Too_Short
+			}
+
+			code := intrinsics.unaligned_load((^u32)(&input[inp]))
+			when ODIN_ENDIAN == .Little {
+				code = intrinsics.byte_swap(code)
+			}
+
+			// Unpack the leading char
+			offset := pack.offsets[0]
+			mask   := pack.masks[0]
+
+			last_chr := model.characters_by_id[(code >> offset) & u32(mask)]
+			output[out] = last_chr
+
+			// Unpack the successor chars
+			for i := 1; i < int(pack.bytes_unpacked); i += 1 {
+				offset = pack.offsets[i]
+				mask   = pack.masks[i]
+
+				index_major := u32(last_chr - model.min_char) * u32(model.successor_count)
+				index_minor := (code >> offset) & u32(mask)
+
+				last_chr = model.successors_reversed[index_major + index_minor]
+
+				output[out + i] = last_chr
+			}
+
+			out += int(pack.bytes_unpacked)
+			inp += int(pack.bytes_packed)
+		}
+	}
+
+	return out, nil
+}
+
+decompress_slice_to_string :: proc(input: []u8, model := DEFAULT_MODEL, allocator := context.allocator) -> (res: string, err: compress.Error) {
+	context.allocator = allocator
+
+	if len(input) == 0 {
+		return "", .Stream_Too_Short
+	}
+
+	max_output_size := decompress_bound(len(input), model)
+
+	buf: [dynamic]u8
+	if !resize(&buf, max_output_size) {
+		return "", .Out_Of_Memory
+	}
+
+	length, result := decompress_slice_to_output_buffer(input, buf[:])
+	resize(&buf, length)
+	return string(buf[:]), result
+}
+decompress :: proc{decompress_slice_to_output_buffer, decompress_slice_to_string}
+
+compress_string_to_buffer :: proc(input: string, output: []u8, model := DEFAULT_MODEL, allocator := context.allocator) -> (size: int, err: compress.Error) {
+	inp, inp_end := 0, len(input)
+	out, out_end := 0, len(output)
+	output := output
+
+	validate_model(model) or_return
+
+	indices := make([]i16, model.max_successor_n + 1)
+	defer delete(indices)
+
+	last_resort := false
+
+	encode: for inp < inp_end {
+		if last_resort {
+			last_resort = false
+
+			if input[inp] & 0x80 == 0x80 {
+				// Non-ASCII case
+				if out + 2 > out_end {
+					return out, .Output_Too_Short
+				}
+
+				// Put in a sentinel byte
+				output[out] = 0x00
+				out += 1
+			} else {
+				// An ASCII byte
+				if out + 1 > out_end {
+					return out, .Output_Too_Short
+				}
+			}
+			output[out] = input[inp]
+			out, inp = out + 1, inp + 1
+		} else {
+			// Find the longest string of known successors
+			indices[0] = model.ids_by_character[input[inp]]
+			last_chr_index := indices[0]
+
+			if last_chr_index < 0 {
+				last_resort = true
+				continue encode
+			}
+
+			rest := inp_end - inp
+			n_consecutive: i8 = 1
+			for ; n_consecutive <= model.max_successor_n; n_consecutive += 1 {
+				if inp_end > 0 && int(n_consecutive) == rest {
+					break
+				}
+
+				current_index := model.ids_by_character[input[inp + int(n_consecutive)]]
+				if current_index < 0 { // '\0' is always -1
+					break
+				}
+
+				successor_index := model.successors_by_bigram[last_chr_index * i16(model.character_count) + current_index]
+				if successor_index < 0 {
+					break
+				}
+
+				indices[n_consecutive] = i16(successor_index)
+				last_chr_index = current_index
+			}
+
+			if n_consecutive < 2 {
+				last_resort = true
+				continue encode
+			}
+
+			pack_n := find_best_encoding(indices, n_consecutive)
+			if pack_n >= 0 {
+				if out + int(model.packs[pack_n].bytes_packed) > out_end {
+					return out, .Output_Too_Short
+				}
+
+				pack := model.packs[pack_n]
+				code := pack.word
+
+				for i := 0; i < int(pack.bytes_unpacked); i += 1 {
+					code |= u32(indices[i]) << pack.offsets[i]
+				}
+
+				// In the little-endian world, we need to swap what's in the register to match the memory representation.
+				when ODIN_ENDIAN == .Little {
+					code = intrinsics.byte_swap(code)
+				}
+				out_ptr := raw_data(output[out:])
+
+				switch pack.bytes_packed {
+				case 4:
+					intrinsics.unaligned_store(transmute(^u32)out_ptr, code)
+				case 2:
+					intrinsics.unaligned_store(transmute(^u16)out_ptr, u16(code))
+				case 1:
+					intrinsics.unaligned_store(transmute(^u8)out_ptr,  u8(code))
+				case:
+					return out, .Unknown_Compression_Method
+				}
+
+				out += int(pack.bytes_packed)
+				inp += int(pack.bytes_unpacked)
+			} else {
+				last_resort = true
+				continue encode
+			}
+		}
+	}
+	return out, nil
+}
+
+compress_string :: proc(input: string, model := DEFAULT_MODEL, allocator := context.allocator) -> (output: []u8, err: compress.Error) {
+	context.allocator = allocator
+
+	if len(input) == 0 {
+		return {}, .Stream_Too_Short
+	}
+
+	max_output_size := compress_bound(len(input))
+
+	buf: [dynamic]u8
+	if !resize(&buf, max_output_size) {
+		return {}, .Out_Of_Memory
+	}
+
+	length, result := compress_string_to_buffer(input, buf[:])
+	resize(&buf, length)
+	return buf[:length], result
+}
+compress :: proc{compress_string_to_buffer, compress_string}
@@ -47,10 +47,10 @@ Options :: struct {
 	level: u8,
 }

-Error     :: compress.Error
-E_General :: compress.General_Error
-E_ZLIB    :: compress.ZLIB_Error
-E_Deflate :: compress.Deflate_Error
+Error         :: compress.Error
+General_Error :: compress.General_Error
+ZLIB_Error    :: compress.ZLIB_Error
+Deflate_Error :: compress.Deflate_Error

 DEFLATE_MAX_CHUNK_SIZE   :: 65535
 DEFLATE_MAX_LITERAL_SIZE :: 65535
@@ -111,9 +111,9 @@ ZFAST_MASK :: ((1 << ZFAST_BITS) - 1)
 */
 Huffman_Table :: struct {
 	fast:        [1 << ZFAST_BITS]u16,
-	firstcode:   [16]u16,
+	firstcode:   [17]u16,
 	maxcode:     [17]int,
-	firstsymbol: [16]u16,
+	firstsymbol: [17]u16,
 	size:        [288]u8,
 	value:       [288]u16,
 }
@@ -244,7 +244,7 @@ allocate_huffman_table :: proc(allocator := context.allocator) -> (z: ^Huffman_T
@(optimization_mode="speed")
 build_huffman :: proc(z: ^Huffman_Table, code_lengths: []u8) -> (err: Error) {
 	sizes:     [HUFFMAN_MAX_BITS+1]int
-	next_code: [HUFFMAN_MAX_BITS]int
+	next_code: [HUFFMAN_MAX_BITS+1]int

 	k := int(0)

@@ -256,21 +256,21 @@ build_huffman :: proc(z: ^Huffman_Table, code_lengths: []u8) -> (err: Error) {
 	}
 	sizes[0] = 0

-	for i in 1..<(HUFFMAN_MAX_BITS+1) {
+	for i in 1 ..< HUFFMAN_MAX_BITS {
 		if sizes[i] > (1 << uint(i)) {
-			return E_Deflate.Huffman_Bad_Sizes
+			return .Huffman_Bad_Sizes
 		}
 	}
 	code := int(0)

-	for i in 1..<HUFFMAN_MAX_BITS {
+	for i in 1 ..= HUFFMAN_MAX_BITS {
 		next_code[i]     = code
 		z.firstcode[i]   = u16(code)
 		z.firstsymbol[i] = u16(k)
 		code = code + sizes[i]
 		if sizes[i] != 0 {
 			if code - 1 >= (1 << u16(i)) {
-				return E_Deflate.Huffman_Bad_Code_Lengths
+				return .Huffman_Bad_Code_Lengths
 			}
 		}
 		z.maxcode[i] = code << (HUFFMAN_MAX_BITS - uint(i))
@@ -314,15 +314,15 @@ decode_huffman_slowpath :: proc(z: ^$C, t: ^Huffman_Table) -> (r: u16, err: Erro
 		s += 1
 	}
 	if s >= 16 {
-		return 0, E_Deflate.Bad_Huffman_Code
+		return 0, .Bad_Huffman_Code
 	}
 	// code size is s, so:
 	b := (k >> (16-s)) - int(t.firstcode[s]) + int(t.firstsymbol[s])
 	if b >= size_of(t.size) {
-		return 0, E_Deflate.Bad_Huffman_Code
+		return 0, .Bad_Huffman_Code
 	}
 	if t.size[b] != s {
-		return 0, E_Deflate.Bad_Huffman_Code
+		return 0, .Bad_Huffman_Code
 	}

 	compress.consume_bits_lsb(z, s)
@@ -335,11 +335,11 @@ decode_huffman_slowpath :: proc(z: ^$C, t: ^Huffman_Table) -> (r: u16, err: Erro
 decode_huffman :: proc(z: ^$C, t: ^Huffman_Table) -> (r: u16, err: Error) #no_bounds_check {
 	if z.num_bits < 16 {
 		if z.num_bits > 63 {
-			return 0, E_ZLIB.Code_Buffer_Malformed
+			return 0, .Code_Buffer_Malformed
 		}
 		compress.refill_lsb(z)
 		if z.num_bits > 63 {
-			return 0, E_General.Stream_Too_Short
+			return 0, .Stream_Too_Short
 		}
 	}
 	#no_bounds_check b := t.fast[z.code_buffer & ZFAST_MASK]
@@ -361,7 +361,7 @@ parse_huffman_block :: proc(z: ^$C, z_repeat, z_offset: ^Huffman_Table) -> (err:
 		if value < 256 {
 			e := write_byte(z, u8(value))
 			if e != .None {
-				return E_General.Output_Too_Short
+				return .Output_Too_Short
 			}
 		} else {
 			if value == 256 {
@@ -377,7 +377,7 @@ parse_huffman_block :: proc(z: ^$C, z_repeat, z_offset: ^Huffman_Table) -> (err:

 			value, e = decode_huffman(z, z_offset)
 			if e != nil {
-				return E_Deflate.Bad_Huffman_Code
+				return .Bad_Huffman_Code
 			}

 			distance := Z_DIST_BASE[value]
@@ -387,7 +387,7 @@ parse_huffman_block :: proc(z: ^$C, z_repeat, z_offset: ^Huffman_Table) -> (err:

 			if z.bytes_written < i64(distance) {
 				// Distance is longer than we've decoded so far.
-				return E_Deflate.Bad_Distance
+				return .Bad_Distance
 			}

 			/*
@@ -405,14 +405,14 @@ parse_huffman_block :: proc(z: ^$C, z_repeat, z_offset: ^Huffman_Table) -> (err:
 					c := z.output.buf[z.bytes_written - i64(distance)]
 					e := repl_byte(z, length, c)
 					if e != .None {
-						return E_General.Output_Too_Short
+						return .Output_Too_Short
 					}
 				}
 			} else {
 				if length > 0 {
 					e := repl_bytes(z, length, distance)
 					if e != .None {
-						return E_General.Output_Too_Short
+						return .Output_Too_Short
 					}
 				}
 			}
@@ -432,25 +432,25 @@ inflate_from_context :: proc(using ctx: ^compress.Context_Memory_Input, raw := f
 	if !raw {
 		size, size_err := compress.input_size(ctx)
 		if size < 6 || size_err != nil {
-			return E_General.Stream_Too_Short
+			return .Stream_Too_Short
 		}

 		cmf, _ := compress.read_u8(ctx)

 		method := Compression_Method(cmf & 0xf)
 		if method != .DEFLATE {
-			return E_General.Unknown_Compression_Method
+			return .Unknown_Compression_Method
 		}

 		if cinfo := (cmf >> 4) & 0xf; cinfo > 7 {
-			return E_ZLIB.Unsupported_Window_Size
+			return .Unsupported_Window_Size
 		}
 		flg, _ := compress.read_u8(ctx)

 		fcheck := flg & 0x1f
 		fcheck_computed := (cmf << 8 | flg) & 0x1f
 		if fcheck != fcheck_computed {
-			return E_General.Checksum_Failed
+			return .Checksum_Failed
 		}

 		/*
@@ -458,7 +458,7 @@ inflate_from_context :: proc(using ctx: ^compress.Context_Memory_Input, raw := f
 			They're application specific and PNG doesn't use them.
 		*/
 		if fdict := (flg >> 5) & 1; fdict != 0 {
-			return E_ZLIB.FDICT_Unsupported
+			return .FDICT_Unsupported
 		}

 		// flevel  := Compression_Level((flg >> 6) & 3);
@@ -485,7 +485,7 @@ inflate_from_context :: proc(using ctx: ^compress.Context_Memory_Input, raw := f
 		output_hash := hash.adler32(ctx.output.buf[:])

 		if output_hash != u32(adler) {
-			return E_General.Checksum_Failed
+			return .Checksum_Failed
 		}
 	}
 	return nil
@@ -538,23 +538,24 @@ inflate_raw :: proc(z: ^$C, expected_output_size := -1, allocator := context.all
 		final = compress.read_bits_lsb(z, 1)
 		type  = compress.read_bits_lsb(z, 2)

-		// fmt.printf("Final: %v | Type: %v\n", final, type);
+		// fmt.printf("Final: %v | Type: %v\n", final, type)

 		switch type {
 		case 0:
+			// fmt.printf("Method 0: STORED\n")
 			// Uncompressed block

 			// Discard bits until next byte boundary
 			compress.discard_to_next_byte_lsb(z)

-			uncompressed_len := i16(compress.read_bits_lsb(z, 16))
-			length_check     := i16(compress.read_bits_lsb(z, 16))
+			uncompressed_len := u16(compress.read_bits_lsb(z, 16))
+			length_check     := u16(compress.read_bits_lsb(z, 16))

-			// fmt.printf("LEN: %v, ~LEN: %v, NLEN: %v, ~NLEN: %v\n", uncompressed_len, ~uncompressed_len, length_check, ~length_check);
+			// fmt.printf("LEN: %v, ~LEN: %v, NLEN: %v, ~NLEN: %v\n", uncompressed_len, ~uncompressed_len, length_check, ~length_check)


 			if ~uncompressed_len != length_check {
-				return E_Deflate.Len_Nlen_Mismatch
+				return .Len_Nlen_Mismatch
 			}

 			/*
@@ -567,10 +568,12 @@ inflate_raw :: proc(z: ^$C, expected_output_size := -1, allocator := context.all
 				write_byte(z, u8(lit))
 				uncompressed_len -= 1
 			}
+			assert(uncompressed_len == 0)
+
 		case 3:
-			return E_Deflate.BType_3
+			return .BType_3
 		case:
-			// log.debugf("Err: %v | Final: %v | Type: %v\n", err, final, type);
+			// fmt.printf("Err: %v | Final: %v | Type: %v\n", err, final, type)
 			if type == 1 {
 				// Use fixed code lengths.
 				build_huffman(z_repeat, Z_FIXED_LENGTH[:]) or_return
@@ -601,7 +604,7 @@ inflate_raw :: proc(z: ^$C, expected_output_size := -1, allocator := context.all
 					c = decode_huffman(z, codelength_ht) or_return

 					if c < 0 || c >= 19 {
-						return E_Deflate.Huffman_Bad_Code_Lengths
+						return .Huffman_Bad_Code_Lengths
 					}
 					if c < 16 {
 						lencodes[n] = u8(c)
@@ -613,7 +616,7 @@ inflate_raw :: proc(z: ^$C, expected_output_size := -1, allocator := context.all
 						case 16:
 							c = u16(compress.read_bits_no_refill_lsb(z, 2) + 3)
 							if n == 0 {
-								return E_Deflate.Huffman_Bad_Code_Lengths
+								return .Huffman_Bad_Code_Lengths
 							}
 							fill = lencodes[n - 1]
 						case 17:
@@ -621,11 +624,11 @@ inflate_raw :: proc(z: ^$C, expected_output_size := -1, allocator := context.all
 						case 18:
 							c = u16(compress.read_bits_no_refill_lsb(z, 7) + 11)
 						case:
-								return E_Deflate.Huffman_Bad_Code_Lengths
+								return .Huffman_Bad_Code_Lengths
 						}

 						if ntot - n < u32(c) {
-							return E_Deflate.Huffman_Bad_Code_Lengths
+							return .Huffman_Bad_Code_Lengths
 						}

 						nc := n + u32(c)
@@ -636,7 +639,7 @@ inflate_raw :: proc(z: ^$C, expected_output_size := -1, allocator := context.all
 				}

 				if n != ntot {
-					return E_Deflate.Huffman_Bad_Code_Lengths
+					return .Huffman_Bad_Code_Lengths
 				}

 				build_huffman(z_repeat, lencodes[:hlit])     or_return
@@ -674,4 +677,4 @@ inflate_from_byte_array_raw :: proc(input: []u8, buf: ^bytes.Buffer, raw := fals
 	return inflate_raw(z=&ctx, expected_output_size=expected_output_size)
 }

-inflate :: proc{inflate_from_context, inflate_from_byte_array};
+inflate :: proc{inflate_from_context, inflate_from_byte_array}
@@ -1,216 +0,0 @@
-package container
-
-import "core:mem"
-import "core:runtime"
-
-Array :: struct($T: typeid) {
-	data:      ^T,
-	len:       int,
-	cap:       int,
-	allocator: mem.Allocator,
-}
-
-ARRAY_DEFAULT_CAPACITY :: 16
-
-/*
-array_init :: proc {
-	array_init_none,
-	array_init_len,
-	array_init_len_cap,
-}
-array_init
-array_delete
-array_len
-array_cap
-array_space
-array_slice
-array_get
-array_get_ptr
-array_set
-array_reserve
-array_resize
-array_push = array_append :: proc{
-	array_push_back,
-	array_push_back_elems,
-}
-array_push_front
-array_pop_back
-array_pop_front
-array_consume
-array_trim
-array_clear
-array_clone
-array_set_capacity
-array_grow
-*/
-
-
-array_init_none :: proc(a: ^$A/Array, allocator := context.allocator) {
-	array_init_len_cap(a, 0, ARRAY_DEFAULT_CAPACITY, allocator)
-}
-array_init_len :: proc(a: ^$A/Array, len: int, allocator := context.allocator) {
-	array_init_len_cap(a, len, len, allocator)
-}
-array_init_len_cap :: proc(a: ^$A/Array($T), len: int, cap: int, allocator := context.allocator) {
-	a.allocator = allocator
-	a.data = (^T)(mem.alloc(size_of(T)*cap, align_of(T), a.allocator))
-	a.len = len
-	a.cap = cap
-}
-
-array_init :: proc{array_init_none, array_init_len, array_init_len_cap}
-
-array_delete :: proc(a: $A/Array) {
-	mem.free(a.data, a.allocator)
-}
-
-array_len :: proc(a: $A/Array) -> int {
-	return a.len
-}
-
-array_cap :: proc(a: $A/Array) -> int {
-	return a.cap
-}
-
-array_space :: proc(a: $A/Array) -> int {
-	return a.cap - a.len
-}
-
-array_slice :: proc(a: $A/Array($T)) -> []T {
-	s := mem.Raw_Slice{a.data, a.len}
-	return transmute([]T)s
-}
-
-array_cap_slice :: proc(a: $A/Array($T)) -> []T {
-	s := mem.Raw_Slice{a.data, a.cap}
-	return transmute([]T)s
-}
-
-array_get :: proc(a: $A/Array($T), index: int, loc := #caller_location) -> T {
-	runtime.bounds_check_error_loc(loc, index, array_len(a))
-	return (^T)(uintptr(a.data) + size_of(T)*uintptr(index))^
-}
-array_get_ptr :: proc(a: $A/Array($T), index: int, loc := #caller_location) -> ^T {
-	runtime.bounds_check_error_loc(loc, index, array_len(a))
-	return (^T)(uintptr(a.data) + size_of(T)*uintptr(index))
-}
-
-array_set :: proc(a: ^$A/Array($T), index: int, item: T, loc := #caller_location)  {
-	runtime.bounds_check_error_loc(loc, index, array_len(a^))
-	(^T)(uintptr(a.data) + size_of(T)*uintptr(index))^ = item
-}
-
-
-array_reserve :: proc(a: ^$A/Array, capacity: int) {
-	if capacity > a.len {
-		array_set_capacity(a, capacity)
-	}
-}
-
-array_resize :: proc(a: ^$A/Array, length: int) {
-	if length > a.len {
-		array_set_capacity(a, length)
-	}
-	a.len = length
-}
-
-
-
-array_push_back :: proc(a: ^$A/Array($T), item: T) {
-	if array_space(a^) == 0 {
-		array_grow(a)
-	}
-
-	a.len += 1
-	array_set(a, a.len-1, item)
-}
-
-array_push_front :: proc(a: ^$A/Array($T), item: T) {
-	if array_space(a^) == 0 {
-		array_grow(a)
-	}
-
-	a.len += 1
-	data := array_slice(a^)
-	copy(data[1:], data[:])
-	data[0] = item
-}
-
-array_pop_back :: proc(a: ^$A/Array($T), loc := #caller_location) -> T {
-	assert(condition=a.len > 0, loc=loc)
-	item := array_get(a^, a.len-1)
-	a.len -= 1
-	return item
-}
-
-array_pop_front :: proc(a: ^$A/Array($T), loc := #caller_location) -> T {
-	assert(condition=a.len > 0, loc=loc)
-	item := array_get(a^, 0)
-	s := array_slice(a^)
-	copy(s[:], s[1:])
-	a.len -= 1
-	return item
-}
-
-
-array_consume :: proc(a: ^$A/Array($T), count: int, loc := #caller_location) {
-	assert(condition=a.len >= count, loc=loc)
-	a.len -= count
-}
-
-
-array_trim :: proc(a: ^$A/Array($T)) {
-	array_set_capacity(a, a.len)
-}
-
-array_clear :: proc(a: ^$A/Array($T)) {
-	array_resize(a, 0)
-}
-
-array_clone :: proc(a: $A/Array($T), allocator := context.allocator) -> A {
-	res: A
-	array_init(&res, array_len(a), array_len(a), allocator)
-	copy(array_slice(res), array_slice(a))
-	return res
-}
-
-array_push_back_elems :: proc(a: ^$A/Array($T), items: ..T) {
-	if array_space(a^) < len(items) {
-		array_grow(a, a.len + len(items))
-	}
-	offset := a.len
-	data := array_cap_slice(a^)
-	n := copy(data[a.len:], items)
-	a.len += n
-}
-
-array_push   :: proc{array_push_back, array_push_back_elems}
-array_append :: proc{array_push_back, array_push_back_elems}
-
-array_set_capacity :: proc(a: ^$A/Array($T), new_capacity: int) {
-	if new_capacity == a.cap {
-		return
-	}
-
-	if new_capacity < a.len {
-		array_resize(a, new_capacity)
-	}
-
-	new_data: ^T
-	if new_capacity > 0 {
-		if a.allocator.procedure == nil {
-			a.allocator = context.allocator
-		}
-		new_data = (^T)(mem.alloc(size_of(T)*new_capacity, align_of(T), a.allocator))
-		if new_data != nil {
-			mem.copy(new_data, a.data, size_of(T)*a.len)
-		}
-	}
-	mem.free(a.data, a.allocator)
-	a.data = new_data
-	a.cap = new_capacity
-}
-array_grow :: proc(a: ^$A/Array, min_capacity: int = 0) {
-	new_capacity := max(array_len(a^)*2 + 8, min_capacity)
-	array_set_capacity(a, new_capacity)
-}
@@ -0,0 +1,266 @@
+package dynamic_bit_array
+
+import "core:intrinsics"
+import "core:mem"
+
+/*
+	Note that these constants are dependent on the backing being a u64.
+*/
+@(private="file")
+INDEX_SHIFT :: 6
+
+@(private="file")
+INDEX_MASK  :: 63
+
+@(private="file")
+NUM_BITS :: 64
+
+Bit_Array :: struct {
+	bits:         [dynamic]u64,
+	bias:         int,
+	max_index:    int,
+	free_pointer: bool,
+}
+
+Bit_Array_Iterator :: struct {
+	array:    ^Bit_Array,
+	word_idx: int,
+	bit_idx:  uint,
+}
+
+/*
+	In:
+		- ba:   ^Bit_Array - the array to iterate over
+
+	Out:
+		- it:   ^Bit_Array_Iterator - the iterator that holds iteration state
+*/
+make_iterator :: proc (ba: ^Bit_Array) -> (it: Bit_Array_Iterator) {
+	return Bit_Array_Iterator { array = ba }
+}
+
+/*
+	In:
+		- it:    ^Bit_Array_Iterator - the iterator struct that holds the state.
+
+	Out:
+		- set:    bool - the state of the bit at `index`
+		- index:  int - the next bit of the Bit_Array referenced by `it`.
+		- ok:	  bool - `true` if the iterator returned a valid index,
+			  `false` if there were no more bits
+*/
+iterate_by_all :: proc (it: ^Bit_Array_Iterator) -> (set: bool, index: int, ok: bool) {
+	index = it.word_idx * NUM_BITS + int(it.bit_idx) + it.array.bias
+	if index > it.array.max_index { return false, 0, false }
+
+	word := it.array.bits[it.word_idx] if len(it.array.bits) > it.word_idx else 0
+	set = (word >> it.bit_idx & 1) == 1
+
+	it.bit_idx += 1
+	if it.bit_idx >= NUM_BITS {
+		it.bit_idx = 0
+		it.word_idx += 1
+	}
+
+	return set, index, true
+}
+
+/*
+	In:
+		- it:     ^Bit_Array_Iterator - the iterator struct that holds the state.
+
+	Out:
+		- index:  int - the next set bit of the Bit_Array referenced by `it`.
+		- ok:	  bool - `true` if the iterator returned a valid index,
+			  `false` if there were no more bits set
+*/
+iterate_by_set :: proc (it: ^Bit_Array_Iterator) -> (index: int, ok: bool) {
+	return iterate_internal_(it, true)
+}
+
+/*
+	In:
+		- it:	  ^Bit_Array_Iterator - the iterator struct that holds the state.
+
+	Out:
+		- index:  int - the next unset bit of the Bit_Array referenced by `it`.
+		- ok:	  bool - `true` if the iterator returned a valid index,
+			  `false` if there were no more unset bits
+*/
+iterate_by_unset:: proc (it: ^Bit_Array_Iterator) -> (index: int, ok: bool) {
+	return iterate_internal_(it, false)
+}
+
+@(private="file")
+iterate_internal_ :: proc (it: ^Bit_Array_Iterator, $ITERATE_SET_BITS: bool) -> (index: int, ok: bool) {
+	word := it.array.bits[it.word_idx] if len(it.array.bits) > it.word_idx else 0
+	when ! ITERATE_SET_BITS { word = ~word }
+
+	// if the word is empty or we have already gone over all the bits in it,
+	// b.bit_idx is greater than the index of any set bit in the word,
+	// meaning that word >> b.bit_idx == 0.
+	for it.word_idx < len(it.array.bits) && word >> it.bit_idx == 0 {
+		it.word_idx += 1
+		it.bit_idx = 0
+		word = it.array.bits[it.word_idx] if len(it.array.bits) > it.word_idx else 0
+		when ! ITERATE_SET_BITS { word = ~word }
+	}
+
+	// if we are iterating the set bits, reaching the end of the array means we have no more bits to check
+	when ITERATE_SET_BITS {
+		if it.word_idx >= len(it.array.bits) {
+			return 0, false
+		}
+	}
+
+	// reaching here means that the word has some set bits
+	it.bit_idx += uint(intrinsics.count_trailing_zeros(word >> it.bit_idx))
+	index = it.word_idx * NUM_BITS + int(it.bit_idx) + it.array.bias
+
+	it.bit_idx += 1
+	if it.bit_idx >= NUM_BITS {
+		it.bit_idx = 0
+		it.word_idx += 1
+	}
+	return index, index <= it.array.max_index
+}
+
+
+/*
+	In:
+		- ba:    ^Bit_Array - a pointer to the Bit Array
+		- index: The bit index. Can be an enum member.
+
+	Out:
+		- res:   The bit you're interested in.
+		- ok:    Whether the index was valid. Returns `false` if the index is smaller than the bias.
+
+	The `ok` return value may be ignored.
+*/
+get :: proc(ba: ^Bit_Array, #any_int index: uint, allocator := context.allocator) -> (res: bool, ok: bool) {
+	idx := int(index) - ba.bias
+
+	if ba == nil || int(index) < ba.bias { return false, false }
+	context.allocator = allocator
+
+	leg_index := idx >> INDEX_SHIFT
+	bit_index := idx &  INDEX_MASK
+
+	/*
+		If we `get` a bit that doesn't fit in the Bit Array, it's naturally `false`.
+		This early-out prevents unnecessary resizing.
+	*/
+	if leg_index + 1 > len(ba.bits) { return false, true }
+
+	val := u64(1 << uint(bit_index))
+	res = ba.bits[leg_index] & val == val
+
+	return res, true
+}
+
+/*
+	In:
+		- ba:    ^Bit_Array - a pointer to the Bit Array
+		- index: The bit index. Can be an enum member.
+
+	Out:
+		- ok:    Whether or not we managed to set requested bit.
+
+	`set` automatically resizes the Bit Array to accommodate the requested index if needed.
+*/
+set :: proc(ba: ^Bit_Array, #any_int index: uint, allocator := context.allocator) -> (ok: bool) {
+
+	idx := int(index) - ba.bias
+
+	if ba == nil || int(index) < ba.bias { return false }
+	context.allocator = allocator
+
+	leg_index := idx >> INDEX_SHIFT
+	bit_index := idx &  INDEX_MASK
+
+	resize_if_needed(ba, leg_index) or_return
+
+	ba.max_index = max(idx, ba.max_index)
+	ba.bits[leg_index] |= 1 << uint(bit_index)
+	return true
+}
+
+/*
+	In:
+		- ba:    ^Bit_Array - a pointer to the Bit Array
+		- index: The bit index. Can be an enum member.
+
+	Out:
+		- ok:    Whether or not we managed to unset requested bit.
+
+	`unset` automatically resizes the Bit Array to accommodate the requested index if needed.
+*/
+unset :: proc(ba: ^Bit_Array, #any_int index: uint, allocator := context.allocator) -> (ok: bool) {
+
+	idx := int(index) - ba.bias
+
+	if ba == nil || int(index) < ba.bias { return false }
+	context.allocator = allocator
+
+	leg_index := idx >> INDEX_SHIFT
+	bit_index := idx &  INDEX_MASK
+
+	resize_if_needed(ba, leg_index) or_return
+
+	ba.max_index = max(idx, ba.max_index)
+	ba.bits[leg_index] &= ~(1 << uint(bit_index))
+	return true
+}
+
+/*
+	A helper function to create a Bit Array with optional bias, in case your smallest index is non-zero (including negative).
+*/
+create :: proc(max_index: int, min_index := 0, allocator := context.allocator) -> (res: ^Bit_Array, ok: bool) #optional_ok {
+	context.allocator = allocator
+	size_in_bits := max_index - min_index
+
+	if size_in_bits < 1 { return {}, false }
+
+	legs := size_in_bits >> INDEX_SHIFT
+
+	res = new(Bit_Array)
+	res.bias         = min_index
+	res.max_index    = max_index
+	res.free_pointer = true
+	return res, resize_if_needed(res, legs)
+}
+
+/*
+	Sets all bits to `false`.
+*/
+clear :: proc(ba: ^Bit_Array) {
+	if ba == nil { return }
+	mem.zero_slice(ba.bits[:])
+}
+
+/*
+	Releases the memory used by the Bit Array.
+*/
+destroy :: proc(ba: ^Bit_Array) {
+	if ba == nil { return }
+	delete(ba.bits)
+	if ba.free_pointer { // Only free if this Bit_Array was created using `create`, not when on the stack.
+		free(ba)
+	}
+}
+
+/*
+	Resizes the Bit Array. For internal use.
+	If you want to reserve the memory for a given-sized Bit Array up front, you can use `create`.
+*/
+@(private="file")
+resize_if_needed :: proc(ba: ^Bit_Array, legs: int, allocator := context.allocator) -> (ok: bool) {
+	if ba == nil { return false }
+
+	context.allocator = allocator
+
+	if legs + 1 > len(ba.bits) {
+		resize(&ba.bits, legs + 1)
+	}
+	return len(ba.bits) > legs
+}
@@ -0,0 +1,53 @@
+package dynamic_bit_array
+
+/*
+	The Bit Array can be used in several ways:
+
+	-- By default you don't need to instantiate a Bit Array:
+
+		package test
+
+		import "core:fmt"
+		import "core:container/bit_array"
+
+		main :: proc() {
+			using bit_array
+
+			bits: Bit_Array
+
+			// returns `true`
+			fmt.println(set(&bits, 42))
+
+			// returns `false`, `false`, because this Bit Array wasn't created to allow negative indices.
+			was_set, was_retrieved := get(&bits, -1)
+			fmt.println(was_set, was_retrieved) 
+			destroy(&bits)
+		}
+
+	-- A Bit Array can optionally allow for negative indices, if the mininum value was given during creation:
+
+		package test
+
+		import "core:fmt"
+		import "core:container/bit_array"
+
+		main :: proc() {
+			Foo :: enum int {
+				Negative_Test = -42,
+				Bar           = 420,
+				Leaves        = 69105,
+			}
+
+			using bit_array
+
+			bits := create(int(max(Foo)), int(min(Foo)))
+			defer destroy(bits)
+
+			fmt.printf("Set(Bar):           %v\n",     set(bits, Foo.Bar))
+			fmt.printf("Get(Bar):           %v, %v\n", get(bits, Foo.Bar))
+			fmt.printf("Set(Negative_Test): %v\n",     set(bits, Foo.Negative_Test))
+			fmt.printf("Get(Leaves):        %v, %v\n", get(bits, Foo.Leaves))
+			fmt.printf("Get(Negative_Test): %v, %v\n", get(bits, Foo.Negative_Test))
+			fmt.printf("Freed.\n")
+		}
+*/
@@ -1,80 +0,0 @@
-package container
-
-import "core:mem"
-
-Bloom_Hash_Proc :: #type proc(data: []byte) -> u32
-
-Bloom_Hash :: struct {
-	hash_proc: Bloom_Hash_Proc,
-	next:     ^Bloom_Hash,
-}
-
-Bloom_Filter :: struct {
-	allocator: mem.Allocator,
-	hash:      ^Bloom_Hash,
-	bits:      []byte,
-}
-
-bloom_filter_init :: proc(b: ^Bloom_Filter, size: int, allocator := context.allocator) {
-	b.allocator = allocator
-	b.bits = make([]byte, size, allocator)
-}
-
-bloom_filter_destroy :: proc(b: ^Bloom_Filter) {
-	context.allocator = b.allocator
-	delete(b.bits)
-	for b.hash != nil {
-		hash := b.hash
-		b.hash = b.hash.next
-		free(hash)
-	}
-}
-
-bloom_filter_add_hash_proc :: proc(b: ^Bloom_Filter, hash_proc: Bloom_Hash_Proc) {
-	context.allocator = b.allocator
-	h := new(Bloom_Hash)
-	h.hash_proc = hash_proc
-
-	head := &b.hash
-	for head^ != nil {
-		head = &(head^.next)
-	}
-	head^ = h
-}
-
-bloom_filter_add :: proc(b: ^Bloom_Filter, item: []byte) {
-	#no_bounds_check for h := b.hash; h != nil; h = h.next {
-		hash := h.hash_proc(item)
-		hash %= u32(len(b.bits) * 8)
-		b.bits[hash >> 3] |= 1 << (hash & 3)
-	}
-}
-
-bloom_filter_add_string :: proc(b: ^Bloom_Filter, item: string) {
-	bloom_filter_add(b, transmute([]byte)item)
-}
-
-bloom_filter_add_raw :: proc(b: ^Bloom_Filter, data: rawptr, size: int) {
-	item := mem.slice_ptr((^byte)(data), size)
-	bloom_filter_add(b, item)
-}
-
-bloom_filter_test :: proc(b: ^Bloom_Filter, item: []byte) -> bool {
-	#no_bounds_check for h := b.hash; h != nil; h = h.next {
-		hash := h.hash_proc(item)
-		hash %= u32(len(b.bits) * 8)
-		if (b.bits[hash >> 3] & (1 << (hash & 3)) == 0) {
-			return false
-		}
-	}
-	return true
-}
-
-bloom_filter_test_string :: proc(b: ^Bloom_Filter, item: string) -> bool {
-	return bloom_filter_test(b, transmute([]byte)item)
-}
-
-bloom_filter_test_raw :: proc(b: ^Bloom_Filter, data: rawptr, size: int) -> bool {
-	item := mem.slice_ptr((^byte)(data), size)
-	return bloom_filter_test(b, item)
-}
@@ -0,0 +1,173 @@
+package container_intrusive_list
+
+import "core:intrinsics"
+
+// An intrusive doubly-linked list
+//
+// As this is an intrusive container, a `Node` must be embedded in your own
+// structure which is conventionally called a "link". The use of `push_front`
+// and `push_back` take the address of this node. Retrieving the data
+// associated with the node requires finding the relative offset of the node
+// of the parent structure. The parent type and field name are given to
+// `iterator_*` procedures, or to the built-in `container_of` procedure.
+//
+// This data structure is two-pointers in size:
+// 	8 bytes on 32-bit platforms and 16 bytes on 64-bit platforms
+List :: struct {
+	head: ^Node,
+	tail: ^Node,
+}
+
+
+Node :: struct {
+	next, prev: ^Node,
+}
+
+push_front :: proc(list: ^List, node: ^Node) {
+	if list.head != nil {
+		list.head.prev = node
+		node.prev, node.next = nil, list.head
+		list.head = node
+	} else {
+		list.head, list.tail = node, node
+		node.prev, node.next = nil, nil
+	}
+}
+
+push_back :: proc(list: ^List, node: ^Node) {
+	if list.tail != nil {
+		list.tail.next = node
+		node.prev, node.next = list.tail, nil
+		list.tail = node
+	} else {
+		list.head, list.tail = node, node
+		node.prev, node.next = nil, nil
+	}
+}
+
+remove :: proc(list: ^List, node: ^Node) {
+	if node != nil {
+		if node.next != nil {
+			node.next.prev = node.prev
+		}
+		if node.prev != nil {
+			node.prev.next = node.next
+		}
+		if list.head == node {
+			list.head = node.next
+		}
+		if list.tail == node {
+			list.tail = node.prev
+		}
+	}
+}
+
+remove_by_proc :: proc(list: ^List, to_erase: proc(^Node) -> bool) {
+	for node := list.head; node != nil; {
+		next := node.next
+		if to_erase(node) {
+			if node.next != nil {
+				node.next.prev = node.prev
+			}
+			if node.prev != nil {
+				node.prev.next = node.next
+			}
+			if list.head == node {
+				list.head = node.next
+			}
+			if list.tail == node {
+				list.tail = node.prev
+			}
+		}
+		node = next
+	}
+}
+
+
+is_empty :: proc(list: ^List) -> bool {
+	return list.head == nil
+}
+
+pop_front :: proc(list: ^List) -> ^Node {
+	link := list.head
+	if link == nil {
+		return nil
+	}
+	if link.next != nil {
+		link.next.prev = link.prev
+	}
+	if link.prev != nil {
+		link.prev.next = link.next
+	}
+	if link == list.head {
+		list.head = link.next
+	}
+	if link == list.tail {
+		list.tail = link.prev
+	}
+	return link
+
+}
+pop_back :: proc(list: ^List) -> ^Node {
+	link := list.tail
+	if link == nil {
+		return nil
+	}
+	if link.next != nil {
+		link.next.prev = link.prev
+	}
+	if link.prev != nil {
+		link.prev.next = link.next
+	}
+	if link == list.head {
+		list.head = link.next
+	}
+	if link == list.tail {
+		list.tail = link.prev
+	}
+	return link
+}
+
+
+Iterator :: struct($T: typeid) {
+	curr:   ^Node,
+	offset: uintptr,
+}
+
+iterator_head :: proc(list: List, $T: typeid, $field_name: string) -> Iterator(T)
+	where intrinsics.type_has_field(T, field_name),
+	      intrinsics.type_field_type(T, field_name) == Node {
+	return {list.head, offset_of_by_string(T, field_name)}
+}
+
+iterator_tail :: proc(list: List, $T: typeid, $field_name: string) -> Iterator(T)
+	where intrinsics.type_has_field(T, field_name),
+	      intrinsics.type_field_type(T, field_name) == Node {
+	return {list.tail, offset_of_by_string(T, field_name)}
+}
+
+iterator_from_node :: proc(node: ^Node, $T: typeid, $field_name: string) -> Iterator(T)
+	where intrinsics.type_has_field(T, field_name),
+	      intrinsics.type_field_type(T, field_name) == Node {
+	return {node, offset_of_by_string(T, field_name)}
+}
+
+iterate_next :: proc(it: ^Iterator($T)) -> (ptr: ^T, ok: bool) {
+	node := it.curr
+	if node == nil {
+		return nil, false
+	}
+	it.curr = node.next
+
+	return (^T)(uintptr(node) - it.offset), true
+}
+
+iterate_prev :: proc(it: ^Iterator($T)) -> (ptr: ^T, ok: bool) {
+	node := it.curr
+	if node == nil {
+		return nil, false
+	}
+	it.curr = node.prev
+
+	return (^T)(uintptr(node) - it.offset), true
+}
@@ -0,0 +1,201 @@
+package container_lru
+
+import "core:runtime"
+import "core:intrinsics"
+_ :: runtime
+_ :: intrinsics
+
+Node :: struct($Key, $Value: typeid) where intrinsics.type_is_valid_map_key(Key) {
+	prev, next: ^Node(Key, Value),
+	key:   Key,
+	value: Value,
+}
+
+// Cache is an LRU cache. It automatically removes entries as new entries are
+// added if the capacity is reached. Entries are removed based on how recently
+// they were used where the oldest entries are removed first.
+Cache :: struct($Key, $Value: typeid) where intrinsics.type_is_valid_map_key(Key) {
+	head: ^Node(Key, Value),
+	tail: ^Node(Key, Value),
+
+	entries: map[Key]^Node(Key, Value),
+
+	count:    int,
+	capacity: int,
+
+	node_allocator: runtime.Allocator,
+
+	on_remove: proc(key: Key, value: Value, user_data: rawptr),
+	on_remove_user_data: rawptr,
+}
+
+// init initializes a Cache
+init :: proc(c: ^$C/Cache($Key, $Value), capacity: int, entries_allocator := context.allocator, node_allocator := context.allocator) {
+	c.entries.allocator = entries_allocator
+	c.node_allocator = node_allocator
+	c.capacity = capacity
+}
+
+// destroy deinitializes a Cachem
+destroy :: proc(c: ^$C/Cache($Key, $Value), call_on_remove: bool) {
+	clear(c, call_on_remove)
+	delete(c.entries)
+}
+
+// clear the contents of a Cache
+clear :: proc(c: ^$C/Cache($Key, $Value), call_on_remove: bool) {
+	for _, node in c.entries {
+		if call_on_remove {
+			_call_on_remove(c, node)
+		}
+		free(node, c.node_allocator)
+	}
+	runtime.clear(&c.entries)
+	c.head = nil
+	c.tail = nil
+	c.count = 0
+}
+
+// set the given key value pair. This operation updates the recent usage of the item.
+set :: proc(c: ^$C/Cache($Key, $Value), key: Key, value: Value) -> runtime.Allocator_Error {
+	if e, ok := c.entries[key]; ok {
+		e.value = value
+		_pop_node(c, e)
+		_push_front_node(c, e)
+		return nil
+	}
+
+	e : ^Node(Key, Value) = nil
+	assert(c.count <= c.capacity)
+	if c.count == c.capacity {
+		e = c.tail
+		_remove_node(c, e)
+	}
+	else {
+		c.count += 1
+		e = new(Node(Key, Value), c.node_allocator) or_return
+	}
+
+	e.key = key
+	e.value = value
+	_push_front_node(c, e)
+	c.entries[key] = e
+
+	return nil
+}
+
+// get a value from the cache from a given key. This operation updates the usage of the item.
+get :: proc(c: ^$C/Cache($Key, $Value), key: Key) -> (value: Value, ok: bool) #optional_ok {
+	e: ^Node(Key, Value)
+	e, ok = c.entries[key]
+	if !ok {
+		return
+	}
+	_pop_node(c, e)
+	_push_front_node(c, e)
+	return e.value, true
+}
+
+// get_ptr gets the pointer to a value the cache from a given key. This operation updates the usage of the item.
+get_ptr :: proc(c: ^$C/Cache($Key, $Value), key: Key) -> (value: ^Value, ok: bool) #optional_ok {
+	e: ^Node(Key, Value)
+	e, ok = c.entries[key]
+	if !ok {
+		return
+	}
+	_pop_node(c, e)
+	_push_front_node(c, e)
+	return &e.value, true
+}
+
+// peek gets the value from the cache from a given key without updating the recent usage.
+peek :: proc(c: ^$C/Cache($Key, $Value), key: Key) -> (value: Value, ok: bool) #optional_ok {
+	e: ^Node(Key, Value)
+	e, ok = c.entries[key]
+	if !ok {
+		return
+	}
+	return e.value, true
+}
+
+// exists checks for the existence of a value from a given key without updating the recent usage.
+exists :: proc(c: ^$C/Cache($Key, $Value), key: Key) -> bool {
+	return key in c.entries
+}
+
+// remove removes an item from the cache.
+remove :: proc(c: ^$C/Cache($Key, $Value), key: Key) -> bool {
+	e, ok := c.entries[key]
+	if !ok {
+		return false
+	}
+	_remove_node(c, e)
+	free(node, c.node_allocator)
+	c.count -= 1
+	return true
+}
+
+
+@(private)
+_remove_node :: proc(c: ^$C/Cache($Key, $Value), node: ^Node(Key, Value)) {
+	if c.head == node {
+		c.head = node.next
+	}
+	if c.tail == node {
+		c.tail = node.prev
+	}
+	if node.prev != nil {
+		node.prev.next = node.next
+	}
+	if node.next != nil {
+		node.next.prev = node.prev
+	}
+	node.prev = nil
+	node.next = nil
+
+	delete_key(&c.entries, node.key)
+
+	_call_on_remove(c, node)
+}
+
+@(private)
+_call_on_remove :: proc(c: ^$C/Cache($Key, $Value), node: ^Node(Key, Value)) {
+	if c.on_remove != nil {
+		c.on_remove(node.key, node.value, c.on_remove_user_data)
+	}
+}
+
+@(private)
+_push_front_node :: proc(c: ^$C/Cache($Key, $Value), e: ^Node(Key, Value)) {
+	if c.head != nil {
+		e.next = c.head
+		e.next.prev = e
+	}
+	c.head = e
+	if c.tail == nil {
+		c.tail = e
+	}
+	e.prev = nil
+}
+
+@(private)
+_pop_node :: proc(c: ^$C/Cache($Key, $Value), e: ^Node(Key, Value)) {
+	if e == nil {
+		return
+	}
+	if c.head == e {
+		c.head = e.next
+	}
+	if c.tail == e {
+		c.tail = e.prev
+	}
+	if e.prev != nil {
+		e.prev.next = e.next
+	}
+
+	if e.next != nil {
+		e.next.prev = e.prev
+	}
+	e.prev = nil
+	e.next = nil
+}
@@ -1,377 +0,0 @@
-package container
-
-import "core:intrinsics"
-_ :: intrinsics
-
-
-Map :: struct($Key, $Value: typeid) where intrinsics.type_is_valid_map_key(Key) {
-	hash: Array(int),
-	entries: Array(Map_Entry(Key, Value)),
-}
-
-Map_Entry :: struct($Key, $Value: typeid) where intrinsics.type_is_valid_map_key(Key) {
-	hash:  uintptr,
-	next:  int,
-	key:   Key,
-	value: Value,
-}
-
-
-/*
-map_init :: proc{
-	map_init_none,
-	map_init_cap,
-}
-map_delete
-
-map_has
-map_get
-map_get_default
-map_get_ptr
-map_set
-map_remove
-map_reserve
-map_clear
-
-// Multi Map
-
-multi_map_find_first
-multi_map_find_next
-multi_map_count
-multi_map_get :: proc{
-	multi_map_get_array,
-	multi_map_get_slice,
-};
-multi_map_get_as_slice
-multi_map_insert
-multi_map_remove
-multi_map_remove_all
-
-*/
-
-map_init :: proc{map_init_none, map_init_cap}
-
-map_init_none :: proc(m: ^$M/Map($Key, $Value), allocator := context.allocator) {
-	m.hash.allocator = allocator
-	m.entries.allocator = allocator
-}
-
-map_init_cap :: proc(m: ^$M/Map($Key, $Value), cap: int, allocator := context.allocator) {
-	m.hash.allocator = allocator
-	m.entries.allocator = allocator
-	map_reserve(m, cap)
-}
-
-map_delete :: proc(m: $M/Map($Key, $Value)) {
-	array_delete(m.hash)
-	array_delete(m.entries)
-}
-
-
-map_has :: proc(m: $M/Map($Key, $Value), key: Key) -> bool {
-	return _map_find_or_fail(m, key) >= 0
-}
-
-map_get :: proc(m: $M/Map($Key, $Value), key: Key) -> (res: Value, ok: bool) #optional_ok {
-	i := _map_find_or_fail(m, key)
-	if i < 0 {
-		return {}, false
-	}
-	return array_get(m.entries, i).value, true
-}
-
-map_get_default :: proc(m: $M/Map($Key, $Value), key: Key, default: Value) -> (res: Value, ok: bool) #optional_ok {
-	i := _map_find_or_fail(m, key)
-	if i < 0 {
-		return default, false
-	}
-	return array_get(m.entries, i).value, true
-}
-
-map_get_ptr :: proc(m: $M/Map($Key, $Value), key: Key) -> ^Value {
-	i := _map_find_or_fail(m, key)
-	if i < 0 {
-		return nil
-	}
-	return array_get_ptr(m.entries, i).value
-}
-
-map_set :: proc(m: ^$M/Map($Key, $Value), key: Key, value: Value) {
-	if array_len(m.hash) == 0 {
-		_map_grow(m)
-	}
-
-	i := _map_find_or_make(m, key)
-	array_get_ptr(m.entries, i).value = value
-	if _map_full(m^) {
-		_map_grow(m)
-	}
-}
-
-map_remove :: proc(m: ^$M/Map($Key, $Value), key: Key) {
-	fr := _map_find_key(m^, key)
-	if fr.entry_index >= 0 {
-		_map_erase(m, fr)
-	}
-}
-
-
-map_reserve :: proc(m: ^$M/Map($Key, $Value), new_size: int) {
-	nm: M
-	map_init(&nm, m.hash.allocator)
-	array_resize(&nm.hash, new_size)
-	array_reserve(&nm.entries, array_len(m.entries))
-
-	for i in 0..<new_size {
-		array_set(&nm.hash, i, -1)
-	}
-	for i in 0..<array_len(m.entries) {
-		e := array_get(m.entries, i)
-		multi_map_insert(&nm, e.key, e.value)
-	}
-
-	map_delete(m^)
-	m^ = nm
-}
-
-map_clear :: proc(m: ^$M/Map($Key, $Value)) {
-	array_clear(&m.hash)
-	array_clear(&m.entries)
-}
-
-
-
-multi_map_find_first :: proc(m: $M/Map($Key, $Value), key: Key) -> ^Map_Entry(Key, Value) {
-	i := _map_find_or_fail(m, key)
-	if i < 0 {
-		return nil
-	}
-	return array_get_ptr(m.entries, i)
-}
-
-multi_map_find_next :: proc(m: $M/Map($Key, $Value), e: ^Map_Entry(Key, Value)) -> ^Map_Entry(Key, Value) {
-	i := e.next
-	for i >= 0 {
-		it := array_get_ptr(m.entries, i)
-		if it.hash == e.hash && it.key == e.key {
-			return it
-		}
-		i = it.next
-	}
-	return nil
-}
-
-multi_map_count :: proc(m: $M/Map($Key, $Value), key: Key) -> int {
-	n := 0
-	e := multi_map_find_first(m, key)
-	for e != nil {
-		n += 1
-		e = multi_map_find_next(m, e)
-	}
-	return n
-}
-
-multi_map_get :: proc{multi_map_get_array, multi_map_get_slice}
-
-multi_map_get_array :: proc(m: $M/Map($Key, $Value), key: Key, items: ^Array(Value)) {
-	if items == nil {
-		return
-	}
-	e := multi_map_find_first(m, key)
-	for e != nil {
-		array_append(items, e.value)
-		e = multi_map_find_next(m, e)
-	}
-}
-
-multi_map_get_slice :: proc(m: $M/Map($Key, $Value), key: Key, items: []Value) {
-	e := multi_map_find_first(m, key)
-	i := 0
-	for e != nil && i < len(items) {
-		items[i] = e.value
-		i += 1
-		e = multi_map_find_next(m, e)
-	}
-}
-
-multi_map_get_as_slice :: proc(m: $M/Map($Key, $Value), key: Key) -> []Value {
-	items: Array(Value)
-	array_init(&items, 0)
-
-	e := multi_map_find_first(m, key)
-	for e != nil {
-		array_append(&items, e.value)
-		e = multi_map_find_next(m, e)
-	}
-
-	return array_slice(items)
-}
-
-
-multi_map_insert :: proc(m: ^$M/Map($Key, $Value), key: Key, value: Value) {
-	if array_len(m.hash) == 0 {
-		_map_grow(m)
-	}
-
-	i := _map_make(m, key)
-	array_get_ptr(m.entries, i).value = value
-	if _map_full(m^) {
-		_map_grow(m)
-	}
-}
-
-multi_map_remove :: proc(m: ^$M/Map($Key, $Value), e: ^Map_Entry(Key, Value)) {
-	fr := _map_find_entry(m, e)
-	if fr.entry_index >= 0 {
-		_map_erase(m, fr)
-	}
-}
-
-multi_map_remove_all :: proc(m: ^$M/Map($Key, $Value), key: Key) {
-	for map_exist(m^, key) {
-		map_remove(m, key)
-	}
-}
-
-
-/// Internal
-
-
-Map_Find_Result :: struct {
-	hash_index:  int,
-	entry_prev:  int,
-	entry_index: int,
-}
-
-_map_add_entry :: proc(m: ^$M/Map($Key, $Value), key: Key) -> int where intrinsics.type_is_valid_map_key(Key) {
-	hasher := intrinsics.type_hasher_proc(Key)
-
-	e: Map_Entry(Key, Value)
-	e.key = key
-	e.hash = hasher(&e.key, 0)
-	e.next = -1
-	idx := array_len(m.entries)
-	array_push(&m.entries, e)
-	return idx
-}
-
-_map_erase :: proc(m: ^$M/Map, fr: Map_Find_Result) {
-	if fr.entry_prev < 0 {
-		array_set(&m.hash, fr.hash_index, array_get(m.entries, fr.entry_index).next)
-	} else {
-		array_get_ptr(m.entries, fr.entry_prev).next = array_get(m.entries, fr.entry_index).next
-	}
-
-	if fr.entry_index == array_len(m.entries)-1 {
-		array_pop_back(&m.entries)
-		return
-	}
-
-	array_set(&m.entries, fr.entry_index, array_get(m.entries, array_len(m.entries)-1))
-	last := _map_find_key(m^, array_get(m.entries, fr.entry_index).key)
-
-	if last.entry_prev < 0 {
-		array_get_ptr(m.entries, last.entry_prev).next = fr.entry_index
-	} else {
-		array_set(&m.hash, last.hash_index, fr.entry_index)
-	}
-}
-
-
-_map_find_key :: proc(m: $M/Map($Key, $Value), key: Key) -> Map_Find_Result where intrinsics.type_is_valid_map_key(Key) {
-	fr: Map_Find_Result
-	fr.hash_index = -1
-	fr.entry_prev = -1
-	fr.entry_index = -1
-
-	if array_len(m.hash) == 0 {
-		return fr
-	}
-
-	hasher := intrinsics.type_hasher_proc(Key)
-
-	key := key
-	hash := hasher(&key, 0)
-
-	fr.hash_index = int(hash % uintptr(array_len(m.hash)))
-	fr.entry_index = array_get(m.hash, fr.hash_index)
-	for fr.entry_index >= 0 {
-		it := array_get_ptr(m.entries, fr.entry_index)
-		if it.hash == hash && it.key == key {
-			return fr
-		}
-		fr.entry_prev = fr.entry_index
-		fr.entry_index = it.next
-	}
-	return fr
-}
-
-_map_find_entry :: proc(m: ^$M/Map($Key, $Value), e: ^Map_Entry(Key, Value)) -> Map_Find_Result {
-	fr: Map_Find_Result
-	fr.hash_index = -1
-	fr.entry_prev = -1
-	fr.entry_index = -1
-
-	if array_len(m.hash) == 0 {
-		return fr
-	}
-
-	fr.hash_index = int(e.hash % uintptr(array_len(m.hash)))
-	fr.entry_index = array_get(m.hash, fr.hash_index)
-	for fr.entry_index >= 0 {
-		it := array_get_ptr(m.entries, fr.entry_index)
-		if it == e {
-			return fr
-		}
-		fr.entry_prev = fr.entry_index
-		fr.entry_index = it.next
-	}
-	return fr
-}
-
-_map_find_or_fail :: proc(m: $M/Map($Key, $Value), key: Key) -> int {
-	return _map_find_key(m, key).entry_index
-}
-_map_find_or_make :: proc(m: ^$M/Map($Key, $Value), key: Key) -> int {
-	fr := _map_find_key(m^, key)
-	if fr.entry_index >= 0 {
-		return fr.entry_index
-	}
-
-	i := _map_add_entry(m, key)
-	if fr.entry_prev < 0 {
-		array_set(&m.hash, fr.hash_index, i)
-	} else {
-		array_get_ptr(m.entries, fr.entry_prev).next = i
-	}
-	return i
-}
-
-
-_map_make :: proc(m: ^$M/Map($Key, $Value), key: Key) -> int {
-	fr := _map_find_key(m^, key)
-	i := _map_add_entry(m, key)
-
-	if fr.entry_prev < 0 {
-		array_set(&m.hash, fr.hash_index, i)
-	} else {
-		array_get_ptr(m.entries, fr.entry_prev).next = i
-	}
-
-	array_get_ptr(m.entries, i).next = fr.entry_index
-
-	return i
-}
-
-
-_map_full :: proc(m: $M/Map($Key, $Value)) -> bool {
-	// TODO(bill): Determine good max load factor
-	return array_len(m.entries) >= (array_len(m.hash) / 4)*3
-}
-
-_map_grow :: proc(m: ^$M/Map($Key, $Value)) {
-	new_size := array_len(m.entries) * 4 + 7 // TODO(bill): Determine good grow rate
-	map_reserve(m, new_size)
-}
-
-
@@ -1,121 +0,0 @@
-package container
-
-Priority_Queue :: struct($T: typeid) {
-	data: Array(T),
-	len: int,
-	priority: proc(item: T) -> int,
-}
-
-priority_queue_init_none :: proc(q: ^$Q/Priority_Queue($T), f: proc(item: T) -> int, allocator := context.allocator) {
-	queue_init_len(q, f, 0, allocator)
-}
-priority_queue_init_len :: proc(q: ^$Q/Priority_Queue($T), f: proc(item: T) -> int, len: int, allocator := context.allocator) {
-	queue_init_len_cap(q, f, 0, 16, allocator)
-}
-priority_queue_init_len_cap :: proc(q: ^$Q/Priority_Queue($T), f: proc(item: T) -> int, len: int, cap: int, allocator := context.allocator) {
-	array_init(&q.data, len, cap, allocator)
-	q.len = len
-	q.priority = f
-}
-
-priority_queue_init :: proc{priority_queue_init_none, priority_queue_init_len, priority_queue_init_len_cap}
-
-
-priority_queue_delete :: proc(q: $Q/Priority_Queue($T)) {
-	array_delete(q.data)
-}
-
-priority_queue_clear :: proc(q: ^$Q/Priority_Queue($T)) {
-	q.len = 0
-}
-
-priority_queue_len :: proc(q: $Q/Priority_Queue($T)) -> int {
-	return q.len
-}
-
-priority_queue_cap :: proc(q: $Q/Priority_Queue($T)) -> int {
-	return array_cap(q.data)
-}
-
-priority_queue_space :: proc(q: $Q/Priority_Queue($T)) -> int {
-	return array_len(q.data) - q.len
-}
-
-priority_queue_reserve :: proc(q: ^$Q/Priority_Queue($T), capacity: int) {
-	if capacity > q.len {
-		array_resize(&q.data, new_capacity)
-	}
-}
-
-priority_queue_resize :: proc(q: ^$Q/Priority_Queue($T), length: int) {
-	if length > q.len {
-		array_resize(&q.data, new_capacity)
-	}
-	q.len = length
-}
-
-_priority_queue_grow :: proc(q: ^$Q/Priority_Queue($T), min_capacity: int = 0) {
-	new_capacity := max(array_len(q.data)*2 + 8, min_capacity)
-	array_resize(&q.data, new_capacity)
-}
-
-
-priority_queue_push :: proc(q: ^$Q/Priority_Queue($T), item: T) {
-	if array_len(q.data) - q.len == 0 {
-		_priority_queue_grow(q)
-	}
-
-	s := array_slice(q.data)
-	s[q.len] = item
-
-	i := q.len
-	for i > 0 {
-		p := (i - 1) / 2
-		if q.priority(s[p]) <= q.priority(item) { 
-			break 
-		}
-		s[i] = s[p]
-		i = p
-	}
-
-	q.len += 1
-	if q.len > 0 { 
-		s[i] = item 
-	} 
-}
-
-
-
-priority_queue_pop :: proc(q: ^$Q/Priority_Queue($T)) -> T {
-	assert(q.len > 0)
-
-	s := array_slice(q.data)
-	min := s[0]
-	root := s[q.len-1]
-	q.len -= 1
-
-	i := 0
-	for i * 2 + 1 < q.len {
-		a := i * 2 + 1
-		b := i * 2 + 2
-		c := b < q.len && q.priority(s[b]) < q.priority(s[a]) ? b : a
-
-		if q.priority(s[c]) >= q.priority(root) {
-			break
-		}
-		s[i] = s[c]
-		i = c
-	}
-
-	if q.len > 0 {
-		s[i] = root
-	}
-	return min
-}
-
-priority_queue_peek :: proc(q: ^$Q/Priority_Queue($T)) -> T {
-	assert(q.len > 0)
-
-	s := array_slice(q.data)
-	return s[0]
-}
@@ -0,0 +1,142 @@
+package container_priority_queue
+
+import "core:builtin"
+
+Priority_Queue :: struct($T: typeid) {
+	queue: [dynamic]T,
+	
+	less:  proc(a, b: T) -> bool,
+	swap:  proc(q: []T, i, j: int),
+}
+
+DEFAULT_CAPACITY :: 16
+
+default_swap_proc :: proc($T: typeid) -> proc(q: []T, i, j: int) {
+	return proc(q: []T, i, j: int) {
+		q[i], q[j] = q[j], q[i]
+	}
+}
+
+init :: proc(pq: ^$Q/Priority_Queue($T), less: proc(a, b: T) -> bool, swap: proc(q: []T, i, j: int), capacity := DEFAULT_CAPACITY, allocator := context.allocator) {
+	if pq.queue.allocator.procedure == nil {
+		pq.queue.allocator = allocator
+	}
+	reserve(pq, capacity)
+	pq.less = less
+	pq.swap = swap
+}
+
+init_from_dynamic_array :: proc(pq: ^$Q/Priority_Queue($T), queue: [dynamic]T, less: proc(a, b: T) -> bool, swap: proc(q: []T, i, j: int)) {
+	pq.queue = queue
+	pq.less = less
+	pq.swap = swap
+	n := builtin.len(pq.queue)
+	for i := n/2 - 1; i >= 0; i -= 1 {
+		_shift_down(pq, i, n)
+	}
+}
+
+destroy :: proc(pq: ^$Q/Priority_Queue($T)) {
+	clear(pq)
+	delete(pq.queue)
+}
+
+reserve :: proc(pq: ^$Q/Priority_Queue($T), capacity: int) {
+	builtin.reserve(&pq.queue, capacity)
+}
+clear :: proc(pq: ^$Q/Priority_Queue($T)) {
+	builtin.clear(&pq.queue)
+}
+len :: proc(pq: $Q/Priority_Queue($T)) -> int {
+	return builtin.len(pq.queue)
+}
+cap :: proc(pq: $Q/Priority_Queue($T)) -> int {
+	return builtin.cap(pq.queue)
+}
+
+_shift_down :: proc(pq: ^$Q/Priority_Queue($T), i0, n: int) -> bool {
+	// O(n log n)
+	if 0 > i0 || i0 > n {
+		return false
+	}
+	
+	i := i0
+	queue := pq.queue[:]
+	
+	for {
+		j1 := 2*i + 1
+		if j1 < 0 || j1 >= n {
+			break
+		}
+		j := j1
+		if j2 := j1+1; j2 < n && pq.less(queue[j2], queue[j1]) {
+			j = j2
+		}
+		if !pq.less(queue[j], queue[i]) {
+			break
+		}
+		
+		pq.swap(queue, i, j)
+		i = j
+	}
+	return i > i0
+}
+
+_shift_up :: proc(pq: ^$Q/Priority_Queue($T), j: int) {
+	j := j
+	queue := pq.queue[:]
+	for 0 <= j {
+		i := (j-1)/2
+		if i == j || !pq.less(queue[j], queue[i]) {
+			break
+		}
+		pq.swap(queue, i, j)
+		j = i
+	}
+}
+
+// NOTE(bill): When an element at index 'i' has changed its value, this will fix the
+// the heap ordering. This is using a basic "heapsort" with shift up and a shift down parts.
+fix :: proc(pq: ^$Q/Priority_Queue($T), i: int) {
+	if !_shift_down(pq, i, builtin.len(pq.queue)) {
+		_shift_up(pq, i)
+	}
+}
+
+push :: proc(pq: ^$Q/Priority_Queue($T), value: T) {
+	append(&pq.queue, value)
+	_shift_up(pq, builtin.len(pq.queue)-1)
+}
+
+pop :: proc(pq: ^$Q/Priority_Queue($T), loc := #caller_location) -> (value: T) {
+	assert(condition=builtin.len(pq.queue)>0, loc=loc)
+	
+	n := builtin.len(pq.queue)-1
+	pq.swap(pq.queue[:], 0, n)
+	_shift_down(pq, 0, n)
+	return builtin.pop(&pq.queue)
+}
+
+pop_safe :: proc(pq: ^$Q/Priority_Queue($T), loc := #caller_location) -> (value: T, ok: bool) {
+	if builtin.len(pq.queue) > 0 {
+		n := builtin.len(pq.queue)-1
+		pq.swap(pq.queue[:], 0, n)
+		_shift_down(pq, 0, n)
+		return builtin.pop_safe(&pq.queue)
+	}
+	return
+}
+
+remove :: proc(pq: ^$Q/Priority_Queue($T), i: int) -> (value: T, ok: bool) {
+	n := builtin.len(pq.queue)
+	if 0 <= i && i < n {
+		if n != i {
+			pq.swap(pq.queue[:], i, n)
+			_shift_down(pq, i, n)
+			_shift_up(pq, i)
+		}
+		value, ok = builtin.pop_safe(&pq.queue)
+	}
+	return
+}
+
@@ -1,175 +0,0 @@
-package container
-
-Queue :: struct($T: typeid) {
-	data: Array(T),
-	len: int,
-	offset: int,
-}
-
-/*
-queue_init :: proc{
-	queue_init_none,
-	queue_init_len,
-	queue_init_len_cap,
-}
-queue_delete
-queue_clear
-queue_len
-queue_cap
-queue_space
-queue_get
-queue_set
-queue_reserve
-queue_resize
-queue_push :: proc{
-	queue_push_back, 
-	queue_push_elems,
-};
-queue_push_front
-queue_pop_front
-queue_pop_back
-queue_consume
-*/
-
-queue_init_none :: proc(q: ^$Q/Queue($T), allocator := context.allocator) {
-	queue_init_len(q, 0, allocator)
-}
-queue_init_len :: proc(q: ^$Q/Queue($T), len: int, allocator := context.allocator) {
-	queue_init_len_cap(q, 0, 16, allocator)
-}
-queue_init_len_cap :: proc(q: ^$Q/Queue($T), len: int, cap: int, allocator := context.allocator) {
-	array_init(&q.data, len, cap, allocator)
-	q.len = len
-	q.offset = 0
-}
-
-queue_init :: proc{queue_init_none, queue_init_len, queue_init_len_cap}
-
-queue_delete :: proc(q: $Q/Queue($T)) {
-	array_delete(q.data)
-}
-
-queue_clear :: proc(q: ^$Q/Queue($T)) {
-	q.len = 0
-}
-
-queue_len :: proc(q: $Q/Queue($T)) -> int {
-	return q.len
-}
-
-queue_cap :: proc(q: $Q/Queue($T)) -> int {
-	return array_cap(q.data)
-}
-
-queue_space :: proc(q: $Q/Queue($T)) -> int {
-	return array_len(q.data) - q.len
-}
-
-queue_get :: proc(q: $Q/Queue($T), index: int) -> T {
-	i := (index + q.offset) % array_len(q.data)
-	data := array_slice(q.data)
-	return data[i]
-}
-
-queue_set :: proc(q: ^$Q/Queue($T), index: int, item: T)  {
-	i := (index + q.offset) % array_len(q.data)
-	data := array_slice(q.data)
-	data[i] = item
-}
-
-
-queue_reserve :: proc(q: ^$Q/Queue($T), capacity: int) {
-	if capacity > q.len {
-		_queue_increase_capacity(q, capacity)
-	}
-}
-
-queue_resize :: proc(q: ^$Q/Queue($T), length: int) {
-	if length > q.len {
-		_queue_increase_capacity(q, length)
-	}
-	q.len = length
-}
-
-queue_push_back :: proc(q: ^$Q/Queue($T), item: T) {
-	if queue_space(q^) == 0 {
-		_queue_grow(q)
-	}
-
-	queue_set(q, q.len, item)
-	q.len += 1
-}
-
-queue_push_front :: proc(q: ^$Q/Queue($T), item: T) {
-	if queue_space(q^) == 0 {
-		_queue_grow(q)
-	}
-
-	q.offset = (q.offset - 1 + array_len(q.data)) % array_len(q.data)
-	q.len += 1
-	queue_set(q, 0, item)
-}
-
-queue_pop_front :: proc(q: ^$Q/Queue($T)) -> T {
-	assert(q.len > 0)
-	item := queue_get(q^, 0)
-	q.offset = (q.offset + 1) % array_len(q.data)
-	q.len -= 1
-	if q.len == 0 {
-		q.offset = 0
-	}
-	return item
-}
-
-queue_pop_back :: proc(q: ^$Q/Queue($T)) -> T {
-	assert(q.len > 0)
-	item := queue_get(q^, q.len-1)
-	q.len -= 1
-	return item
-}
-
-queue_consume :: proc(q: ^$Q/Queue($T), count: int) {
-	q.offset = (q.offset + count) & array_len(q.data)
-	q.len -= count
-}
-
-
-queue_push_elems :: proc(q: ^$Q/Queue($T), items: ..T) {
-	if queue_space(q^) < len(items) {
-		_queue_grow(q, q.len + len(items))
-	}
-	size := array_len(q.data)
-	insert := (q.offset + q.len) % size
-
-	to_insert := len(items)
-	if insert + to_insert > size {
-		to_insert = size - insert
-	}
-
-	the_items := items[:]
-
-	data := array_slice(q.data)
-
-	q.len += copy(data[insert:][:to_insert], the_items)
-	the_items = the_items[to_insert:]
-	q.len += copy(data[:], the_items)
-}
-
-queue_push :: proc{queue_push_back, queue_push_elems}
-
-
-
-_queue_increase_capacity :: proc(q: ^$Q/Queue($T), new_capacity: int) {
-	end := array_len(q.data)
-	array_resize(&q.data, new_capacity)
-	if q.offset + q.len > end {
-		end_items := q.len + end
-		data := array_slice(q.data)
-		copy(data[new_capacity-end_items:][:end_items], data[q.offset:][:end_items])
-		q.offset += new_capacity - end
-	}
-}
-_queue_grow :: proc(q: ^$Q/Queue($T), min_capacity: int = 0) {
-	new_capacity := max(array_len(q.data)*2 + 8, min_capacity)
-	_queue_increase_capacity(q, new_capacity)
-}
@@ -0,0 +1,238 @@
+package container_queue
+
+import "core:builtin"
+import "core:runtime"
+_ :: runtime
+
+// Dynamically resizable double-ended queue/ring-buffer
+Queue :: struct($T: typeid) {
+	data:   [dynamic]T,
+	len:    uint,
+	offset: uint,
+}
+
+DEFAULT_CAPACITY :: 16
+
+// Procedure to initialize a queue
+init :: proc(q: ^$Q/Queue($T), capacity := DEFAULT_CAPACITY, allocator := context.allocator) -> bool {
+	if q.data.allocator.procedure == nil {
+		q.data.allocator = allocator
+	}
+	clear(q)
+	return reserve(q, capacity)
+}
+
+// Procedure to initialize a queue from a fixed backing slice
+init_from_slice :: proc(q: ^$Q/Queue($T), backing: []T) -> bool {
+	clear(q)
+	q.data = transmute([dynamic]T)runtime.Raw_Dynamic_Array{
+		data = raw_data(backing),
+		len = builtin.len(backing),
+		cap = builtin.len(backing),
+		allocator = {procedure=runtime.nil_allocator_proc, data=nil},
+	}
+	return true
+}
+
+// Procedure to destroy a queue
+destroy :: proc(q: ^$Q/Queue($T)) {
+	delete(q.data)
+}
+
+// The length of the queue
+len :: proc(q: $Q/Queue($T)) -> int {
+	return int(q.len)
+}
+
+// The current capacity of the queue
+cap :: proc(q: $Q/Queue($T)) -> int {
+	return builtin.len(q.data)
+}
+
+// Remaining space in the queue (cap-len)
+space :: proc(q: $Q/Queue($T)) -> int {
+	return builtin.len(q.data) - int(q.len)
+}
+
+// Reserve enough space for at least the specified capacity
+reserve :: proc(q: ^$Q/Queue($T), capacity: int) -> bool {
+	if uint(capacity) > q.len {
+		return _grow(q, uint(capacity)) 
+	}
+	return true
+}
+
+
+get :: proc(q: ^$Q/Queue($T), #any_int i: int, loc := #caller_location) -> T {
+	runtime.bounds_check_error_loc(loc, i, builtin.len(q.data))
+
+	idx := (uint(i)+q.offset)%builtin.len(q.data)
+	return q.data[idx]
+}
+
+front :: proc(q: ^$Q/Queue($T)) -> T {
+	return q.data[q.offset]
+}
+front_ptr :: proc(q: ^$Q/Queue($T)) -> ^T {
+	return &q.data[q.offset]
+}
+
+back :: proc(q: ^$Q/Queue($T)) -> T {
+	idx := (q.offset+uint(q.len))%builtin.len(q.data)
+	return q.data[idx]
+}
+back_ptr :: proc(q: ^$Q/Queue($T)) -> ^T {
+	idx := (q.offset+uint(q.len))%builtin.len(q.data)
+	return &q.data[idx]
+}
+
+set :: proc(q: ^$Q/Queue($T), #any_int i: int, val: T, loc := #caller_location) {
+	runtime.bounds_check_error_loc(loc, i, builtin.len(q.data))
+	
+	idx := (uint(i)+q.offset)%builtin.len(q.data)
+	q.data[idx] = val
+}
+get_ptr :: proc(q: ^$Q/Queue($T), #any_int i: int, loc := #caller_location) -> ^T {
+	runtime.bounds_check_error_loc(loc, i, builtin.len(q.data))
+	
+	idx := (uint(i)+q.offset)%builtin.len(q.data)
+	return &q.data[idx]
+}
+
+peek_front :: proc(q: ^$Q/Queue($T), loc := #caller_location) -> ^T {
+	runtime.bounds_check_error_loc(loc, 0, builtin.len(q.data))
+	idx := q.offset%builtin.len(q.data)
+	return &q.data[idx]
+}
+
+peek_back :: proc(q: ^$Q/Queue($T), loc := #caller_location) -> ^T {
+	runtime.bounds_check_error_loc(loc, int(q.len - 1), builtin.len(q.data))
+	idx := (uint(q.len - 1)+q.offset)%builtin.len(q.data)
+	return &q.data[idx]
+}
+
+// Push an element to the back of the queue
+push_back :: proc(q: ^$Q/Queue($T), elem: T) -> bool {
+	if space(q^) == 0 {
+		_grow(q) or_return
+	}
+	idx := (q.offset+uint(q.len))%builtin.len(q.data)
+	q.data[idx] = elem
+	q.len += 1
+	return true
+}
+
+// Push an element to the front of the queue
+push_front :: proc(q: ^$Q/Queue($T), elem: T) -> bool {
+	if space(q^) == 0 {
+		_grow(q) or_return
+	}	
+	q.offset = uint(q.offset - 1 + builtin.len(q.data)) % builtin.len(q.data)
+	q.len += 1
+	q.data[q.offset] = elem
+	return true
+}
+
+
+// Pop an element from the back of the queue
+pop_back :: proc(q: ^$Q/Queue($T), loc := #caller_location) -> (elem: T) {
+	assert(condition=q.len > 0, loc=loc)
+	q.len -= 1
+	idx := (q.offset+uint(q.len))%builtin.len(q.data)
+	elem = q.data[idx]
+	return
+}
+// Safely pop an element from the back of the queue
+pop_back_safe :: proc(q: ^$Q/Queue($T)) -> (elem: T, ok: bool) {
+	if q.len > 0 {
+		q.len -= 1
+		idx := (q.offset+uint(q.len))%builtin.len(q.data)
+		elem = q.data[idx]
+		ok = true
+	}
+	return
+}
+
+// Pop an element from the front of the queue
+pop_front :: proc(q: ^$Q/Queue($T), loc := #caller_location) -> (elem: T) {
+	assert(condition=q.len > 0, loc=loc)
+	elem = q.data[q.offset]
+	q.offset = (q.offset+1)%builtin.len(q.data)
+	q.len -= 1
+	return
+}
+// Safely pop an element from the front of the queue
+pop_front_safe :: proc(q: ^$Q/Queue($T)) -> (elem: T, ok: bool) {
+	if q.len > 0 {
+		elem = q.data[q.offset]
+		q.offset = (q.offset+1)%builtin.len(q.data)
+		q.len -= 1
+		ok = true
+	}
+	return
+}
+
+// Push multiple elements to the front of the queue
+push_back_elems :: proc(q: ^$Q/Queue($T), elems: ..T) -> bool {
+	n := uint(builtin.len(elems))
+	if space(q^) < int(n) {
+		_grow(q, q.len + n) or_return
+	}
+	
+	sz := uint(builtin.len(q.data))
+	insert_from := (q.offset + q.len) % sz
+	insert_to := n
+	if insert_from + insert_to > sz {
+		insert_to = sz - insert_from
+	}
+	copy(q.data[insert_from:], elems[:insert_to])
+	copy(q.data[:insert_from], elems[insert_to:])
+	q.len += n
+	return true
+}
+
+// Consume `n` elements from the front of the queue
+consume_front :: proc(q: ^$Q/Queue($T), n: int, loc := #caller_location) {
+	assert(condition=int(q.len) >= n, loc=loc)
+	if n > 0 {
+		nu := uint(n)
+		q.offset = (q.offset + nu) % builtin.len(q.data)
+		q.len -= nu	
+	}
+}
+
+// Consume `n` elements from the back of the queue
+consume_back :: proc(q: ^$Q/Queue($T), n: int, loc := #caller_location) {
+	assert(condition=int(q.len) >= n, loc=loc)
+	if n > 0 {
+		q.len -= uint(n)
+	}
+}
+
+
+
+append_elem  :: push_back
+append_elems :: push_back_elems
+push   :: proc{push_back, push_back_elems}
+append :: proc{push_back, push_back_elems}
+
+
+// Clear the contents of the queue
+clear :: proc(q: ^$Q/Queue($T)) {
+	q.len = 0
+	q.offset = 0
+}
+
+
+// Internal growinh procedure
+_grow :: proc(q: ^$Q/Queue($T), min_capacity: uint = 0) -> bool {
+	new_capacity := max(min_capacity, uint(8), uint(builtin.len(q.data))*2)
+	n := uint(builtin.len(q.data))
+	builtin.resize(&q.data, int(new_capacity)) or_return
+	if q.offset + q.len > n {
+		diff := n - q.offset
+		copy(q.data[new_capacity-diff:], q.data[q.offset:][:diff])
+		q.offset += new_capacity - n
+	}
+	return true
+}
@@ -1,74 +0,0 @@
-package container
-
-
-Ring :: struct($T: typeid) {
-	next, prev: ^Ring(T),
-	value: T,
-}
-
-ring_init :: proc(r: ^$R/Ring) -> ^R {
-	r.prev, r.next = r, r
-	return r
-}
-
-ring_next :: proc(r: ^$R/Ring) -> ^R {
-	if r.next == nil {
-		return ring_init(r)
-	}
-	return r.next
-}
-ring_prev :: proc(r: ^$R/Ring) -> ^R {
-	if r.prev == nil {
-		return ring_init(r)
-	}
-	return r.prev
-}
-
-
-ring_move :: proc(r: ^$R/Ring, n: int) -> ^R {
-  r := r
-	if r.next == nil {
-		return ring_init(r)
-	}
-
-	switch {
-	case n < 0:
-		for _ in n..<0 {
-			r = r.prev
-		}
-	case n > 0:
-		for _ in 0..<n {
-			r = r.next
-		}
-	}
-	return r
-}
-
-ring_link :: proc(r, s: ^$R/Ring) -> ^R {
-	n := ring_next(r)
-	if s != nil {
-		p := ring_prev(s)
-		r.next = s
-		s.prev = r
-		n.prev = p
-		p.next = n
-	}
-	return n
-}
-ring_unlink :: proc(r: ^$R/Ring, n: int) -> ^R {
-	if n <= 0 {
-		return nil
-	}
-	return ring_link(r, ring_move(r, n+1))
-}
-ring_len :: proc(r: ^$R/Ring) -> int {
-	n := 0
-	if r != nil {
-		n = 1
-		for p := ring_next(r); p != r; p = p.next {
-			n += 1
-		}
-	}
-	return n
-}
-
@@ -1,240 +0,0 @@
-package container
-
-Set :: struct {
-	hash:    Array(int),
-	entries: Array(Set_Entry),
-}
-
-Set_Entry :: struct {
-	key:   u64,
-	next:  int,
-}
-
-
-/*
-set_init :: proc{
-	set_init_none,
-	set_init_cap,
-}
-set_delete
-
-set_in
-set_not_in
-set_add
-set_remove
-set_reserve
-set_clear
-*/
-
-set_init :: proc{set_init_none, set_init_cap}
-
-set_init_none :: proc(m: ^Set, allocator := context.allocator) {
-	m.hash.allocator = allocator
-	m.entries.allocator = allocator
-}
-
-set_init_cap :: proc(m: ^Set, cap: int, allocator := context.allocator) {
-	m.hash.allocator = allocator
-	m.entries.allocator = allocator
-	set_reserve(m, cap)
-}
-
-set_delete :: proc(m: Set) {
-	array_delete(m.hash)
-	array_delete(m.entries)
-}
-
-
-set_in :: proc(m: Set, key: u64) -> bool {
-	return _set_find_or_fail(m, key) >= 0
-}
-set_not_in :: proc(m: Set, key: u64) -> bool {
-	return _set_find_or_fail(m, key) < 0
-}
-
-set_add :: proc(m: ^Set, key: u64) {
-	if array_len(m.hash) == 0 {
-		_set_grow(m)
-	}
-
-	_ = _set_find_or_make(m, key)
-	if _set_full(m^) {
-		_set_grow(m)
-	}
-}
-
-set_remove :: proc(m: ^Set, key: u64) {
-	fr := _set_find_key(m^, key)
-	if fr.entry_index >= 0 {
-		_set_erase(m, fr)
-	}
-}
-
-
-set_reserve :: proc(m: ^Set, new_size: int) {
-	nm: Set
-	set_init(&nm, m.hash.allocator)
-	array_resize(&nm.hash, new_size)
-	array_reserve(&nm.entries, array_len(m.entries))
-
-	for i in 0..<new_size {
-		array_set(&nm.hash, i, -1)
-	}
-	for i in 0..<array_len(m.entries) {
-		e := array_get(m.entries, i)
-		set_add(&nm, e.key)
-	}
-
-	set_delete(m^)
-	m^ = nm
-}
-
-set_clear :: proc(m: ^Set) {
-	array_clear(&m.hash)
-	array_clear(&m.entries)
-}
-
-
-set_equal :: proc(a, b: Set) -> bool {
-	a_entries := array_slice(a.entries)
-	b_entries := array_slice(b.entries)
-	if len(a_entries) != len(b_entries) {
-		return false
-	}
-	for e in a_entries {
-		if set_not_in(b, e.key) {
-			return false
-		}
-	}
-
-	return true
-}
-
-
-
-/// Internal
-
-_set_add_entry :: proc(m: ^Set, key: u64) -> int {
-	e: Set_Entry
-	e.key = key
-	e.next = -1
-	idx := array_len(m.entries)
-	array_push(&m.entries, e)
-	return idx
-}
-
-_set_erase :: proc(m: ^Set, fr: Map_Find_Result) {
-	if fr.entry_prev < 0 {
-		array_set(&m.hash, fr.hash_index, array_get(m.entries, fr.entry_index).next)
-	} else {
-		array_get_ptr(m.entries, fr.entry_prev).next = array_get(m.entries, fr.entry_index).next
-	}
-
-	if fr.entry_index == array_len(m.entries)-1 {
-		array_pop_back(&m.entries)
-		return
-	}
-
-	array_set(&m.entries, fr.entry_index, array_get(m.entries, array_len(m.entries)-1))
-	last := _set_find_key(m^, array_get(m.entries, fr.entry_index).key)
-
-	if last.entry_prev < 0 {
-		array_get_ptr(m.entries, last.entry_prev).next = fr.entry_index
-	} else {
-		array_set(&m.hash, last.hash_index, fr.entry_index)
-	}
-}
-
-
-_set_find_key :: proc(m: Set, key: u64) -> Map_Find_Result {
-	fr: Map_Find_Result
-	fr.hash_index = -1
-	fr.entry_prev = -1
-	fr.entry_index = -1
-
-	if array_len(m.hash) == 0 {
-		return fr
-	}
-
-	fr.hash_index = int(key % u64(array_len(m.hash)))
-	fr.entry_index = array_get(m.hash, fr.hash_index)
-	for fr.entry_index >= 0 {
-		it := array_get_ptr(m.entries, fr.entry_index)
-		if it.key == key {
-			return fr
-		}
-		fr.entry_prev = fr.entry_index
-		fr.entry_index = it.next
-	}
-	return fr
-}
-
-_set_find_entry :: proc(m: ^Set, e: ^Set_Entry) -> Map_Find_Result {
-	fr: Map_Find_Result
-	fr.hash_index = -1
-	fr.entry_prev = -1
-	fr.entry_index = -1
-
-	if array_len(m.hash) == 0 {
-		return fr
-	}
-
-	fr.hash_index = int(e.key % u64(array_len(m.hash)))
-	fr.entry_index = array_get(m.hash, fr.hash_index)
-	for fr.entry_index >= 0 {
-		it := array_get_ptr(m.entries, fr.entry_index)
-		if it == e {
-			return fr
-		}
-		fr.entry_prev = fr.entry_index
-		fr.entry_index = it.next
-	}
-	return fr
-}
-
-_set_find_or_fail :: proc(m: Set, key: u64) -> int {
-	return _set_find_key(m, key).entry_index
-}
-_set_find_or_make :: proc(m: ^Set, key: u64) -> int {
-	fr := _set_find_key(m^, key)
-	if fr.entry_index >= 0 {
-		return fr.entry_index
-	}
-
-	i := _set_add_entry(m, key)
-	if fr.entry_prev < 0 {
-		array_set(&m.hash, fr.hash_index, i)
-	} else {
-		array_get_ptr(m.entries, fr.entry_prev).next = i
-	}
-	return i
-}
-
-
-_set_make :: proc(m: ^Set, key: u64) -> int {
-	fr := _set_find_key(m^, key)
-	i := _set_add_entry(m, key)
-
-	if fr.entry_prev < 0 {
-		array_set(&m.hash, fr.hash_index, i)
-	} else {
-		array_get_ptr(m.entries, fr.entry_prev).next = i
-	}
-
-	array_get_ptr(m.entries, i).next = fr.entry_index
-
-	return i
-}
-
-
-_set_full :: proc(m: Set) -> bool {
-	// TODO(bill): Determine good max load factor
-	return array_len(m.entries) >= (array_len(m.hash) / 4)*3
-}
-
-_set_grow :: proc(m: ^Set) {
-	new_size := array_len(m.entries) * 4 + 7 // TODO(bill): Determine good grow rate
-	set_reserve(m, new_size)
-}
-
-
@@ -1,95 +0,0 @@
-package container
-
-Small_Array :: struct($N: int, $T: typeid) where N >= 0 {
-	data: [N]T,
-	len:  int,
-}
-
-
-small_array_len :: proc(a: $A/Small_Array) -> int {
-	return a.len
-}
-
-small_array_cap :: proc(a: $A/Small_Array) -> int {
-	return len(a.data)
-}
-
-small_array_space :: proc(a: $A/Small_Array) -> int {
-	return len(a.data) - a.len
-}
-
-small_array_slice :: proc(a: ^$A/Small_Array($N, $T)) -> []T {
-	return a.data[:a.len]
-}
-
-
-small_array_get :: proc(a: $A/Small_Array($N, $T), index: int, loc := #caller_location) -> T {
-	return a.data[index]
-}
-small_array_get_ptr :: proc(a: $A/Small_Array($N, $T), index: int, loc := #caller_location) -> ^T {
-	return &a.data[index]
-}
-
-small_array_set :: proc(a: ^$A/Small_Array($N, $T), index: int, item: T, loc := #caller_location) {
-	a.data[index] = item
-}
-
-small_array_resize :: proc(a: ^$A/Small_Array, length: int) {
-	a.len = min(length, len(a.data))
-}
-
-
-small_array_push_back :: proc(a: ^$A/Small_Array($N, $T), item: T) -> bool {
-	if a.len < len(a.data) {
-		a.len += 1
-		a.data[a.len-1] = item
-		return true
-	}
-	return false
-}
-
-small_array_push_front :: proc(a: ^$A/Small_Array($N, $T), item: T) -> bool {
-	if a.len < len(a.data) {
-		a.len += 1
-		data := small_array_slice(a)
-		copy(data[1:], data[:])
-		data[0] = item
-		return true
-	}
-	return false
-}
-
-small_array_pop_back :: proc(a: ^$A/Small_Array($N, $T), loc := #caller_location) -> T {
-	assert(condition=a.len > 0, loc=loc)
-	item := a.data[a.len-1]
-	a.len -= 1
-	return item
-}
-
-small_array_pop_front :: proc(a: ^$A/Small_Array($N, $T), loc := #caller_location) -> T {
-	assert(condition=a.len > 0, loc=loc)
-	item := a.data[0]
-	s := small_array_slice(a)
-	copy(s[:], s[1:])
-	a.len -= 1
-	return item
-}
-
-
-small_array_consume :: proc(a: ^$A/Small_Array($N, $T), count: int, loc := #caller_location) {
-	assert(condition=a.len >= count, loc=loc)
-	a.len -= count
-}
-
-small_array_clear :: proc(a: ^$A/Small_Array($N, $T)) {
-	small_array_resize(a, 0)
-}
-
-small_array_push_back_elems :: proc(a: ^$A/Small_Array($N, $T), items: ..T) {
-	n := copy(a.data[a.len:], items[:])
-	a.len += n
-}
-
-small_array_push   :: proc{small_array_push_back, small_array_push_back_elems}
-small_array_append :: proc{small_array_push_back, small_array_push_back_elems}
-
@@ -0,0 +1,117 @@
+package container_small_array
+
+import "core:builtin"
+
+Small_Array :: struct($N: int, $T: typeid) where N >= 0 {
+	data: [N]T,
+	len:  int,
+}
+
+
+len :: proc(a: $A/Small_Array) -> int {
+	return a.len
+}
+
+cap :: proc(a: $A/Small_Array) -> int {
+	return builtin.len(a.data)
+}
+
+space :: proc(a: $A/Small_Array) -> int {
+	return builtin.len(a.data) - a.len
+}
+
+slice :: proc(a: ^$A/Small_Array($N, $T)) -> []T {
+	return a.data[:a.len]
+}
+
+
+get :: proc(a: $A/Small_Array($N, $T), index: int) -> T {
+	return a.data[index]
+}
+get_ptr :: proc(a: ^$A/Small_Array($N, $T), index: int) -> ^T {
+	return &a.data[index]
+}
+
+set :: proc(a: ^$A/Small_Array($N, $T), index: int, item: T) {
+	a.data[index] = item
+}
+
+resize :: proc(a: ^$A/Small_Array, length: int) {
+	a.len = min(length, builtin.len(a.data))
+}
+
+
+push_back :: proc(a: ^$A/Small_Array($N, $T), item: T) -> bool {
+	if a.len < cap(a^) {
+		a.data[a.len] = item
+		a.len += 1
+		return true
+	}
+	return false
+}
+
+push_front :: proc(a: ^$A/Small_Array($N, $T), item: T) -> bool {
+	if a.len < cap(a^) {
+		a.len += 1
+		data := slice(a)
+		copy(data[1:], data[:])
+		data[0] = item
+		return true
+	}
+	return false
+}
+
+pop_back :: proc(a: ^$A/Small_Array($N, $T), loc := #caller_location) -> T {
+	assert(condition=(N > 0 && a.len > 0), loc=loc)
+	item := a.data[a.len-1]
+	a.len -= 1
+	return item
+}
+
+pop_front :: proc(a: ^$A/Small_Array($N, $T), loc := #caller_location) -> T {
+	assert(condition=(N > 0 && a.len > 0), loc=loc)
+	item := a.data[0]
+	s := slice(a)
+	copy(s[:], s[1:])
+	a.len -= 1
+	return item
+}
+
+pop_back_safe :: proc(a: ^$A/Small_Array($N, $T)) -> (item: T, ok: bool) {
+	if N > 0 && a.len > 0 {
+		item = a.data[a.len-1]
+		a.len -= 1
+		ok = true
+	}
+	return
+}
+
+pop_front_safe :: proc(a: ^$A/Small_Array($N, $T)) -> (item: T, ok: bool) {
+	if N > 0 && a.len > 0 {
+		item = a.data[0]
+		s := slice(a)
+		copy(s[:], s[1:])
+		a.len -= 1
+		ok = true
+	} 
+	return
+}
+
+consume :: proc(a: ^$A/Small_Array($N, $T), count: int, loc := #caller_location) {
+	assert(condition=a.len >= count, loc=loc)
+	a.len -= count
+}
+
+clear :: proc(a: ^$A/Small_Array($N, $T)) {
+	resize(a, 0)
+}
+
+push_back_elems :: proc(a: ^$A/Small_Array($N, $T), items: ..T) {
+	n := copy(a.data[a.len:], items[:])
+	a.len += n
+}
+
+append_elem  :: push_back
+append_elems :: push_back_elems
+push   :: proc{push_back, push_back_elems}
+append :: proc{push_back, push_back_elems}
@@ -0,0 +1,98 @@
+// The following is a generic O(V+E) topological sorter implementation.
+// This is the fastest known method for topological sorting and Odin's
+// map type is being used to accelerate lookups.
+package container_topological_sort
+
+import "core:intrinsics"
+import "core:runtime"
+_ :: intrinsics
+_ :: runtime
+
+
+Relations :: struct($K: typeid) where intrinsics.type_is_valid_map_key(K) {
+	dependents:   map[K]bool,
+	dependencies: int,
+}
+
+Sorter :: struct(K: typeid) where intrinsics.type_is_valid_map_key(K)  {
+	relations: map[K]Relations(K),
+	dependents_allocator: runtime.Allocator,
+}
+
+@(private="file")
+make_relations :: proc(sorter: ^$S/Sorter($K)) -> (r: Relations(K)) {
+	r.dependents.allocator = sorter.dependents_allocator
+	return
+}
+
+
+init :: proc(sorter: ^$S/Sorter($K)) {
+	sorter.relations = make(map[K]Relations(K))
+	sorter.dependents_allocator = context.allocator
+}
+
+destroy :: proc(sorter: ^$S/Sorter($K)) {
+	for _, v in &sorter.relations {
+		delete(v.dependents)
+	}
+	delete(sorter.relations)
+}
+
+add_key :: proc(sorter: ^$S/Sorter($K), key: K) -> bool {
+	if key in sorter.relations {
+		return false
+	}
+	sorter.relations[key] = make_relations(sorter)
+	return true
+}
+
+add_dependency :: proc(sorter: ^$S/Sorter($K), key, dependency: K) -> bool {
+	if key == dependency {
+		return false
+	}
+
+	find := &sorter.relations[dependency]
+	if find == nil {
+		find = map_insert(&sorter.relations, dependency, make_relations(sorter))
+	}
+
+	if find.dependents[key] {
+		return true
+	}
+	find.dependents[key] = true
+
+ 	find = &sorter.relations[key]
+	if find == nil {
+		find = map_insert(&sorter.relations, key, make_relations(sorter))
+	}
+
+	find.dependencies += 1
+
+	return true
+}
+
+sort :: proc(sorter: ^$S/Sorter($K)) -> (sorted, cycled: [dynamic]K) {
+	relations := &sorter.relations
+
+	for k, v in relations {
+		if v.dependencies == 0 {
+			append(&sorted, k)
+		}
+	}
+
+	for root in &sorted do for k, _ in relations[root].dependents {
+		relation := &relations[k]
+		relation.dependencies -= 1
+		if relation.dependencies == 0 {
+			append(&sorted, k)
+		}
+	}
+
+	for k, v in relations {
+		if v.dependencies != 0 {
+			append(&cycled, k)
+		}
+	}
+
+	return
+}
@@ -32,9 +32,11 @@ Please see the chart below for the options.

 #### High level API
 Each hash algorithm contains a procedure group named `hash`, or if the algorithm provides more than one digest size `hash_<size>`\*.  
-Included in these groups are four procedures.
+Included in these groups are six procedures.
 * `hash_string` - Hash a given string and return the computed hash. Just calls `hash_bytes` internally
 * `hash_bytes` - Hash a given byte slice and return the computed hash
+* `hash_string_to_buffer` - Hash a given string and put the computed hash in the second proc parameter. Just calls `hash_bytes_to_buffer` internally
+* `hash_bytes_to_buffer` - Hash a given string and put the computed hash in the second proc parameter. The destination buffer has to be at least as big as the digest size of the hash
 * `hash_stream` - Takes a stream from io.Stream and returns the computed hash from it
 * `hash_file` - Takes a file handle and returns the computed hash from it. A second optional boolean parameter controls if the file is streamed (this is the default) or read at once (set to true)

@@ -59,6 +61,10 @@ main :: proc() {
    // Compute the hash, using the high level API
    computed_hash := md4.hash(input)

+    // Variant that takes a destination buffer, instead of returning the computed hash
+    hash := make([]byte, md4.DIGEST_SIZE) // @note: Destination buffer has to be at least as big as the digest size of the hash
+    md4.hash(input, hash[:])
+
    // Compute the hash, using the low level API
    ctx: md4.Md4_Context
    computed_hash_low: [16]byte
@@ -75,7 +81,7 @@ The crypto package is not thread-safe at the moment. This may change in the futu
 ### Disclaimer
 The algorithms were ported out of curiosity and due to interest in the field.
 We have not had any of the code verified by a third party or tested/fuzzed by any automatic means.
-Whereever we were able to find official test vectors, those were used to verify the implementation.
+Wherever we were able to find official test vectors, those were used to verify the implementation.
 We do not recommend using them in a production environment, without any additional testing and/or verification.

 ### ToDo
@@ -30,6 +30,6 @@ equivalence.

 For the most part, alterations to the base fiat-crypto generated code was
 kept to a minimum, to aid auditability.  This results in a somewhat
-ideosyncratic style, and in some cases minor performance penalties.
+idiosyncratic style, and in some cases minor performance penalties.

 [1]: https://github.com/mit-plv/fiat-crypto
@@ -22,7 +22,7 @@ fe_from_bytes :: #force_inline proc (out1: ^Tight_Field_Element, arg1: []byte, a

 	assert(len(arg1) == 16)

-	when ODIN_ARCH == "386" || ODIN_ARCH == "amd64" {
+	when ODIN_ARCH == .i386 || ODIN_ARCH == .amd64 {
 		// While it may be unwise to do deserialization here on our
 		// own when fiat-crypto provides equivalent functionality,
 		// doing it this way provides a little under 3x performance
@@ -52,7 +52,7 @@ keccakf :: proc "contextless" (st: ^[25]u64) {
    t: u64       = ---
    bc: [5]u64   = ---

-    when ODIN_ENDIAN != "little" {
+    when ODIN_ENDIAN != .Little {
        v: uintptr = ---
        for i = 0; i < 25; i += 1 {
            v := uintptr(&st[i])
@@ -98,7 +98,7 @@ keccakf :: proc "contextless" (st: ^[25]u64) {
        st[0] ~= keccakf_rndc[r]
    }

-    when ODIN_ENDIAN != "little" {
+    when ODIN_ENDIAN != .Little {
        for i = 0; i < 25; i += 1 {
            v = uintptr(&st[i])
            t = st[i]
@@ -17,16 +17,21 @@ import "core:io"
    High level API
 */

+DIGEST_SIZE_224 :: 28
+DIGEST_SIZE_256 :: 32
+DIGEST_SIZE_384 :: 48
+DIGEST_SIZE_512 :: 64
+
 // hash_string_224 will hash the given input and return the
 // computed hash
-hash_string_224 :: proc "contextless" (data: string) -> [28]byte {
+hash_string_224 :: proc "contextless" (data: string) -> [DIGEST_SIZE_224]byte {
    return hash_bytes_224(transmute([]byte)(data))
 }

 // hash_bytes_224 will hash the given input and return the
 // computed hash
-hash_bytes_224 :: proc "contextless" (data: []byte) -> [28]byte {
-    hash: [28]byte
+hash_bytes_224 :: proc "contextless" (data: []byte) -> [DIGEST_SIZE_224]byte {
+    hash: [DIGEST_SIZE_224]byte
    ctx: Blake256_Context
    ctx.is224 = true
    init(&ctx)
@@ -35,10 +40,29 @@ hash_bytes_224 :: proc "contextless" (data: []byte) -> [28]byte {
    return hash
 }

+// hash_string_to_buffer_224 will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer_224 :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer_224(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer_224 will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer_224 :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE_224, "Size of destination buffer is smaller than the digest size")
+    ctx: Blake256_Context
+    ctx.is224 = true
+    init(&ctx)
+    update(&ctx, data)
+    final(&ctx, hash)
+}
+
 // hash_stream_224 will read the stream in chunks and compute a
 // hash from its contents
-hash_stream_224 :: proc(s: io.Stream) -> ([28]byte, bool) {
-    hash: [28]byte
+hash_stream_224 :: proc(s: io.Stream) -> ([DIGEST_SIZE_224]byte, bool) {
+    hash: [DIGEST_SIZE_224]byte
    ctx: Blake256_Context
    ctx.is224 = true
    init(&ctx)
@@ -57,7 +81,7 @@ hash_stream_224 :: proc(s: io.Stream) -> ([28]byte, bool) {

 // hash_file_224 will read the file provided by the given handle
 // and compute a hash
-hash_file_224 :: proc(hd: os.Handle, load_at_once := false) -> ([28]byte, bool) {
+hash_file_224 :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE_224]byte, bool) {
    if !load_at_once {
        return hash_stream_224(os.stream_from_handle(hd))
    } else {
@@ -65,7 +89,7 @@ hash_file_224 :: proc(hd: os.Handle, load_at_once := false) -> ([28]byte, bool)
            return hash_bytes_224(buf[:]), ok
        }
    }
-    return [28]byte{}, false
+    return [DIGEST_SIZE_224]byte{}, false
 }

 hash_224 :: proc {
@@ -73,18 +97,20 @@ hash_224 :: proc {
    hash_file_224,
    hash_bytes_224,
    hash_string_224,
+    hash_bytes_to_buffer_224,
+    hash_string_to_buffer_224,
 }

 // hash_string_256 will hash the given input and return the
 // computed hash
-hash_string_256 :: proc "contextless" (data: string) -> [32]byte {
+hash_string_256 :: proc "contextless" (data: string) -> [DIGEST_SIZE_256]byte {
    return hash_bytes_256(transmute([]byte)(data))
 }

 // hash_bytes_256 will hash the given input and return the
 // computed hash
-hash_bytes_256 :: proc "contextless" (data: []byte) -> [32]byte {
-    hash: [32]byte
+hash_bytes_256 :: proc "contextless" (data: []byte) -> [DIGEST_SIZE_256]byte {
+    hash: [DIGEST_SIZE_256]byte
    ctx: Blake256_Context
    ctx.is224 = false
    init(&ctx)
@@ -93,10 +119,29 @@ hash_bytes_256 :: proc "contextless" (data: []byte) -> [32]byte {
    return hash
 }

+// hash_string_to_buffer_256 will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer_256 :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer_256(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer_256 will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer_256 :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE_256, "Size of destination buffer is smaller than the digest size")
+    ctx: Blake256_Context
+    ctx.is224 = false
+    init(&ctx)
+    update(&ctx, data)
+    final(&ctx, hash)
+}
+
 // hash_stream_256 will read the stream in chunks and compute a
 // hash from its contents
-hash_stream_256 :: proc(s: io.Stream) -> ([32]byte, bool) {
-    hash: [32]byte
+hash_stream_256 :: proc(s: io.Stream) -> ([DIGEST_SIZE_256]byte, bool) {
+    hash: [DIGEST_SIZE_256]byte
    ctx: Blake256_Context
    ctx.is224 = false
    init(&ctx)
@@ -115,7 +160,7 @@ hash_stream_256 :: proc(s: io.Stream) -> ([32]byte, bool) {

 // hash_file_256 will read the file provided by the given handle
 // and compute a hash
-hash_file_256 :: proc(hd: os.Handle, load_at_once := false) -> ([32]byte, bool) {
+hash_file_256 :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE_256]byte, bool) {
    if !load_at_once {
        return hash_stream_256(os.stream_from_handle(hd))
    } else {
@@ -123,7 +168,7 @@ hash_file_256 :: proc(hd: os.Handle, load_at_once := false) -> ([32]byte, bool)
            return hash_bytes_256(buf[:]), ok
        }
    }
-    return [32]byte{}, false
+    return [DIGEST_SIZE_256]byte{}, false
 }

 hash_256 :: proc {
@@ -131,18 +176,20 @@ hash_256 :: proc {
    hash_file_256,
    hash_bytes_256,
    hash_string_256,
+    hash_bytes_to_buffer_256,
+    hash_string_to_buffer_256,
 }

 // hash_string_384 will hash the given input and return the
 // computed hash
-hash_string_384 :: proc "contextless" (data: string) -> [48]byte {
+hash_string_384 :: proc "contextless" (data: string) -> [DIGEST_SIZE_384]byte {
    return hash_bytes_384(transmute([]byte)(data))
 }

 // hash_bytes_384 will hash the given input and return the
 // computed hash
-hash_bytes_384 :: proc "contextless" (data: []byte) -> [48]byte {
-    hash: [48]byte
+hash_bytes_384 :: proc "contextless" (data: []byte) -> [DIGEST_SIZE_384]byte {
+    hash: [DIGEST_SIZE_384]byte
    ctx: Blake512_Context
    ctx.is384 = true
    init(&ctx)
@@ -151,10 +198,29 @@ hash_bytes_384 :: proc "contextless" (data: []byte) -> [48]byte {
    return hash
 }

+// hash_string_to_buffer_384 will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer_384 :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer_384(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer_384 will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer_384 :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE_384, "Size of destination buffer is smaller than the digest size")
+    ctx: Blake512_Context
+    ctx.is384 = true
+    init(&ctx)
+    update(&ctx, data)
+    final(&ctx, hash)
+}
+
 // hash_stream_384 will read the stream in chunks and compute a
 // hash from its contents
-hash_stream_384 :: proc(s: io.Stream) -> ([48]byte, bool) {
-    hash: [48]byte
+hash_stream_384 :: proc(s: io.Stream) -> ([DIGEST_SIZE_384]byte, bool) {
+    hash: [DIGEST_SIZE_384]byte
    ctx: Blake512_Context
    ctx.is384 = true
    init(&ctx)
@@ -173,7 +239,7 @@ hash_stream_384 :: proc(s: io.Stream) -> ([48]byte, bool) {

 // hash_file_384 will read the file provided by the given handle
 // and compute a hash
-hash_file_384 :: proc(hd: os.Handle, load_at_once := false) -> ([48]byte, bool) {
+hash_file_384 :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE_384]byte, bool) {
    if !load_at_once {
        return hash_stream_384(os.stream_from_handle(hd))
    } else {
@@ -181,7 +247,7 @@ hash_file_384 :: proc(hd: os.Handle, load_at_once := false) -> ([48]byte, bool)
            return hash_bytes_384(buf[:]), ok
        }
    }
-    return [48]byte{}, false
+    return [DIGEST_SIZE_384]byte{}, false
 }

 hash_384 :: proc {
@@ -189,18 +255,20 @@ hash_384 :: proc {
    hash_file_384,
    hash_bytes_384,
    hash_string_384,
+    hash_bytes_to_buffer_384,
+    hash_string_to_buffer_384,
 }

 // hash_string_512 will hash the given input and return the
 // computed hash
-hash_string_512 :: proc "contextless" (data: string) -> [64]byte {
+hash_string_512 :: proc "contextless" (data: string) -> [DIGEST_SIZE_512]byte {
    return hash_bytes_512(transmute([]byte)(data))
 }

 // hash_bytes_512 will hash the given input and return the
 // computed hash
-hash_bytes_512 :: proc "contextless" (data: []byte) -> [64]byte {
-    hash: [64]byte
+hash_bytes_512 :: proc "contextless" (data: []byte) -> [DIGEST_SIZE_512]byte {
+    hash: [DIGEST_SIZE_512]byte
    ctx: Blake512_Context
    ctx.is384 = false
    init(&ctx)
@@ -209,10 +277,29 @@ hash_bytes_512 :: proc "contextless" (data: []byte) -> [64]byte {
    return hash
 }

+// hash_string_to_buffer_512 will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer_512 :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer_512(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer_512 will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer_512 :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE_512, "Size of destination buffer is smaller than the digest size")
+    ctx: Blake512_Context
+    ctx.is384 = false
+    init(&ctx)
+    update(&ctx, data)
+    final(&ctx, hash)
+}
+
 // hash_stream_512 will read the stream in chunks and compute a
 // hash from its contents
-hash_stream_512 :: proc(s: io.Stream) -> ([64]byte, bool) {
-    hash: [64]byte
+hash_stream_512 :: proc(s: io.Stream) -> ([DIGEST_SIZE_512]byte, bool) {
+    hash: [DIGEST_SIZE_512]byte
    ctx: Blake512_Context
    ctx.is384 = false
    init(&ctx)
@@ -231,7 +318,7 @@ hash_stream_512 :: proc(s: io.Stream) -> ([64]byte, bool) {

 // hash_file_512 will read the file provided by the given handle
 // and compute a hash
-hash_file_512 :: proc(hd: os.Handle, load_at_once := false) -> ([64]byte, bool) {
+hash_file_512 :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE_512]byte, bool) {
    if !load_at_once {
        return hash_stream_512(os.stream_from_handle(hd))
    } else {
@@ -239,7 +326,7 @@ hash_file_512 :: proc(hd: os.Handle, load_at_once := false) -> ([64]byte, bool)
            return hash_bytes_512(buf[:]), ok
        }
    }
-    return [64]byte{}, false
+    return [DIGEST_SIZE_512]byte{}, false
 }

 hash_512 :: proc {
@@ -247,6 +334,8 @@ hash_512 :: proc {
    hash_file_512,
    hash_bytes_512,
    hash_string_512,
+    hash_bytes_to_buffer_512,
+    hash_string_to_buffer_512,
 }

 /*
@@ -20,16 +20,18 @@ import "../_blake2"
    High level API
 */

+DIGEST_SIZE :: 64
+
 // hash_string will hash the given input and return the
 // computed hash
-hash_string :: proc(data: string) -> [64]byte {
+hash_string :: proc(data: string) -> [DIGEST_SIZE]byte {
    return hash_bytes(transmute([]byte)(data))
 }

 // hash_bytes will hash the given input and return the
 // computed hash
-hash_bytes :: proc(data: []byte) -> [64]byte {
-    hash: [64]byte
+hash_bytes :: proc(data: []byte) -> [DIGEST_SIZE]byte {
+    hash: [DIGEST_SIZE]byte
    ctx: _blake2.Blake2b_Context
    cfg: _blake2.Blake2_Config
    cfg.size = _blake2.BLAKE2B_SIZE
@@ -40,10 +42,32 @@ hash_bytes :: proc(data: []byte) -> [64]byte {
    return hash
 }

+// hash_string_to_buffer will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE, "Size of destination buffer is smaller than the digest size")
+    ctx: _blake2.Blake2b_Context
+    cfg: _blake2.Blake2_Config
+    cfg.size = _blake2.BLAKE2B_SIZE
+    ctx.cfg  = cfg
+    _blake2.init(&ctx)
+    _blake2.update(&ctx, data)
+    _blake2.final(&ctx, hash)
+}
+
+
 // hash_stream will read the stream in chunks and compute a
 // hash from its contents
-hash_stream :: proc(s: io.Stream) -> ([64]byte, bool) {
-    hash: [64]byte
+hash_stream :: proc(s: io.Stream) -> ([DIGEST_SIZE]byte, bool) {
+    hash: [DIGEST_SIZE]byte
    ctx: _blake2.Blake2b_Context
    cfg: _blake2.Blake2_Config
    cfg.size = _blake2.BLAKE2B_SIZE
@@ -64,7 +88,7 @@ hash_stream :: proc(s: io.Stream) -> ([64]byte, bool) {

 // hash_file will read the file provided by the given handle
 // and compute a hash
-hash_file :: proc(hd: os.Handle, load_at_once := false) -> ([64]byte, bool) {
+hash_file :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE]byte, bool) {
    if !load_at_once {
        return hash_stream(os.stream_from_handle(hd))
    } else {
@@ -72,7 +96,7 @@ hash_file :: proc(hd: os.Handle, load_at_once := false) -> ([64]byte, bool) {
            return hash_bytes(buf[:]), ok
        }
    }
-    return [64]byte{}, false
+    return [DIGEST_SIZE]byte{}, false
 }

 hash :: proc {
@@ -80,6 +104,8 @@ hash :: proc {
    hash_file,
    hash_bytes,
    hash_string,
+    hash_bytes_to_buffer,
+    hash_string_to_buffer,
 }

 /*
@@ -20,16 +20,18 @@ import "../_blake2"
    High level API
 */

+DIGEST_SIZE :: 32
+
 // hash_string will hash the given input and return the
 // computed hash
-hash_string :: proc(data: string) -> [32]byte {
+hash_string :: proc(data: string) -> [DIGEST_SIZE]byte {
    return hash_bytes(transmute([]byte)(data))
 }

 // hash_bytes will hash the given input and return the
 // computed hash
-hash_bytes :: proc(data: []byte) -> [32]byte {
-    hash: [32]byte
+hash_bytes :: proc(data: []byte) -> [DIGEST_SIZE]byte {
+    hash: [DIGEST_SIZE]byte
    ctx: _blake2.Blake2s_Context
    cfg: _blake2.Blake2_Config
    cfg.size = _blake2.BLAKE2S_SIZE
@@ -40,10 +42,32 @@ hash_bytes :: proc(data: []byte) -> [32]byte {
    return hash
 }

+
+// hash_string_to_buffer will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE, "Size of destination buffer is smaller than the digest size")
+    ctx: _blake2.Blake2s_Context
+    cfg: _blake2.Blake2_Config
+    cfg.size = _blake2.BLAKE2S_SIZE
+    ctx.cfg  = cfg
+    _blake2.init(&ctx)
+    _blake2.update(&ctx, data)
+    _blake2.final(&ctx, hash)
+}
+
 // hash_stream will read the stream in chunks and compute a
 // hash from its contents
-hash_stream :: proc(s: io.Stream) -> ([32]byte, bool) {
-    hash: [32]byte
+hash_stream :: proc(s: io.Stream) -> ([DIGEST_SIZE]byte, bool) {
+    hash: [DIGEST_SIZE]byte
    ctx: _blake2.Blake2s_Context
    cfg: _blake2.Blake2_Config
    cfg.size = _blake2.BLAKE2S_SIZE
@@ -64,7 +88,7 @@ hash_stream :: proc(s: io.Stream) -> ([32]byte, bool) {

 // hash_file will read the file provided by the given handle
 // and compute a hash
-hash_file :: proc(hd: os.Handle, load_at_once := false) -> ([32]byte, bool) {
+hash_file :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE]byte, bool) {
    if !load_at_once {
        return hash_stream(os.stream_from_handle(hd))
    } else {
@@ -72,7 +96,7 @@ hash_file :: proc(hd: os.Handle, load_at_once := false) -> ([32]byte, bool) {
            return hash_bytes(buf[:]), ok
        }
    }
-    return [32]byte{}, false
+    return [DIGEST_SIZE]byte{}, false
 }

 hash :: proc {
@@ -80,6 +104,8 @@ hash :: proc {
    hash_file,
    hash_bytes,
    hash_string,
+    hash_bytes_to_buffer,
+    hash_string_to_buffer,
 }

 /*
@@ -346,7 +346,7 @@ _do_blocks :: proc (ctx: ^Context, dst, src: []byte, nr_blocks: int) {
 		// Until dedicated assembly can be written leverage the fact that
 		// the callers of this routine ensure that src/dst are valid.

-		when ODIN_ARCH == "386" || ODIN_ARCH == "amd64" {
+		when ODIN_ARCH == .i386 || ODIN_ARCH == .amd64 {
 			// util.PUT_U32_LE/util.U32_LE are not required on little-endian
 			// systems that also happen to not be strict about aligned
 			// memory access.
@@ -18,16 +18,18 @@ import "core:io"
    High level API
 */

+DIGEST_SIZE :: 32
+
 // hash_string will hash the given input and return the
 // computed hash
-hash_string :: proc(data: string) -> [32]byte {
+hash_string :: proc(data: string) -> [DIGEST_SIZE]byte {
    return hash_bytes(transmute([]byte)(data))
 }

 // hash_bytes will hash the given input and return the
 // computed hash
-hash_bytes :: proc(data: []byte) -> [32]byte {
-    hash: [32]byte
+hash_bytes :: proc(data: []byte) -> [DIGEST_SIZE]byte {
+    hash: [DIGEST_SIZE]byte
    ctx: Gost_Context
    init(&ctx)
    update(&ctx, data)
@@ -35,10 +37,28 @@ hash_bytes :: proc(data: []byte) -> [32]byte {
    return hash
 }

+// hash_string_to_buffer will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE, "Size of destination buffer is smaller than the digest size")
+    ctx: Gost_Context
+    init(&ctx)
+    update(&ctx, data)
+    final(&ctx, hash)
+}
+
 // hash_stream will read the stream in chunks and compute a
 // hash from its contents
-hash_stream :: proc(s: io.Stream) -> ([32]byte, bool) {
-    hash: [32]byte
+hash_stream :: proc(s: io.Stream) -> ([DIGEST_SIZE]byte, bool) {
+    hash: [DIGEST_SIZE]byte
    ctx: Gost_Context
    init(&ctx)
    buf := make([]byte, 512)
@@ -56,7 +76,7 @@ hash_stream :: proc(s: io.Stream) -> ([32]byte, bool) {

 // hash_file will read the file provided by the given handle
 // and compute a hash
-hash_file :: proc(hd: os.Handle, load_at_once := false) -> ([32]byte, bool) {
+hash_file :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE]byte, bool) {
    if !load_at_once {
        return hash_stream(os.stream_from_handle(hd))
    } else {
@@ -64,7 +84,7 @@ hash_file :: proc(hd: os.Handle, load_at_once := false) -> ([32]byte, bool) {
            return hash_bytes(buf[:]), ok
        }
    }
-    return [32]byte{}, false
+    return [DIGEST_SIZE]byte{}, false
 }

 hash :: proc {
@@ -72,6 +92,8 @@ hash :: proc {
    hash_file,
    hash_bytes,
    hash_string,
+    hash_bytes_to_buffer,
+    hash_string_to_buffer,
 }

 /*
@@ -17,16 +17,21 @@ import "core:io"
    High level API
 */

+DIGEST_SIZE_224 :: 28
+DIGEST_SIZE_256 :: 32
+DIGEST_SIZE_384 :: 48
+DIGEST_SIZE_512 :: 64
+
 // hash_string_224 will hash the given input and return the
 // computed hash
-hash_string_224 :: proc(data: string) -> [28]byte {
+hash_string_224 :: proc(data: string) -> [DIGEST_SIZE_224]byte {
    return hash_bytes_224(transmute([]byte)(data))
 }

 // hash_bytes_224 will hash the given input and return the
 // computed hash
-hash_bytes_224 :: proc(data: []byte) -> [28]byte {
-    hash: [28]byte
+hash_bytes_224 :: proc(data: []byte) -> [DIGEST_SIZE_224]byte {
+    hash: [DIGEST_SIZE_224]byte
    ctx: Groestl_Context
    ctx.hashbitlen = 224
    init(&ctx)
@@ -35,10 +40,29 @@ hash_bytes_224 :: proc(data: []byte) -> [28]byte {
    return hash
 }

+// hash_string_to_buffer_224 will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer_224 :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer_224(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer_224 will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer_224 :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE_224, "Size of destination buffer is smaller than the digest size")
+    ctx: Groestl_Context
+    ctx.hashbitlen = 224
+    init(&ctx)
+    update(&ctx, data)
+    final(&ctx, hash)
+}
+
 // hash_stream_224 will read the stream in chunks and compute a
 // hash from its contents
-hash_stream_224 :: proc(s: io.Stream) -> ([28]byte, bool) {
-    hash: [28]byte
+hash_stream_224 :: proc(s: io.Stream) -> ([DIGEST_SIZE_224]byte, bool) {
+    hash: [DIGEST_SIZE_224]byte
    ctx: Groestl_Context
    ctx.hashbitlen = 224
    init(&ctx)
@@ -57,7 +81,7 @@ hash_stream_224 :: proc(s: io.Stream) -> ([28]byte, bool) {

 // hash_file_224 will read the file provided by the given handle
 // and compute a hash
-hash_file_224 :: proc(hd: os.Handle, load_at_once := false) -> ([28]byte, bool) {
+hash_file_224 :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE_224]byte, bool) {
    if !load_at_once {
        return hash_stream_224(os.stream_from_handle(hd))
    } else {
@@ -65,7 +89,7 @@ hash_file_224 :: proc(hd: os.Handle, load_at_once := false) -> ([28]byte, bool)
            return hash_bytes_224(buf[:]), ok
        }
    }
-    return [28]byte{}, false
+    return [DIGEST_SIZE_224]byte{}, false
 }

 hash_224 :: proc {
@@ -73,18 +97,20 @@ hash_224 :: proc {
    hash_file_224,
    hash_bytes_224,
    hash_string_224,
+    hash_bytes_to_buffer_224,
+    hash_string_to_buffer_224,
 }

 // hash_string_256 will hash the given input and return the
 // computed hash
-hash_string_256 :: proc(data: string) -> [32]byte {
+hash_string_256 :: proc(data: string) -> [DIGEST_SIZE_256]byte {
    return hash_bytes_256(transmute([]byte)(data))
 }

 // hash_bytes_256 will hash the given input and return the
 // computed hash
-hash_bytes_256 :: proc(data: []byte) -> [32]byte {
-    hash: [32]byte
+hash_bytes_256 :: proc(data: []byte) -> [DIGEST_SIZE_256]byte {
+    hash: [DIGEST_SIZE_256]byte
    ctx: Groestl_Context
    ctx.hashbitlen = 256
    init(&ctx)
@@ -93,10 +119,29 @@ hash_bytes_256 :: proc(data: []byte) -> [32]byte {
    return hash
 }

+// hash_string_to_buffer_256 will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer_256 :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer_256(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer_256 will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer_256 :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE_256, "Size of destination buffer is smaller than the digest size")
+    ctx: Groestl_Context
+    ctx.hashbitlen = 256
+    init(&ctx)
+    update(&ctx, data)
+    final(&ctx, hash)
+}
+
 // hash_stream_256 will read the stream in chunks and compute a
 // hash from its contents
-hash_stream_256 :: proc(s: io.Stream) -> ([32]byte, bool) {
-    hash: [32]byte
+hash_stream_256 :: proc(s: io.Stream) -> ([DIGEST_SIZE_256]byte, bool) {
+    hash: [DIGEST_SIZE_256]byte
    ctx: Groestl_Context
    ctx.hashbitlen = 256
    init(&ctx)
@@ -115,7 +160,7 @@ hash_stream_256 :: proc(s: io.Stream) -> ([32]byte, bool) {

 // hash_file_256 will read the file provided by the given handle
 // and compute a hash
-hash_file_256 :: proc(hd: os.Handle, load_at_once := false) -> ([32]byte, bool) {
+hash_file_256 :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE_256]byte, bool) {
    if !load_at_once {
        return hash_stream_256(os.stream_from_handle(hd))
    } else {
@@ -123,7 +168,7 @@ hash_file_256 :: proc(hd: os.Handle, load_at_once := false) -> ([32]byte, bool)
            return hash_bytes_256(buf[:]), ok
        }
    }
-    return [32]byte{}, false
+    return [DIGEST_SIZE_256]byte{}, false
 }

 hash_256 :: proc {
@@ -131,18 +176,20 @@ hash_256 :: proc {
    hash_file_256,
    hash_bytes_256,
    hash_string_256,
+    hash_bytes_to_buffer_256,
+    hash_string_to_buffer_256,
 }

 // hash_string_384 will hash the given input and return the
 // computed hash
-hash_string_384 :: proc(data: string) -> [48]byte {
+hash_string_384 :: proc(data: string) -> [DIGEST_SIZE_384]byte {
    return hash_bytes_384(transmute([]byte)(data))
 }

 // hash_bytes_384 will hash the given input and return the
 // computed hash
-hash_bytes_384 :: proc(data: []byte) -> [48]byte {
-    hash: [48]byte
+hash_bytes_384 :: proc(data: []byte) -> [DIGEST_SIZE_384]byte {
+    hash: [DIGEST_SIZE_384]byte
    ctx: Groestl_Context
    ctx.hashbitlen = 384
    init(&ctx)
@@ -151,10 +198,29 @@ hash_bytes_384 :: proc(data: []byte) -> [48]byte {
    return hash
 }

+// hash_string_to_buffer_384 will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer_384 :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer_384(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer_384 will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer_384 :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE_384, "Size of destination buffer is smaller than the digest size")
+    ctx: Groestl_Context
+    ctx.hashbitlen = 384
+    init(&ctx)
+    update(&ctx, data)
+    final(&ctx, hash)
+}
+
 // hash_stream_384 will read the stream in chunks and compute a
 // hash from its contents
-hash_stream_384 :: proc(s: io.Stream) -> ([48]byte, bool) {
-    hash: [48]byte
+hash_stream_384 :: proc(s: io.Stream) -> ([DIGEST_SIZE_384]byte, bool) {
+    hash: [DIGEST_SIZE_384]byte
    ctx: Groestl_Context
    ctx.hashbitlen = 384
    init(&ctx)
@@ -173,7 +239,7 @@ hash_stream_384 :: proc(s: io.Stream) -> ([48]byte, bool) {

 // hash_file_384 will read the file provided by the given handle
 // and compute a hash
-hash_file_384 :: proc(hd: os.Handle, load_at_once := false) -> ([48]byte, bool) {
+hash_file_384 :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE_384]byte, bool) {
    if !load_at_once {
        return hash_stream_384(os.stream_from_handle(hd))
    } else {
@@ -181,7 +247,7 @@ hash_file_384 :: proc(hd: os.Handle, load_at_once := false) -> ([48]byte, bool)
            return hash_bytes_384(buf[:]), ok
        }
    }
-    return [48]byte{}, false
+    return [DIGEST_SIZE_384]byte{}, false
 }

 hash_384 :: proc {
@@ -189,18 +255,20 @@ hash_384 :: proc {
    hash_file_384,
    hash_bytes_384,
    hash_string_384,
+    hash_bytes_to_buffer_384,
+    hash_string_to_buffer_384,
 }

 // hash_string_512 will hash the given input and return the
 // computed hash
-hash_string_512 :: proc(data: string) -> [64]byte {
+hash_string_512 :: proc(data: string) -> [DIGEST_SIZE_512]byte {
    return hash_bytes_512(transmute([]byte)(data))
 }

 // hash_bytes_512 will hash the given input and return the
 // computed hash
-hash_bytes_512 :: proc(data: []byte) -> [64]byte {
-    hash: [64]byte
+hash_bytes_512 :: proc(data: []byte) -> [DIGEST_SIZE_512]byte {
+    hash: [DIGEST_SIZE_512]byte
    ctx: Groestl_Context
    ctx.hashbitlen = 512
    init(&ctx)
@@ -209,10 +277,29 @@ hash_bytes_512 :: proc(data: []byte) -> [64]byte {
    return hash
 }

+// hash_string_to_buffer_512 will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer_512 :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer_512(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer_512 will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer_512 :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE_512, "Size of destination buffer is smaller than the digest size")
+    ctx: Groestl_Context
+    ctx.hashbitlen = 512
+    init(&ctx)
+    update(&ctx, data)
+    final(&ctx, hash)
+}
+
 // hash_stream_512 will read the stream in chunks and compute a
 // hash from its contents
-hash_stream_512 :: proc(s: io.Stream) -> ([64]byte, bool) {
-    hash: [64]byte
+hash_stream_512 :: proc(s: io.Stream) -> ([DIGEST_SIZE_512]byte, bool) {
+    hash: [DIGEST_SIZE_512]byte
    ctx: Groestl_Context
    ctx.hashbitlen = 512
    init(&ctx)
@@ -231,7 +318,7 @@ hash_stream_512 :: proc(s: io.Stream) -> ([64]byte, bool) {

 // hash_file_512 will read the file provided by the given handle
 // and compute a hash
-hash_file_512 :: proc(hd: os.Handle, load_at_once := false) -> ([64]byte, bool) {
+hash_file_512 :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE_512]byte, bool) {
    if !load_at_once {
        return hash_stream_512(os.stream_from_handle(hd))
    } else {
@@ -239,7 +326,7 @@ hash_file_512 :: proc(hd: os.Handle, load_at_once := false) -> ([64]byte, bool)
            return hash_bytes_512(buf[:]), ok
        }
    }
-    return [64]byte{}, false
+    return [DIGEST_SIZE_512]byte{}, false
 }

 hash_512 :: proc {
@@ -247,6 +334,8 @@ hash_512 :: proc {
    hash_file_512,
    hash_bytes_512,
    hash_string_512,
+    hash_bytes_to_buffer_512,
+    hash_string_to_buffer_512,
 }

 /*
@@ -17,16 +17,21 @@ import "core:io"
    High level API
 */

+DIGEST_SIZE_224 :: 28
+DIGEST_SIZE_256 :: 32
+DIGEST_SIZE_384 :: 48
+DIGEST_SIZE_512 :: 64
+
 // hash_string_224 will hash the given input and return the
 // computed hash
-hash_string_224 :: proc(data: string) -> [28]byte {
+hash_string_224 :: proc(data: string) -> [DIGEST_SIZE_224]byte {
    return hash_bytes_224(transmute([]byte)(data))
 }

 // hash_bytes_224 will hash the given input and return the
 // computed hash
-hash_bytes_224 :: proc(data: []byte) -> [28]byte {
-    hash: [28]byte
+hash_bytes_224 :: proc(data: []byte) -> [DIGEST_SIZE_224]byte {
+    hash: [DIGEST_SIZE_224]byte
    ctx: Jh_Context
    ctx.hashbitlen = 224
    init(&ctx)
@@ -35,10 +40,29 @@ hash_bytes_224 :: proc(data: []byte) -> [28]byte {
    return hash
 }

+// hash_string_to_buffer_224 will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer_224 :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer_224(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer_224 will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer_224 :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE_224, "Size of destination buffer is smaller than the digest size")
+    ctx: Jh_Context
+    ctx.hashbitlen = 224
+    init(&ctx)
+    update(&ctx, data)
+    final(&ctx, hash)
+}
+
 // hash_stream_224 will read the stream in chunks and compute a
 // hash from its contents
-hash_stream_224 :: proc(s: io.Stream) -> ([28]byte, bool) {
-    hash: [28]byte
+hash_stream_224 :: proc(s: io.Stream) -> ([DIGEST_SIZE_224]byte, bool) {
+    hash: [DIGEST_SIZE_224]byte
    ctx: Jh_Context
    ctx.hashbitlen = 224
    init(&ctx)
@@ -57,7 +81,7 @@ hash_stream_224 :: proc(s: io.Stream) -> ([28]byte, bool) {

 // hash_file_224 will read the file provided by the given handle
 // and compute a hash
-hash_file_224 :: proc(hd: os.Handle, load_at_once := false) -> ([28]byte, bool) {
+hash_file_224 :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE_224]byte, bool) {
    if !load_at_once {
        return hash_stream_224(os.stream_from_handle(hd))
    } else {
@@ -65,7 +89,7 @@ hash_file_224 :: proc(hd: os.Handle, load_at_once := false) -> ([28]byte, bool)
            return hash_bytes_224(buf[:]), ok
        }
    }
-    return [28]byte{}, false
+    return [DIGEST_SIZE_224]byte{}, false
 }

 hash_224 :: proc {
@@ -73,18 +97,20 @@ hash_224 :: proc {
    hash_file_224,
    hash_bytes_224,
    hash_string_224,
+    hash_bytes_to_buffer_224,
+    hash_string_to_buffer_224,
 }

 // hash_string_256 will hash the given input and return the
 // computed hash
-hash_string_256 :: proc(data: string) -> [32]byte {
+hash_string_256 :: proc(data: string) -> [DIGEST_SIZE_256]byte {
    return hash_bytes_256(transmute([]byte)(data))
 }

 // hash_bytes_256 will hash the given input and return the
 // computed hash
-hash_bytes_256 :: proc(data: []byte) -> [32]byte {
-    hash: [32]byte
+hash_bytes_256 :: proc(data: []byte) -> [DIGEST_SIZE_256]byte {
+    hash: [DIGEST_SIZE_256]byte
    ctx: Jh_Context
    ctx.hashbitlen = 256
    init(&ctx)
@@ -93,10 +119,29 @@ hash_bytes_256 :: proc(data: []byte) -> [32]byte {
    return hash
 }

+// hash_string_to_buffer_256 will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer_256 :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer_256(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer_256 will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer_256 :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE_256, "Size of destination buffer is smaller than the digest size")
+    ctx: Jh_Context
+    ctx.hashbitlen = 256
+    init(&ctx)
+    update(&ctx, data)
+    final(&ctx, hash)
+}
+
 // hash_stream_256 will read the stream in chunks and compute a
 // hash from its contents
-hash_stream_256 :: proc(s: io.Stream) -> ([32]byte, bool) {
-    hash: [32]byte
+hash_stream_256 :: proc(s: io.Stream) -> ([DIGEST_SIZE_256]byte, bool) {
+    hash: [DIGEST_SIZE_256]byte
    ctx: Jh_Context
    ctx.hashbitlen = 256
    init(&ctx)
@@ -115,7 +160,7 @@ hash_stream_256 :: proc(s: io.Stream) -> ([32]byte, bool) {

 // hash_file_256 will read the file provided by the given handle
 // and compute a hash
-hash_file_256 :: proc(hd: os.Handle, load_at_once := false) -> ([32]byte, bool) {
+hash_file_256 :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE_256]byte, bool) {
    if !load_at_once {
        return hash_stream_256(os.stream_from_handle(hd))
    } else {
@@ -123,7 +168,7 @@ hash_file_256 :: proc(hd: os.Handle, load_at_once := false) -> ([32]byte, bool)
            return hash_bytes_256(buf[:]), ok
        }
    }
-    return [32]byte{}, false
+    return [DIGEST_SIZE_256]byte{}, false
 }

 hash_256 :: proc {
@@ -131,18 +176,20 @@ hash_256 :: proc {
    hash_file_256,
    hash_bytes_256,
    hash_string_256,
+    hash_bytes_to_buffer_256,
+    hash_string_to_buffer_256,
 }

 // hash_string_384 will hash the given input and return the
 // computed hash
-hash_string_384 :: proc(data: string) -> [48]byte {
+hash_string_384 :: proc(data: string) -> [DIGEST_SIZE_384]byte {
    return hash_bytes_384(transmute([]byte)(data))
 }

 // hash_bytes_384 will hash the given input and return the
 // computed hash
-hash_bytes_384 :: proc(data: []byte) -> [48]byte {
-    hash: [48]byte
+hash_bytes_384 :: proc(data: []byte) -> [DIGEST_SIZE_384]byte {
+    hash: [DIGEST_SIZE_384]byte
    ctx: Jh_Context
    ctx.hashbitlen = 384
    init(&ctx)
@@ -151,10 +198,29 @@ hash_bytes_384 :: proc(data: []byte) -> [48]byte {
    return hash
 }

+// hash_string_to_buffer_384 will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer_384 :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer_384(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer_384 will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer_384 :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE_384, "Size of destination buffer is smaller than the digest size")
+    ctx: Jh_Context
+    ctx.hashbitlen = 384
+    init(&ctx)
+    update(&ctx, data)
+    final(&ctx, hash)
+}
+
 // hash_stream_384 will read the stream in chunks and compute a
 // hash from its contents
-hash_stream_384 :: proc(s: io.Stream) -> ([48]byte, bool) {
-    hash: [48]byte
+hash_stream_384 :: proc(s: io.Stream) -> ([DIGEST_SIZE_384]byte, bool) {
+    hash: [DIGEST_SIZE_384]byte
    ctx: Jh_Context
    ctx.hashbitlen = 384
    init(&ctx)
@@ -173,7 +239,7 @@ hash_stream_384 :: proc(s: io.Stream) -> ([48]byte, bool) {

 // hash_file_384 will read the file provided by the given handle
 // and compute a hash
-hash_file_384 :: proc(hd: os.Handle, load_at_once := false) -> ([48]byte, bool) {
+hash_file_384 :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE_384]byte, bool) {
    if !load_at_once {
        return hash_stream_384(os.stream_from_handle(hd))
    } else {
@@ -181,7 +247,7 @@ hash_file_384 :: proc(hd: os.Handle, load_at_once := false) -> ([48]byte, bool)
            return hash_bytes_384(buf[:]), ok
        }
    }
-    return [48]byte{}, false
+    return [DIGEST_SIZE_384]byte{}, false
 }

 hash_384 :: proc {
@@ -189,18 +255,20 @@ hash_384 :: proc {
    hash_file_384,
    hash_bytes_384,
    hash_string_384,
+    hash_bytes_to_buffer_384,
+    hash_string_to_buffer_384,
 }

 // hash_string_512 will hash the given input and return the
 // computed hash
-hash_string_512 :: proc(data: string) -> [64]byte {
+hash_string_512 :: proc(data: string) -> [DIGEST_SIZE_512]byte {
    return hash_bytes_512(transmute([]byte)(data))
 }

 // hash_bytes_512 will hash the given input and return the
 // computed hash
-hash_bytes_512 :: proc(data: []byte) -> [64]byte {
-    hash: [64]byte
+hash_bytes_512 :: proc(data: []byte) -> [DIGEST_SIZE_512]byte {
+    hash: [DIGEST_SIZE_512]byte
    ctx: Jh_Context
    ctx.hashbitlen = 512
    init(&ctx)
@@ -209,10 +277,29 @@ hash_bytes_512 :: proc(data: []byte) -> [64]byte {
    return hash
 }

+// hash_string_to_buffer_512 will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer_512 :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer_512(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer_512 will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer_512 :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE_512, "Size of destination buffer is smaller than the digest size")
+    ctx: Jh_Context
+    ctx.hashbitlen = 512
+    init(&ctx)
+    update(&ctx, data)
+    final(&ctx, hash)
+}
+
 // hash_stream_512 will read the stream in chunks and compute a
 // hash from its contents
-hash_stream_512 :: proc(s: io.Stream) -> ([64]byte, bool) {
-    hash: [64]byte
+hash_stream_512 :: proc(s: io.Stream) -> ([DIGEST_SIZE_512]byte, bool) {
+    hash: [DIGEST_SIZE_512]byte
    ctx: Jh_Context
    ctx.hashbitlen = 512
    init(&ctx)
@@ -231,7 +318,7 @@ hash_stream_512 :: proc(s: io.Stream) -> ([64]byte, bool) {

 // hash_file_512 will read the file provided by the given handle
 // and compute a hash
-hash_file_512 :: proc(hd: os.Handle, load_at_once := false) -> ([64]byte, bool) {
+hash_file_512 :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE_512]byte, bool) {
    if !load_at_once {
        return hash_stream_512(os.stream_from_handle(hd))
    } else {
@@ -239,7 +326,7 @@ hash_file_512 :: proc(hd: os.Handle, load_at_once := false) -> ([64]byte, bool)
            return hash_bytes_512(buf[:]), ok
        }
    }
-    return [64]byte{}, false
+    return [DIGEST_SIZE_512]byte{}, false
 }

 hash_512 :: proc {
@@ -247,6 +334,8 @@ hash_512 :: proc {
    hash_file_512,
    hash_bytes_512,
    hash_string_512,
+    hash_bytes_to_buffer_512,
+    hash_string_to_buffer_512,
 }

 /*
@@ -21,18 +21,23 @@ import "../_sha3"
    High level API
 */

+DIGEST_SIZE_224 :: 28
+DIGEST_SIZE_256 :: 32
+DIGEST_SIZE_384 :: 48
+DIGEST_SIZE_512 :: 64
+
 // hash_string_224 will hash the given input and return the
 // computed hash
-hash_string_224 :: proc(data: string) -> [28]byte {
+hash_string_224 :: proc(data: string) -> [DIGEST_SIZE_224]byte {
    return hash_bytes_224(transmute([]byte)(data))
 }

 // hash_bytes_224 will hash the given input and return the
 // computed hash
-hash_bytes_224 :: proc(data: []byte) -> [28]byte {
-    hash: [28]byte
+hash_bytes_224 :: proc(data: []byte) -> [DIGEST_SIZE_224]byte {
+    hash: [DIGEST_SIZE_224]byte
    ctx: _sha3.Sha3_Context
-    ctx.mdlen = 28
+    ctx.mdlen = DIGEST_SIZE_224
    ctx.is_keccak = true
    _sha3.init(&ctx)
    _sha3.update(&ctx, data)
@@ -40,12 +45,32 @@ hash_bytes_224 :: proc(data: []byte) -> [28]byte {
    return hash
 }

+// hash_string_to_buffer_224 will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer_224 :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer_224(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer_224 will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer_224 :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE_224, "Size of destination buffer is smaller than the digest size")
+    ctx: _sha3.Sha3_Context
+    ctx.mdlen = DIGEST_SIZE_224
+    ctx.is_keccak = true
+    _sha3.init(&ctx)
+    _sha3.update(&ctx, data)
+    _sha3.final(&ctx, hash)
+}
+
 // hash_stream_224 will read the stream in chunks and compute a
 // hash from its contents
-hash_stream_224 :: proc(s: io.Stream) -> ([28]byte, bool) {
-    hash: [28]byte
+hash_stream_224 :: proc(s: io.Stream) -> ([DIGEST_SIZE_224]byte, bool) {
+    hash: [DIGEST_SIZE_224]byte
    ctx: _sha3.Sha3_Context
-    ctx.mdlen = 28
+    ctx.mdlen = DIGEST_SIZE_224
    ctx.is_keccak = true
    _sha3.init(&ctx)
    buf := make([]byte, 512)
@@ -63,7 +88,7 @@ hash_stream_224 :: proc(s: io.Stream) -> ([28]byte, bool) {

 // hash_file_224 will read the file provided by the given handle
 // and compute a hash
-hash_file_224 :: proc(hd: os.Handle, load_at_once := false) -> ([28]byte, bool) {
+hash_file_224 :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE_224]byte, bool) {
    if !load_at_once {
        return hash_stream_224(os.stream_from_handle(hd))
    } else {
@@ -71,7 +96,7 @@ hash_file_224 :: proc(hd: os.Handle, load_at_once := false) -> ([28]byte, bool)
            return hash_bytes_224(buf[:]), ok
        }
    }
-    return [28]byte{}, false
+    return [DIGEST_SIZE_224]byte{}, false
 }

 hash_224 :: proc {
@@ -79,20 +104,22 @@ hash_224 :: proc {
    hash_file_224,
    hash_bytes_224,
    hash_string_224,
+    hash_bytes_to_buffer_224,
+    hash_string_to_buffer_224,
 }

 // hash_string_256 will hash the given input and return the
 // computed hash
-hash_string_256 :: proc(data: string) -> [32]byte {
+hash_string_256 :: proc(data: string) -> [DIGEST_SIZE_256]byte {
    return hash_bytes_256(transmute([]byte)(data))
 }

 // hash_bytes_256 will hash the given input and return the
 // computed hash
-hash_bytes_256 :: proc(data: []byte) -> [32]byte {
-    hash: [32]byte
+hash_bytes_256 :: proc(data: []byte) -> [DIGEST_SIZE_256]byte {
+    hash: [DIGEST_SIZE_256]byte
    ctx: _sha3.Sha3_Context
-    ctx.mdlen = 32
+    ctx.mdlen = DIGEST_SIZE_256
    ctx.is_keccak = true
    _sha3.init(&ctx)
    _sha3.update(&ctx, data)
@@ -100,12 +127,32 @@ hash_bytes_256 :: proc(data: []byte) -> [32]byte {
    return hash
 }

+// hash_string_to_buffer_256 will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer_256 :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer_256(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer_256 will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer_256 :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE_256, "Size of destination buffer is smaller than the digest size")
+    ctx: _sha3.Sha3_Context
+    ctx.mdlen = DIGEST_SIZE_256
+    ctx.is_keccak = true
+    _sha3.init(&ctx)
+    _sha3.update(&ctx, data)
+    _sha3.final(&ctx, hash)
+}
+
 // hash_stream_256 will read the stream in chunks and compute a
 // hash from its contents
-hash_stream_256 :: proc(s: io.Stream) -> ([32]byte, bool) {
-    hash: [32]byte
+hash_stream_256 :: proc(s: io.Stream) -> ([DIGEST_SIZE_256]byte, bool) {
+    hash: [DIGEST_SIZE_256]byte
    ctx: _sha3.Sha3_Context
-    ctx.mdlen = 32
+    ctx.mdlen = DIGEST_SIZE_256
    ctx.is_keccak = true
    _sha3.init(&ctx)
    buf := make([]byte, 512)
@@ -123,7 +170,7 @@ hash_stream_256 :: proc(s: io.Stream) -> ([32]byte, bool) {

 // hash_file_256 will read the file provided by the given handle
 // and compute a hash
-hash_file_256 :: proc(hd: os.Handle, load_at_once := false) -> ([32]byte, bool) {
+hash_file_256 :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE_256]byte, bool) {
    if !load_at_once {
        return hash_stream_256(os.stream_from_handle(hd))
    } else {
@@ -131,7 +178,7 @@ hash_file_256 :: proc(hd: os.Handle, load_at_once := false) -> ([32]byte, bool)
            return hash_bytes_256(buf[:]), ok
        }
    }
-    return [32]byte{}, false
+    return [DIGEST_SIZE_256]byte{}, false
 }

 hash_256 :: proc {
@@ -139,20 +186,22 @@ hash_256 :: proc {
    hash_file_256,
    hash_bytes_256,
    hash_string_256,
+    hash_bytes_to_buffer_256,
+    hash_string_to_buffer_256,
 }

 // hash_string_384 will hash the given input and return the
 // computed hash
-hash_string_384 :: proc(data: string) -> [48]byte {
+hash_string_384 :: proc(data: string) -> [DIGEST_SIZE_384]byte {
    return hash_bytes_384(transmute([]byte)(data))
 }

 // hash_bytes_384 will hash the given input and return the
 // computed hash
-hash_bytes_384 :: proc(data: []byte) -> [48]byte {
-    hash: [48]byte
+hash_bytes_384 :: proc(data: []byte) -> [DIGEST_SIZE_384]byte {
+    hash: [DIGEST_SIZE_384]byte
    ctx: _sha3.Sha3_Context
-    ctx.mdlen = 48
+    ctx.mdlen = DIGEST_SIZE_384
    ctx.is_keccak = true
    _sha3.init(&ctx)
    _sha3.update(&ctx, data)
@@ -160,12 +209,32 @@ hash_bytes_384 :: proc(data: []byte) -> [48]byte {
    return hash
 }

+// hash_string_to_buffer_384 will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer_384 :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer_384(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer_384 will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer_384 :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE_384, "Size of destination buffer is smaller than the digest size")
+    ctx: _sha3.Sha3_Context
+    ctx.mdlen = DIGEST_SIZE_384
+    ctx.is_keccak = true
+    _sha3.init(&ctx)
+    _sha3.update(&ctx, data)
+    _sha3.final(&ctx, hash)
+}
+
 // hash_stream_384 will read the stream in chunks and compute a
 // hash from its contents
-hash_stream_384 :: proc(s: io.Stream) -> ([48]byte, bool) {
-    hash: [48]byte
+hash_stream_384 :: proc(s: io.Stream) -> ([DIGEST_SIZE_384]byte, bool) {
+    hash: [DIGEST_SIZE_384]byte
    ctx: _sha3.Sha3_Context
-    ctx.mdlen = 48
+    ctx.mdlen = DIGEST_SIZE_384
    ctx.is_keccak = true
    _sha3.init(&ctx)
    buf := make([]byte, 512)
@@ -183,7 +252,7 @@ hash_stream_384 :: proc(s: io.Stream) -> ([48]byte, bool) {

 // hash_file_384 will read the file provided by the given handle
 // and compute a hash
-hash_file_384 :: proc(hd: os.Handle, load_at_once := false) -> ([48]byte, bool) {
+hash_file_384 :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE_384]byte, bool) {
    if !load_at_once {
        return hash_stream_384(os.stream_from_handle(hd))
    } else {
@@ -191,7 +260,7 @@ hash_file_384 :: proc(hd: os.Handle, load_at_once := false) -> ([48]byte, bool)
            return hash_bytes_384(buf[:]), ok
        }
    }
-    return [48]byte{}, false
+    return [DIGEST_SIZE_384]byte{}, false
 }

 hash_384 :: proc {
@@ -199,20 +268,22 @@ hash_384 :: proc {
    hash_file_384,
    hash_bytes_384,
    hash_string_384,
+    hash_bytes_to_buffer_384,
+    hash_string_to_buffer_384,
 }

 // hash_string_512 will hash the given input and return the
 // computed hash
-hash_string_512 :: proc(data: string) -> [64]byte {
+hash_string_512 :: proc(data: string) -> [DIGEST_SIZE_512]byte {
    return hash_bytes_512(transmute([]byte)(data))
 }

 // hash_bytes_512 will hash the given input and return the
 // computed hash
-hash_bytes_512 :: proc(data: []byte) -> [64]byte {
-    hash: [64]byte
+hash_bytes_512 :: proc(data: []byte) -> [DIGEST_SIZE_512]byte {
+    hash: [DIGEST_SIZE_512]byte
    ctx: _sha3.Sha3_Context
-    ctx.mdlen = 64
+    ctx.mdlen = DIGEST_SIZE_512
    ctx.is_keccak = true
    _sha3.init(&ctx)
    _sha3.update(&ctx, data)
@@ -220,12 +291,32 @@ hash_bytes_512 :: proc(data: []byte) -> [64]byte {
    return hash
 }

+// hash_string_to_buffer_512 will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer_512 :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer_512(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer_512 will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer_512 :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE_512, "Size of destination buffer is smaller than the digest size")
+    ctx: _sha3.Sha3_Context
+    ctx.mdlen = DIGEST_SIZE_512
+    ctx.is_keccak = true
+    _sha3.init(&ctx)
+    _sha3.update(&ctx, data)
+    _sha3.final(&ctx, hash)
+}
+
 // hash_stream_512 will read the stream in chunks and compute a
 // hash from its contents
-hash_stream_512 :: proc(s: io.Stream) -> ([64]byte, bool) {
-    hash: [64]byte
+hash_stream_512 :: proc(s: io.Stream) -> ([DIGEST_SIZE_512]byte, bool) {
+    hash: [DIGEST_SIZE_512]byte
    ctx: _sha3.Sha3_Context
-    ctx.mdlen = 64
+    ctx.mdlen = DIGEST_SIZE_512
    ctx.is_keccak = true
    _sha3.init(&ctx)
    buf := make([]byte, 512)
@@ -243,7 +334,7 @@ hash_stream_512 :: proc(s: io.Stream) -> ([64]byte, bool) {

 // hash_file_512 will read the file provided by the given handle
 // and compute a hash
-hash_file_512 :: proc(hd: os.Handle, load_at_once := false) -> ([64]byte, bool) {
+hash_file_512 :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE_512]byte, bool) {
    if !load_at_once {
        return hash_stream_512(os.stream_from_handle(hd))
    } else {
@@ -251,7 +342,7 @@ hash_file_512 :: proc(hd: os.Handle, load_at_once := false) -> ([64]byte, bool)
            return hash_bytes_512(buf[:]), ok
        }
    }
-    return [64]byte{}, false
+    return [DIGEST_SIZE_512]byte{}, false
 }

 hash_512 :: proc {
@@ -259,13 +350,15 @@ hash_512 :: proc {
    hash_file_512,
    hash_bytes_512,
    hash_string_512,
+    hash_bytes_to_buffer_512,
+    hash_string_to_buffer_512,
 }

 /*
    Low level API
 */

-Sha3_Context :: _sha3.Sha3_Context
+Keccak_Context :: _sha3.Sha3_Context

 init :: proc(ctx: ^_sha3.Sha3_Context) {
    ctx.is_keccak = true
@@ -17,16 +17,18 @@ import "core:io"
    High level API
 */

+DIGEST_SIZE :: 16
+
 // hash_string will hash the given input and return the
 // computed hash
-hash_string :: proc(data: string) -> [16]byte {
+hash_string :: proc(data: string) -> [DIGEST_SIZE]byte {
    return hash_bytes(transmute([]byte)(data))
 }

 // hash_bytes will hash the given input and return the
 // computed hash
-hash_bytes :: proc(data: []byte) -> [16]byte {
-	hash: [16]byte
+hash_bytes :: proc(data: []byte) -> [DIGEST_SIZE]byte {
+	hash: [DIGEST_SIZE]byte
 	ctx: Md2_Context
    // init(&ctx) No-op
    update(&ctx, data)
@@ -34,10 +36,28 @@ hash_bytes :: proc(data: []byte) -> [16]byte {
    return hash
 }

+// hash_string_to_buffer will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer :: proc(data: string, hash: []byte) {
+	hash_bytes_to_buffer(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer :: proc(data, hash: []byte) {
+	assert(len(hash) >= DIGEST_SIZE, "Size of destination buffer is smaller than the digest size")
+    ctx: Md2_Context
+    // init(&ctx) No-op
+    update(&ctx, data)
+    final(&ctx, hash)
+}
+
 // hash_stream will read the stream in chunks and compute a
 // hash from its contents
-hash_stream :: proc(s: io.Stream) -> ([16]byte, bool) {
-	hash: [16]byte
+hash_stream :: proc(s: io.Stream) -> ([DIGEST_SIZE]byte, bool) {
+	hash: [DIGEST_SIZE]byte
 	ctx: Md2_Context
 	// init(&ctx) No-op
 	buf := make([]byte, 512)
@@ -55,7 +75,7 @@ hash_stream :: proc(s: io.Stream) -> ([16]byte, bool) {

 // hash_file will read the file provided by the given handle
 // and compute a hash
-hash_file :: proc(hd: os.Handle, load_at_once := false) -> ([16]byte, bool) {
+hash_file :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE]byte, bool) {
 	if !load_at_once {
        return hash_stream(os.stream_from_handle(hd))
    } else {
@@ -63,7 +83,7 @@ hash_file :: proc(hd: os.Handle, load_at_once := false) -> ([16]byte, bool) {
            return hash_bytes(buf[:]), ok
        }
    }
-    return [16]byte{}, false
+    return [DIGEST_SIZE]byte{}, false
 }

 hash :: proc {
@@ -71,6 +91,8 @@ hash :: proc {
    hash_file,
    hash_bytes,
    hash_string,
+    hash_bytes_to_buffer,
+    hash_string_to_buffer,
 }

 /*
@@ -86,7 +108,7 @@ update :: proc(ctx: ^Md2_Context, data: []byte) {
 	for i := 0; i < len(data); i += 1 {
 		ctx.data[ctx.datalen] = data[i]
 		ctx.datalen += 1
-		if (ctx.datalen == 16) {
+		if (ctx.datalen == DIGEST_SIZE) {
 			transform(ctx, ctx.data[:])
 			ctx.datalen = 0
 		}
@@ -94,14 +116,14 @@ update :: proc(ctx: ^Md2_Context, data: []byte) {
 }

 final :: proc(ctx: ^Md2_Context, hash: []byte) {
-	to_pad := byte(16 - ctx.datalen)
-    for ctx.datalen < 16 {
+	to_pad := byte(DIGEST_SIZE - ctx.datalen)
+    for ctx.datalen < DIGEST_SIZE {
        ctx.data[ctx.datalen] = to_pad
 		ctx.datalen += 1
    }
 	transform(ctx, ctx.data[:])
 	transform(ctx, ctx.checksum[:])
-    for i := 0; i < 16; i += 1 {
+    for i := 0; i < DIGEST_SIZE; i += 1 {
        hash[i] = ctx.state[i]
    }
 }
@@ -111,9 +133,9 @@ final :: proc(ctx: ^Md2_Context, hash: []byte) {
 */

 Md2_Context :: struct {
-    data:     [16]byte,
-    state:    [16 * 3]byte,
-    checksum: [16]byte,
+    data:     [DIGEST_SIZE]byte,
+    state:    [DIGEST_SIZE * 3]byte,
+    checksum: [DIGEST_SIZE]byte,
    datalen:  int,
 }

@@ -140,20 +162,20 @@ PI_TABLE := [?]byte {

 transform :: proc(ctx: ^Md2_Context, data: []byte) {
    j,k,t: byte
-	for j = 0; j < 16; j += 1 {
-		ctx.state[j + 16] = data[j]
-		ctx.state[j + 16 * 2] = (ctx.state[j + 16] ~ ctx.state[j])
+	for j = 0; j < DIGEST_SIZE; j += 1 {
+		ctx.state[j + DIGEST_SIZE] = data[j]
+		ctx.state[j + DIGEST_SIZE * 2] = (ctx.state[j + DIGEST_SIZE] ~ ctx.state[j])
 	}
 	t = 0
-	for j = 0; j < 16 + 2; j += 1 {
-		for k = 0; k < 16 * 3; k += 1 {
+	for j = 0; j < DIGEST_SIZE + 2; j += 1 {
+		for k = 0; k < DIGEST_SIZE * 3; k += 1 {
 			ctx.state[k] ~= PI_TABLE[t]
 			t = ctx.state[k]
 		}
 		t = (t + j) & 0xff
 	}
-	t = ctx.checksum[16 - 1]
-	for j = 0; j < 16; j += 1 {
+	t = ctx.checksum[DIGEST_SIZE - 1]
+	for j = 0; j < DIGEST_SIZE; j += 1 {
 		ctx.checksum[j] ~= PI_TABLE[data[j] ~ t]
 		t = ctx.checksum[j]
 	}
@@ -21,16 +21,18 @@ import "../util"
    High level API
 */

+DIGEST_SIZE :: 16
+
 // hash_string will hash the given input and return the
 // computed hash
-hash_string :: proc(data: string) -> [16]byte {
+hash_string :: proc(data: string) -> [DIGEST_SIZE]byte {
    return hash_bytes(transmute([]byte)(data))
 }

 // hash_bytes will hash the given input and return the
 // computed hash
-hash_bytes :: proc(data: []byte) -> [16]byte {
-    hash: [16]byte
+hash_bytes :: proc(data: []byte) -> [DIGEST_SIZE]byte {
+    hash: [DIGEST_SIZE]byte
    ctx: Md4_Context
    init(&ctx)
    update(&ctx, data)
@@ -38,10 +40,28 @@ hash_bytes :: proc(data: []byte) -> [16]byte {
    return hash
 }

+// hash_string_to_buffer will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE, "Size of destination buffer is smaller than the digest size")
+    ctx: Md4_Context
+    init(&ctx)
+    update(&ctx, data)
+    final(&ctx, hash)
+}
+
 // hash_stream will read the stream in chunks and compute a
 // hash from its contents
-hash_stream :: proc(s: io.Stream) -> ([16]byte, bool) {
-    hash: [16]byte
+hash_stream :: proc(s: io.Stream) -> ([DIGEST_SIZE]byte, bool) {
+    hash: [DIGEST_SIZE]byte
    ctx: Md4_Context
    init(&ctx)
    buf := make([]byte, 512)
@@ -59,7 +79,7 @@ hash_stream :: proc(s: io.Stream) -> ([16]byte, bool) {

 // hash_file will read the file provided by the given handle
 // and compute a hash
-hash_file :: proc(hd: os.Handle, load_at_once := false) -> ([16]byte, bool) {
+hash_file :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE]byte, bool) {
    if !load_at_once {
        return hash_stream(os.stream_from_handle(hd))
    } else {
@@ -67,7 +87,7 @@ hash_file :: proc(hd: os.Handle, load_at_once := false) -> ([16]byte, bool) {
            return hash_bytes(buf[:]), ok
        }
    }
-    return [16]byte{}, false
+    return [DIGEST_SIZE]byte{}, false
 }

 hash :: proc {
@@ -75,6 +95,8 @@ hash :: proc {
    hash_file,
    hash_bytes,
    hash_string,
+    hash_bytes_to_buffer,
+    hash_string_to_buffer,
 }

 /*
@@ -171,9 +193,9 @@ HH :: #force_inline proc "contextless"(a, b, c, d, x: u32, s : int) -> u32 {

 transform :: proc(ctx: ^Md4_Context, data: []byte) {
    a, b, c, d, i, j: u32
-    m: [16]u32
+    m: [DIGEST_SIZE]u32

-    for i, j = 0, 0; i < 16; i += 1 {
+    for i, j = 0, 0; i < DIGEST_SIZE; i += 1 {
        m[i] = u32(data[j]) | (u32(data[j + 1]) << 8) | (u32(data[j + 2]) << 16) | (u32(data[j + 3]) << 24)
        j += 4
    }
@@ -20,16 +20,18 @@ import "../util"
    High level API
 */

+DIGEST_SIZE :: 16
+
 // hash_string will hash the given input and return the
 // computed hash
-hash_string :: proc(data: string) -> [16]byte {
+hash_string :: proc(data: string) -> [DIGEST_SIZE]byte {
    return hash_bytes(transmute([]byte)(data))
 }

 // hash_bytes will hash the given input and return the
 // computed hash
-hash_bytes :: proc(data: []byte) -> [16]byte {
-    hash: [16]byte
+hash_bytes :: proc(data: []byte) -> [DIGEST_SIZE]byte {
+    hash: [DIGEST_SIZE]byte
    ctx: Md5_Context
    init(&ctx)
    update(&ctx, data)
@@ -37,10 +39,28 @@ hash_bytes :: proc(data: []byte) -> [16]byte {
    return hash
 }

+// hash_string_to_buffer will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE, "Size of destination buffer is smaller than the digest size")
+    ctx: Md5_Context
+    init(&ctx)
+    update(&ctx, data)
+    final(&ctx, hash)
+}
+
 // hash_stream will read the stream in chunks and compute a
 // hash from its contents
-hash_stream :: proc(s: io.Stream) -> ([16]byte, bool) {
-    hash: [16]byte
+hash_stream :: proc(s: io.Stream) -> ([DIGEST_SIZE]byte, bool) {
+    hash: [DIGEST_SIZE]byte
    ctx: Md5_Context
    init(&ctx)
    buf := make([]byte, 512)
@@ -58,7 +78,7 @@ hash_stream :: proc(s: io.Stream) -> ([16]byte, bool) {

 // hash_file will read the file provided by the given handle
 // and compute a hash
-hash_file :: proc(hd: os.Handle, load_at_once := false) -> ([16]byte, bool) {
+hash_file :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE]byte, bool) {
    if !load_at_once {
        return hash_stream(os.stream_from_handle(hd))
    } else {
@@ -66,7 +86,7 @@ hash_file :: proc(hd: os.Handle, load_at_once := false) -> ([16]byte, bool) {
            return hash_bytes(buf[:]), ok
        }
    }
-    return [16]byte{}, false
+    return [DIGEST_SIZE]byte{}, false
 }

 hash :: proc {
@@ -74,6 +94,8 @@ hash :: proc {
    hash_file,
    hash_bytes,
    hash_string,
+    hash_bytes_to_buffer,
+    hash_string_to_buffer,
 }

 /*
@@ -176,9 +198,9 @@ II :: #force_inline proc "contextless" (a, b, c, d, m: u32, s: int, t: u32) -> u

 transform :: proc(ctx: ^Md5_Context, data: []byte) {
    i, j: u32
-    m: [16]u32
+    m: [DIGEST_SIZE]u32

-    for i, j = 0, 0; i < 16; i+=1 {
+    for i, j = 0, 0; i < DIGEST_SIZE; i+=1 {
        m[i] = u32(data[j]) + u32(data[j + 1]) << 8 + u32(data[j + 2]) << 16 + u32(data[j + 3]) << 24
        j += 4
    }
@@ -1,6 +1,6 @@
 package crypto

-when ODIN_OS != "linux" {
+when ODIN_OS != .Linux && ODIN_OS != .OpenBSD && ODIN_OS != .Windows {
 	_rand_bytes :: proc (dst: []byte) {
 		unimplemented("crypto: rand_bytes not supported on this OS")
 	}
@@ -0,0 +1,12 @@
+package crypto
+
+import "core:c"
+
+foreign import libc "system:c"
+foreign libc {
+	arc4random_buf :: proc "c" (buf: rawptr, nbytes: c.size_t) ---
+}
+
+_rand_bytes :: proc (dst: []byte) {
+	arc4random_buf(raw_data(dst), len(dst))
+}
@@ -0,0 +1,23 @@
+package crypto
+
+import win32 "core:sys/windows"
+import "core:os"
+import "core:fmt"
+
+_rand_bytes :: proc(dst: []byte) {
+	ret := (os.Errno)(win32.BCryptGenRandom(nil, raw_data(dst), u32(len(dst)), win32.BCRYPT_USE_SYSTEM_PREFERRED_RNG))
+	if ret != os.ERROR_NONE {
+		switch ret {
+			case os.ERROR_INVALID_HANDLE:
+				// The handle to the first parameter is invalid.
+				// This should not happen here, since we explicitly pass nil to it
+				panic("crypto: BCryptGenRandom Invalid handle for hAlgorithm")
+			case os.ERROR_INVALID_PARAMETER:
+				// One of the parameters was invalid
+				panic("crypto: BCryptGenRandom Invalid parameter")
+			case:
+				// Unknown error
+				panic(fmt.tprintf("crypto: BCryptGenRandom failed: %d\n", ret))
+		}
+	}
+}
@@ -19,16 +19,21 @@ import "../util"
    High level API
 */

+DIGEST_SIZE_128 :: 16
+DIGEST_SIZE_160 :: 20
+DIGEST_SIZE_256 :: 32
+DIGEST_SIZE_320 :: 40
+
 // hash_string_128 will hash the given input and return the
 // computed hash
-hash_string_128 :: proc(data: string) -> [16]byte {
+hash_string_128 :: proc(data: string) -> [DIGEST_SIZE_128]byte {
    return hash_bytes_128(transmute([]byte)(data))
 }

 // hash_bytes_128 will hash the given input and return the
 // computed hash
-hash_bytes_128 :: proc(data: []byte) -> [16]byte {
-    hash: [16]byte
+hash_bytes_128 :: proc(data: []byte) -> [DIGEST_SIZE_128]byte {
+    hash: [DIGEST_SIZE_128]byte
    ctx: Ripemd128_Context
    init(&ctx)
    update(&ctx, data)
@@ -36,10 +41,28 @@ hash_bytes_128 :: proc(data: []byte) -> [16]byte {
    return hash
 }

+// hash_string_to_buffer_128 will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer_128 :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer_128(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer_128 will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer_128 :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE_128, "Size of destination buffer is smaller than the digest size")
+    ctx: Ripemd128_Context
+    init(&ctx)
+    update(&ctx, data)
+    final(&ctx, hash)
+}
+
 // hash_stream_128 will read the stream in chunks and compute a
 // hash from its contents
-hash_stream_128 :: proc(s: io.Stream) -> ([16]byte, bool) {
-    hash: [16]byte
+hash_stream_128 :: proc(s: io.Stream) -> ([DIGEST_SIZE_128]byte, bool) {
+    hash: [DIGEST_SIZE_128]byte
    ctx: Ripemd128_Context
    init(&ctx)
    buf := make([]byte, 512)
@@ -57,7 +80,7 @@ hash_stream_128 :: proc(s: io.Stream) -> ([16]byte, bool) {

 // hash_file_128 will read the file provided by the given handle
 // and compute a hash
-hash_file_128 :: proc(hd: os.Handle, load_at_once := false) -> ([16]byte, bool) {
+hash_file_128 :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE_128]byte, bool) {
    if !load_at_once {
        return hash_stream_128(os.stream_from_handle(hd))
    } else {
@@ -65,7 +88,7 @@ hash_file_128 :: proc(hd: os.Handle, load_at_once := false) -> ([16]byte, bool)
            return hash_bytes_128(buf[:]), ok
        }
    }
-    return [16]byte{}, false
+    return [DIGEST_SIZE_128]byte{}, false
 }

 hash_128 :: proc {
@@ -73,18 +96,20 @@ hash_128 :: proc {
    hash_file_128,
    hash_bytes_128,
    hash_string_128,
+    hash_bytes_to_buffer_128,
+    hash_string_to_buffer_128,
 }

 // hash_string_160 will hash the given input and return the
 // computed hash
-hash_string_160 :: proc(data: string) -> [20]byte {
+hash_string_160 :: proc(data: string) -> [DIGEST_SIZE_160]byte {
    return hash_bytes_160(transmute([]byte)(data))
 }

 // hash_bytes_160 will hash the given input and return the
 // computed hash
-hash_bytes_160 :: proc(data: []byte) -> [20]byte {
-    hash: [20]byte
+hash_bytes_160 :: proc(data: []byte) -> [DIGEST_SIZE_160]byte {
+    hash: [DIGEST_SIZE_160]byte
    ctx: Ripemd160_Context
    init(&ctx)
    update(&ctx, data)
@@ -92,10 +117,28 @@ hash_bytes_160 :: proc(data: []byte) -> [20]byte {
    return hash
 }

+// hash_string_to_buffer_160 will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer_160 :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer_160(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer_160 will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer_160 :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE_160, "Size of destination buffer is smaller than the digest size")
+    ctx: Ripemd160_Context
+    init(&ctx)
+    update(&ctx, data)
+    final(&ctx, hash)
+}
+
 // hash_stream_160 will read the stream in chunks and compute a
 // hash from its contents
-hash_stream_160 :: proc(s: io.Stream) -> ([20]byte, bool) {
-    hash: [20]byte
+hash_stream_160 :: proc(s: io.Stream) -> ([DIGEST_SIZE_160]byte, bool) {
+    hash: [DIGEST_SIZE_160]byte
    ctx: Ripemd160_Context
    init(&ctx)
    buf := make([]byte, 512)
@@ -113,7 +156,7 @@ hash_stream_160 :: proc(s: io.Stream) -> ([20]byte, bool) {

 // hash_file_160 will read the file provided by the given handle
 // and compute a hash
-hash_file_160 :: proc(hd: os.Handle, load_at_once := false) -> ([20]byte, bool) {
+hash_file_160 :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE_160]byte, bool) {
    if !load_at_once {
        return hash_stream_160(os.stream_from_handle(hd))
    } else {
@@ -121,7 +164,7 @@ hash_file_160 :: proc(hd: os.Handle, load_at_once := false) -> ([20]byte, bool)
            return hash_bytes_160(buf[:]), ok
        }
    }
-    return [20]byte{}, false
+    return [DIGEST_SIZE_160]byte{}, false
 }

 hash_160 :: proc {
@@ -129,18 +172,20 @@ hash_160 :: proc {
    hash_file_160,
    hash_bytes_160,
    hash_string_160,
+    hash_bytes_to_buffer_160,
+    hash_string_to_buffer_160,
 }

 // hash_string_256 will hash the given input and return the
 // computed hash
-hash_string_256 :: proc(data: string) -> [32]byte {
+hash_string_256 :: proc(data: string) -> [DIGEST_SIZE_256]byte {
    return hash_bytes_256(transmute([]byte)(data))
 }

 // hash_bytes_256 will hash the given input and return the
 // computed hash
-hash_bytes_256 :: proc(data: []byte) -> [32]byte {
-    hash: [32]byte
+hash_bytes_256 :: proc(data: []byte) -> [DIGEST_SIZE_256]byte {
+    hash: [DIGEST_SIZE_256]byte
    ctx: Ripemd256_Context
    init(&ctx)
    update(&ctx, data)
@@ -148,10 +193,28 @@ hash_bytes_256 :: proc(data: []byte) -> [32]byte {
    return hash
 }

+// hash_string_to_buffer_256 will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer_256 :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer_256(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer_256 will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer_256 :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE_256, "Size of destination buffer is smaller than the digest size")
+    ctx: Ripemd256_Context
+    init(&ctx)
+    update(&ctx, data)
+    final(&ctx, hash)
+}
+
 // hash_stream_256 will read the stream in chunks and compute a
 // hash from its contents
-hash_stream_256 :: proc(s: io.Stream) -> ([32]byte, bool) {
-    hash: [32]byte
+hash_stream_256 :: proc(s: io.Stream) -> ([DIGEST_SIZE_256]byte, bool) {
+    hash: [DIGEST_SIZE_256]byte
    ctx: Ripemd256_Context
    init(&ctx)
    buf := make([]byte, 512)
@@ -169,7 +232,7 @@ hash_stream_256 :: proc(s: io.Stream) -> ([32]byte, bool) {

 // hash_file_256 will read the file provided by the given handle
 // and compute a hash
-hash_file_256 :: proc(hd: os.Handle, load_at_once := false) -> ([32]byte, bool) {
+hash_file_256 :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE_256]byte, bool) {
    if !load_at_once {
        return hash_stream_256(os.stream_from_handle(hd))
    } else {
@@ -177,7 +240,7 @@ hash_file_256 :: proc(hd: os.Handle, load_at_once := false) -> ([32]byte, bool)
            return hash_bytes_256(buf[:]), ok
        }
    }
-    return [32]byte{}, false
+    return [DIGEST_SIZE_256]byte{}, false
 }

 hash_256 :: proc {
@@ -185,18 +248,20 @@ hash_256 :: proc {
    hash_file_256,
    hash_bytes_256,
    hash_string_256,
+    hash_bytes_to_buffer_256,
+    hash_string_to_buffer_256,
 }

 // hash_string_320 will hash the given input and return the
 // computed hash
-hash_string_320 :: proc(data: string) -> [40]byte {
+hash_string_320 :: proc(data: string) -> [DIGEST_SIZE_320]byte {
    return hash_bytes_320(transmute([]byte)(data))
 }

 // hash_bytes_320 will hash the given input and return the
 // computed hash
-hash_bytes_320 :: proc(data: []byte) -> [40]byte {
-    hash: [40]byte
+hash_bytes_320 :: proc(data: []byte) -> [DIGEST_SIZE_320]byte {
+    hash: [DIGEST_SIZE_320]byte
    ctx: Ripemd320_Context
    init(&ctx)
    update(&ctx, data)
@@ -204,10 +269,28 @@ hash_bytes_320 :: proc(data: []byte) -> [40]byte {
    return hash
 }

+// hash_string_to_buffer_320 will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer_320 :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer_320(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer_320 will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer_320 :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE_320, "Size of destination buffer is smaller than the digest size")
+    ctx: Ripemd320_Context
+    init(&ctx)
+    update(&ctx, data)
+    final(&ctx, hash)
+}
+
 // hash_stream_320 will read the stream in chunks and compute a
 // hash from its contents
-hash_stream_320 :: proc(s: io.Stream) -> ([40]byte, bool) {
-    hash: [40]byte
+hash_stream_320 :: proc(s: io.Stream) -> ([DIGEST_SIZE_320]byte, bool) {
+    hash: [DIGEST_SIZE_320]byte
    ctx: Ripemd320_Context
    init(&ctx)
    buf := make([]byte, 512)
@@ -225,7 +308,7 @@ hash_stream_320 :: proc(s: io.Stream) -> ([40]byte, bool) {

 // hash_file_320 will read the file provided by the given handle
 // and compute a hash
-hash_file_320 :: proc(hd: os.Handle, load_at_once := false) -> ([40]byte, bool) {
+hash_file_320 :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE_320]byte, bool) {
    if !load_at_once {
        return hash_stream_320(os.stream_from_handle(hd))
    } else {
@@ -233,7 +316,7 @@ hash_file_320 :: proc(hd: os.Handle, load_at_once := false) -> ([40]byte, bool)
            return hash_bytes_320(buf[:]), ok
        }
    }
-    return [40]byte{}, false
+    return [DIGEST_SIZE_320]byte{}, false
 }

 hash_320 :: proc {
@@ -241,6 +324,8 @@ hash_320 :: proc {
    hash_file_320,
    hash_bytes_320,
    hash_string_320,
+    hash_bytes_to_buffer_320,
+    hash_string_to_buffer_320,
 }

 /*
@@ -19,16 +19,19 @@ import "../util"
 /*
    High level API
 */
+
+DIGEST_SIZE :: 20
+
 // hash_string will hash the given input and return the
 // computed hash
-hash_string :: proc(data: string) -> [20]byte {
+hash_string :: proc(data: string) -> [DIGEST_SIZE]byte {
    return hash_bytes(transmute([]byte)(data))
 }

 // hash_bytes will hash the given input and return the
 // computed hash
-hash_bytes :: proc(data: []byte) -> [20]byte {
-    hash: [20]byte
+hash_bytes :: proc(data: []byte) -> [DIGEST_SIZE]byte {
+    hash: [DIGEST_SIZE]byte
    ctx: Sha1_Context
    init(&ctx)
    update(&ctx, data)
@@ -36,10 +39,28 @@ hash_bytes :: proc(data: []byte) -> [20]byte {
    return hash
 }

+// hash_string_to_buffer will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE, "Size of destination buffer is smaller than the digest size")
+    ctx: Sha1_Context
+    init(&ctx)
+    update(&ctx, data)
+    final(&ctx, hash)
+}
+
 // hash_stream will read the stream in chunks and compute a
 // hash from its contents
-hash_stream :: proc(s: io.Stream) -> ([20]byte, bool) {
-    hash: [20]byte
+hash_stream :: proc(s: io.Stream) -> ([DIGEST_SIZE]byte, bool) {
+    hash: [DIGEST_SIZE]byte
    ctx: Sha1_Context
    init(&ctx)
    buf := make([]byte, 512)
@@ -57,7 +78,7 @@ hash_stream :: proc(s: io.Stream) -> ([20]byte, bool) {

 // hash_file will read the file provided by the given handle
 // and compute a hash
-hash_file :: proc(hd: os.Handle, load_at_once := false) -> ([20]byte, bool) {
+hash_file :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE]byte, bool) {
    if !load_at_once {
        return hash_stream(os.stream_from_handle(hd))
    } else {
@@ -65,7 +86,7 @@ hash_file :: proc(hd: os.Handle, load_at_once := false) -> ([20]byte, bool) {
            return hash_bytes(buf[:]), ok
        }
    }
-    return [20]byte{}, false
+    return [DIGEST_SIZE]byte{}, false
 }

 hash :: proc {
@@ -73,6 +94,8 @@ hash :: proc {
    hash_file,
    hash_bytes,
    hash_string,
+    hash_bytes_to_buffer,
+    hash_string_to_buffer,
 }

 /*
@@ -21,16 +21,21 @@ import "../util"
    High level API
 */

+DIGEST_SIZE_224 :: 28
+DIGEST_SIZE_256 :: 32
+DIGEST_SIZE_384 :: 48
+DIGEST_SIZE_512 :: 64
+
 // hash_string_224 will hash the given input and return the
 // computed hash
-hash_string_224 :: proc(data: string) -> [28]byte {
+hash_string_224 :: proc(data: string) -> [DIGEST_SIZE_224]byte {
    return hash_bytes_224(transmute([]byte)(data))
 }

 // hash_bytes_224 will hash the given input and return the
 // computed hash
-hash_bytes_224 :: proc(data: []byte) -> [28]byte {
-    hash: [28]byte
+hash_bytes_224 :: proc(data: []byte) -> [DIGEST_SIZE_224]byte {
+    hash: [DIGEST_SIZE_224]byte
 	ctx: Sha256_Context
    ctx.is224 = true
    init(&ctx)
@@ -39,10 +44,29 @@ hash_bytes_224 :: proc(data: []byte) -> [28]byte {
    return hash
 }

+// hash_string_to_buffer_224 will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer_224 :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer_224(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer_224 will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer_224 :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE_224, "Size of destination buffer is smaller than the digest size")
+    ctx: Sha256_Context
+    ctx.is224 = true
+    init(&ctx)
+    update(&ctx, data)
+    final(&ctx, hash)
+}
+
 // hash_stream_224 will read the stream in chunks and compute a
 // hash from its contents
-hash_stream_224 :: proc(s: io.Stream) -> ([28]byte, bool) {
-	hash: [28]byte
+hash_stream_224 :: proc(s: io.Stream) -> ([DIGEST_SIZE_224]byte, bool) {
+	hash: [DIGEST_SIZE_224]byte
    ctx: Sha512_Context
    ctx.is384 = false
    init(&ctx)
@@ -61,7 +85,7 @@ hash_stream_224 :: proc(s: io.Stream) -> ([28]byte, bool) {

 // hash_file_224 will read the file provided by the given handle
 // and compute a hash
-hash_file_224 :: proc(hd: os.Handle, load_at_once := false) -> ([28]byte, bool) {
+hash_file_224 :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE_224]byte, bool) {
    if !load_at_once {
        return hash_stream_224(os.stream_from_handle(hd))
    } else {
@@ -69,7 +93,7 @@ hash_file_224 :: proc(hd: os.Handle, load_at_once := false) -> ([28]byte, bool)
            return hash_bytes_224(buf[:]), ok
        }
    }
-    return [28]byte{}, false
+    return [DIGEST_SIZE_224]byte{}, false
 }

 hash_224 :: proc {
@@ -77,18 +101,20 @@ hash_224 :: proc {
    hash_file_224,
    hash_bytes_224,
    hash_string_224,
+    hash_bytes_to_buffer_224,
+    hash_string_to_buffer_224,
 }

 // hash_string_256 will hash the given input and return the
 // computed hash
-hash_string_256 :: proc(data: string) -> [32]byte {
+hash_string_256 :: proc(data: string) -> [DIGEST_SIZE_256]byte {
    return hash_bytes_256(transmute([]byte)(data))
 }

 // hash_bytes_256 will hash the given input and return the
 // computed hash
-hash_bytes_256 :: proc(data: []byte) -> [32]byte {
-    hash: [32]byte
+hash_bytes_256 :: proc(data: []byte) -> [DIGEST_SIZE_256]byte {
+    hash: [DIGEST_SIZE_256]byte
 	ctx: Sha256_Context
    ctx.is224 = false
    init(&ctx)
@@ -97,10 +123,29 @@ hash_bytes_256 :: proc(data: []byte) -> [32]byte {
    return hash
 }

+// hash_string_to_buffer_256 will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer_256 :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer_256(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer_256 will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer_256 :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE_256, "Size of destination buffer is smaller than the digest size")
+    ctx: Sha256_Context
+    ctx.is224 = false
+    init(&ctx)
+    update(&ctx, data)
+    final(&ctx, hash)
+}
+
 // hash_stream_256 will read the stream in chunks and compute a
 // hash from its contents
-hash_stream_256 :: proc(s: io.Stream) -> ([32]byte, bool) {
-	hash: [32]byte
+hash_stream_256 :: proc(s: io.Stream) -> ([DIGEST_SIZE_256]byte, bool) {
+	hash: [DIGEST_SIZE_256]byte
    ctx: Sha512_Context
    ctx.is384 = false
    init(&ctx)
@@ -119,7 +164,7 @@ hash_stream_256 :: proc(s: io.Stream) -> ([32]byte, bool) {

 // hash_file_256 will read the file provided by the given handle
 // and compute a hash
-hash_file_256 :: proc(hd: os.Handle, load_at_once := false) -> ([32]byte, bool) {
+hash_file_256 :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE_256]byte, bool) {
    if !load_at_once {
        return hash_stream_256(os.stream_from_handle(hd))
    } else {
@@ -127,7 +172,7 @@ hash_file_256 :: proc(hd: os.Handle, load_at_once := false) -> ([32]byte, bool)
            return hash_bytes_256(buf[:]), ok
        }
    }
-    return [32]byte{}, false
+    return [DIGEST_SIZE_256]byte{}, false
 }

 hash_256 :: proc {
@@ -135,18 +180,20 @@ hash_256 :: proc {
    hash_file_256,
    hash_bytes_256,
    hash_string_256,
+    hash_bytes_to_buffer_256,
+    hash_string_to_buffer_256,
 }

 // hash_string_384 will hash the given input and return the
 // computed hash
-hash_string_384 :: proc(data: string) -> [48]byte {
+hash_string_384 :: proc(data: string) -> [DIGEST_SIZE_384]byte {
    return hash_bytes_384(transmute([]byte)(data))
 }

 // hash_bytes_384 will hash the given input and return the
 // computed hash
-hash_bytes_384 :: proc(data: []byte) -> [48]byte {
-    hash: [48]byte
+hash_bytes_384 :: proc(data: []byte) -> [DIGEST_SIZE_384]byte {
+    hash: [DIGEST_SIZE_384]byte
 	ctx: Sha512_Context
    ctx.is384 = true
    init(&ctx)
@@ -155,10 +202,29 @@ hash_bytes_384 :: proc(data: []byte) -> [48]byte {
    return hash
 }

+// hash_string_to_buffer_384 will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer_384 :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer_384(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer_384 will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer_384 :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE_384, "Size of destination buffer is smaller than the digest size")
+    ctx: Sha512_Context
+    ctx.is384 = true
+    init(&ctx)
+    update(&ctx, data)
+    final(&ctx, hash)
+}
+
 // hash_stream_384 will read the stream in chunks and compute a
 // hash from its contents
-hash_stream_384 :: proc(s: io.Stream) -> ([48]byte, bool) {
-	hash: [48]byte
+hash_stream_384 :: proc(s: io.Stream) -> ([DIGEST_SIZE_384]byte, bool) {
+	hash: [DIGEST_SIZE_384]byte
    ctx: Sha512_Context
    ctx.is384 = true
    init(&ctx)
@@ -177,7 +243,7 @@ hash_stream_384 :: proc(s: io.Stream) -> ([48]byte, bool) {

 // hash_file_384 will read the file provided by the given handle
 // and compute a hash
-hash_file_384 :: proc(hd: os.Handle, load_at_once := false) -> ([48]byte, bool) {
+hash_file_384 :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE_384]byte, bool) {
    if !load_at_once {
        return hash_stream_384(os.stream_from_handle(hd))
    } else {
@@ -185,7 +251,7 @@ hash_file_384 :: proc(hd: os.Handle, load_at_once := false) -> ([48]byte, bool)
            return hash_bytes_384(buf[:]), ok
        }
    }
-    return [48]byte{}, false
+    return [DIGEST_SIZE_384]byte{}, false
 }

 hash_384 :: proc {
@@ -193,18 +259,20 @@ hash_384 :: proc {
    hash_file_384,
    hash_bytes_384,
    hash_string_384,
+    hash_bytes_to_buffer_384,
+    hash_string_to_buffer_384,
 }

 // hash_string_512 will hash the given input and return the
 // computed hash
-hash_string_512 :: proc(data: string) -> [64]byte {
+hash_string_512 :: proc(data: string) -> [DIGEST_SIZE_512]byte {
    return hash_bytes_512(transmute([]byte)(data))
 }

 // hash_bytes_512 will hash the given input and return the
 // computed hash
-hash_bytes_512 :: proc(data: []byte) -> [64]byte {
-    hash: [64]byte
+hash_bytes_512 :: proc(data: []byte) -> [DIGEST_SIZE_512]byte {
+    hash: [DIGEST_SIZE_512]byte
 	ctx: Sha512_Context
    ctx.is384 = false
    init(&ctx)
@@ -213,10 +281,29 @@ hash_bytes_512 :: proc(data: []byte) -> [64]byte {
    return hash
 }

+// hash_string_to_buffer_512 will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer_512 :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer_512(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer_512 will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer_512 :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE_512, "Size of destination buffer is smaller than the digest size")
+    ctx: Sha512_Context
+    ctx.is384 = false
+    init(&ctx)
+    update(&ctx, data)
+    final(&ctx, hash)
+}
+
 // hash_stream_512 will read the stream in chunks and compute a
 // hash from its contents
-hash_stream_512 :: proc(s: io.Stream) -> ([64]byte, bool) {
-    hash: [64]byte
+hash_stream_512 :: proc(s: io.Stream) -> ([DIGEST_SIZE_512]byte, bool) {
+    hash: [DIGEST_SIZE_512]byte
    ctx: Sha512_Context
    ctx.is384 = false
 	init(&ctx)
@@ -235,7 +322,7 @@ hash_stream_512 :: proc(s: io.Stream) -> ([64]byte, bool) {

 // hash_file_512 will read the file provided by the given handle
 // and compute a hash
-hash_file_512 :: proc(hd: os.Handle, load_at_once := false) -> ([64]byte, bool) {
+hash_file_512 :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE_512]byte, bool) {
    if !load_at_once {
        return hash_stream_512(os.stream_from_handle(hd))
    } else {
@@ -243,7 +330,7 @@ hash_file_512 :: proc(hd: os.Handle, load_at_once := false) -> ([64]byte, bool)
            return hash_bytes_512(buf[:]), ok
        }
    }
-    return [64]byte{}, false
+    return [DIGEST_SIZE_512]byte{}, false
 }

 hash_512 :: proc {
@@ -251,6 +338,8 @@ hash_512 :: proc {
    hash_file_512,
    hash_bytes_512,
    hash_string_512,
+    hash_bytes_to_buffer_512,
+    hash_string_to_buffer_512,
 }

 /*
@@ -330,8 +419,10 @@ update :: proc(ctx: ^$T, data: []byte) {
    sha2_transf(ctx, shifted_message, block_nb)

    rem_len = new_len % CURR_BLOCK_SIZE
-    when T == Sha256_Context      {copy(ctx.block[:], shifted_message[block_nb << 6:rem_len])}
-    else when T == Sha512_Context {copy(ctx.block[:], shifted_message[block_nb << 7:rem_len])}
+    if rem_len > 0 {
+        when T == Sha256_Context      {copy(ctx.block[:], shifted_message[block_nb << 6:rem_len])} 
+        else when T == Sha512_Context {copy(ctx.block[:], shifted_message[block_nb << 7:rem_len])}
+    }

    ctx.length = rem_len
    when T == Sha256_Context      {ctx.tot_len += (block_nb + 1) << 6}
@@ -20,30 +20,54 @@ import "../_sha3"
    High level API
 */

+DIGEST_SIZE_224 :: 28
+DIGEST_SIZE_256 :: 32
+DIGEST_SIZE_384 :: 48
+DIGEST_SIZE_512 :: 64
+
 // hash_string_224 will hash the given input and return the
 // computed hash
-hash_string_224 :: proc(data: string) -> [28]byte {
+hash_string_224 :: proc(data: string) -> [DIGEST_SIZE_224]byte {
    return hash_bytes_224(transmute([]byte)(data))
 }

 // hash_bytes_224 will hash the given input and return the
 // computed hash
-hash_bytes_224 :: proc(data: []byte) -> [28]byte {
-    hash: [28]byte
+hash_bytes_224 :: proc(data: []byte) -> [DIGEST_SIZE_224]byte {
+    hash: [DIGEST_SIZE_224]byte
    ctx: _sha3.Sha3_Context
-    ctx.mdlen = 28
+    ctx.mdlen = DIGEST_SIZE_224
    _sha3.init(&ctx)
    _sha3.update(&ctx, data)
    _sha3.final(&ctx, hash[:])
    return hash
 }

+// hash_string_to_buffer_224 will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer_224 :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer_224(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer_224 will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer_224 :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE_224, "Size of destination buffer is smaller than the digest size")
+    ctx: _sha3.Sha3_Context
+    ctx.mdlen = DIGEST_SIZE_224
+    _sha3.init(&ctx)
+    _sha3.update(&ctx, data)
+    _sha3.final(&ctx, hash)
+}
+
 // hash_stream_224 will read the stream in chunks and compute a
 // hash from its contents
-hash_stream_224 :: proc(s: io.Stream) -> ([28]byte, bool) {
-    hash: [28]byte
+hash_stream_224 :: proc(s: io.Stream) -> ([DIGEST_SIZE_224]byte, bool) {
+    hash: [DIGEST_SIZE_224]byte
    ctx: _sha3.Sha3_Context
-    ctx.mdlen = 28
+    ctx.mdlen = DIGEST_SIZE_224
    _sha3.init(&ctx)
    buf := make([]byte, 512)
    defer delete(buf)
@@ -60,7 +84,7 @@ hash_stream_224 :: proc(s: io.Stream) -> ([28]byte, bool) {

 // hash_file_224 will read the file provided by the given handle
 // and compute a hash
-hash_file_224 :: proc(hd: os.Handle, load_at_once := false) -> ([28]byte, bool) {
+hash_file_224 :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE_224]byte, bool) {
    if !load_at_once {
        return hash_stream_224(os.stream_from_handle(hd))
    } else {
@@ -68,7 +92,7 @@ hash_file_224 :: proc(hd: os.Handle, load_at_once := false) -> ([28]byte, bool)
            return hash_bytes_224(buf[:]), ok
        }
    }
-    return [28]byte{}, false
+    return [DIGEST_SIZE_224]byte{}, false
 }

 hash_224 :: proc {
@@ -76,32 +100,53 @@ hash_224 :: proc {
    hash_file_224,
    hash_bytes_224,
    hash_string_224,
+    hash_bytes_to_buffer_224,
+    hash_string_to_buffer_224,
 }

 // hash_string_256 will hash the given input and return the
 // computed hash
-hash_string_256 :: proc(data: string) -> [32]byte {
+hash_string_256 :: proc(data: string) -> [DIGEST_SIZE_256]byte {
    return hash_bytes_256(transmute([]byte)(data))
 }

 // hash_bytes_256 will hash the given input and return the
 // computed hash
-hash_bytes_256 :: proc(data: []byte) -> [32]byte {
-    hash: [32]byte
+hash_bytes_256 :: proc(data: []byte) -> [DIGEST_SIZE_256]byte {
+    hash: [DIGEST_SIZE_256]byte
    ctx: _sha3.Sha3_Context
-    ctx.mdlen = 32
+    ctx.mdlen = DIGEST_SIZE_256
    _sha3.init(&ctx)
    _sha3.update(&ctx, data)
    _sha3.final(&ctx, hash[:])
    return hash
 }

+// hash_string_to_buffer_256 will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer_256 :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer_256(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer_256 will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer_256 :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE_256, "Size of destination buffer is smaller than the digest size")
+    ctx: _sha3.Sha3_Context
+    ctx.mdlen = DIGEST_SIZE_256
+    _sha3.init(&ctx)
+    _sha3.update(&ctx, data)
+    _sha3.final(&ctx, hash)
+}
+
 // hash_stream_256 will read the stream in chunks and compute a
 // hash from its contents
-hash_stream_256 :: proc(s: io.Stream) -> ([32]byte, bool) {
-    hash: [32]byte
+hash_stream_256 :: proc(s: io.Stream) -> ([DIGEST_SIZE_256]byte, bool) {
+    hash: [DIGEST_SIZE_256]byte
    ctx: _sha3.Sha3_Context
-    ctx.mdlen = 32
+    ctx.mdlen = DIGEST_SIZE_256
    _sha3.init(&ctx)
    buf := make([]byte, 512)
    defer delete(buf)
@@ -118,7 +163,7 @@ hash_stream_256 :: proc(s: io.Stream) -> ([32]byte, bool) {

 // hash_file_256 will read the file provided by the given handle
 // and compute a hash
-hash_file_256 :: proc(hd: os.Handle, load_at_once := false) -> ([32]byte, bool) {
+hash_file_256 :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE_256]byte, bool) {
    if !load_at_once {
        return hash_stream_256(os.stream_from_handle(hd))
    } else {
@@ -126,7 +171,7 @@ hash_file_256 :: proc(hd: os.Handle, load_at_once := false) -> ([32]byte, bool)
            return hash_bytes_256(buf[:]), ok
        }
    }
-    return [32]byte{}, false
+    return [DIGEST_SIZE_256]byte{}, false
 }

 hash_256 :: proc {
@@ -134,32 +179,53 @@ hash_256 :: proc {
    hash_file_256,
    hash_bytes_256,
    hash_string_256,
+    hash_bytes_to_buffer_256,
+    hash_string_to_buffer_256,
 }

 // hash_string_384 will hash the given input and return the
 // computed hash
-hash_string_384 :: proc(data: string) -> [48]byte {
+hash_string_384 :: proc(data: string) -> [DIGEST_SIZE_384]byte {
    return hash_bytes_384(transmute([]byte)(data))
 }

 // hash_bytes_384 will hash the given input and return the
 // computed hash
-hash_bytes_384 :: proc(data: []byte) -> [48]byte {
-    hash: [48]byte
+hash_bytes_384 :: proc(data: []byte) -> [DIGEST_SIZE_384]byte {
+    hash: [DIGEST_SIZE_384]byte
    ctx: _sha3.Sha3_Context
-    ctx.mdlen = 48
+    ctx.mdlen = DIGEST_SIZE_384
    _sha3.init(&ctx)
    _sha3.update(&ctx, data)
    _sha3.final(&ctx, hash[:])
    return hash
 }

+// hash_string_to_buffer_384 will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer_384 :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer_384(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer_384 will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer_384 :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE_384, "Size of destination buffer is smaller than the digest size")
+    ctx: _sha3.Sha3_Context
+    ctx.mdlen = DIGEST_SIZE_384
+    _sha3.init(&ctx)
+    _sha3.update(&ctx, data)
+    _sha3.final(&ctx, hash)
+}
+
 // hash_stream_384 will read the stream in chunks and compute a
 // hash from its contents
-hash_stream_384 :: proc(s: io.Stream) -> ([48]byte, bool) {
-    hash: [48]byte
+hash_stream_384 :: proc(s: io.Stream) -> ([DIGEST_SIZE_384]byte, bool) {
+    hash: [DIGEST_SIZE_384]byte
    ctx: _sha3.Sha3_Context
-    ctx.mdlen = 48
+    ctx.mdlen = DIGEST_SIZE_384
    _sha3.init(&ctx)
    buf := make([]byte, 512)
    defer delete(buf)
@@ -176,7 +242,7 @@ hash_stream_384 :: proc(s: io.Stream) -> ([48]byte, bool) {

 // hash_file_384 will read the file provided by the given handle
 // and compute a hash
-hash_file_384 :: proc(hd: os.Handle, load_at_once := false) -> ([48]byte, bool) {
+hash_file_384 :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE_384]byte, bool) {
    if !load_at_once {
        return hash_stream_384(os.stream_from_handle(hd))
    } else {
@@ -184,7 +250,7 @@ hash_file_384 :: proc(hd: os.Handle, load_at_once := false) -> ([48]byte, bool)
            return hash_bytes_384(buf[:]), ok
        }
    }
-    return [48]byte{}, false
+    return [DIGEST_SIZE_384]byte{}, false
 }

 hash_384 :: proc {
@@ -192,32 +258,53 @@ hash_384 :: proc {
    hash_file_384,
    hash_bytes_384,
    hash_string_384,
+    hash_bytes_to_buffer_384,
+    hash_string_to_buffer_384,
 }

 // hash_string_512 will hash the given input and return the
 // computed hash
-hash_string_512 :: proc(data: string) -> [64]byte {
+hash_string_512 :: proc(data: string) -> [DIGEST_SIZE_512]byte {
    return hash_bytes_512(transmute([]byte)(data))
 }

 // hash_bytes_512 will hash the given input and return the
 // computed hash
-hash_bytes_512 :: proc(data: []byte) -> [64]byte {
-    hash: [64]byte
+hash_bytes_512 :: proc(data: []byte) -> [DIGEST_SIZE_512]byte {
+    hash: [DIGEST_SIZE_512]byte
    ctx: _sha3.Sha3_Context
-    ctx.mdlen = 64
+    ctx.mdlen = DIGEST_SIZE_512
    _sha3.init(&ctx)
    _sha3.update(&ctx, data)
    _sha3.final(&ctx, hash[:])
    return hash
 }

+// hash_string_to_buffer_512 will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer_512 :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer_512(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer_512 will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer_512 :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE_512, "Size of destination buffer is smaller than the digest size")
+    ctx: _sha3.Sha3_Context
+    ctx.mdlen = DIGEST_SIZE_512
+    _sha3.init(&ctx)
+    _sha3.update(&ctx, data)
+    _sha3.final(&ctx, hash)
+}
+
 // hash_stream_512 will read the stream in chunks and compute a
 // hash from its contents
-hash_stream_512 :: proc(s: io.Stream) -> ([64]byte, bool) {
-    hash: [64]byte
+hash_stream_512 :: proc(s: io.Stream) -> ([DIGEST_SIZE_512]byte, bool) {
+    hash: [DIGEST_SIZE_512]byte
    ctx: _sha3.Sha3_Context
-    ctx.mdlen = 64
+    ctx.mdlen = DIGEST_SIZE_512
    _sha3.init(&ctx)
    buf := make([]byte, 512)
    defer delete(buf)
@@ -234,7 +321,7 @@ hash_stream_512 :: proc(s: io.Stream) -> ([64]byte, bool) {

 // hash_file_512 will read the file provided by the given handle
 // and compute a hash
-hash_file_512 :: proc(hd: os.Handle, load_at_once := false) -> ([64]byte, bool) {
+hash_file_512 :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE_512]byte, bool) {
    if !load_at_once {
        return hash_stream_512(os.stream_from_handle(hd))
    } else {
@@ -242,7 +329,7 @@ hash_file_512 :: proc(hd: os.Handle, load_at_once := false) -> ([64]byte, bool)
            return hash_bytes_512(buf[:]), ok
        }
    }
-    return [64]byte{}, false
+    return [DIGEST_SIZE_512]byte{}, false
 }

 hash_512 :: proc {
@@ -250,6 +337,8 @@ hash_512 :: proc {
    hash_file_512,
    hash_bytes_512,
    hash_string_512,
+    hash_bytes_to_buffer_512,
+    hash_string_to_buffer_512,
 }

 /*
@@ -20,18 +20,21 @@ import "../_sha3"
    High level API
 */

+DIGEST_SIZE_128 :: 16
+DIGEST_SIZE_256 :: 32
+
 // hash_string_128 will hash the given input and return the
 // computed hash
-hash_string_128 :: proc(data: string) -> [16]byte {
+hash_string_128 :: proc(data: string) -> [DIGEST_SIZE_128]byte {
    return hash_bytes_128(transmute([]byte)(data))
 }

 // hash_bytes_128 will hash the given input and return the
 // computed hash
-hash_bytes_128 :: proc(data: []byte) -> [16]byte {
-    hash: [16]byte
+hash_bytes_128 :: proc(data: []byte) -> [DIGEST_SIZE_128]byte {
+    hash: [DIGEST_SIZE_128]byte
    ctx: _sha3.Sha3_Context
-    ctx.mdlen = 16
+    ctx.mdlen = DIGEST_SIZE_128
    _sha3.init(&ctx)
    _sha3.update(&ctx, data)
    _sha3.shake_xof(&ctx)
@@ -39,12 +42,32 @@ hash_bytes_128 :: proc(data: []byte) -> [16]byte {
    return hash
 }

+// hash_string_to_buffer_128 will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer_128 :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer_128(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer_128 will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer_128 :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE_128, "Size of destination buffer is smaller than the digest size")
+    ctx: _sha3.Sha3_Context
+    ctx.mdlen = DIGEST_SIZE_128
+    _sha3.init(&ctx)
+    _sha3.update(&ctx, data)
+    _sha3.shake_xof(&ctx)
+    _sha3.shake_out(&ctx, hash)
+}
+
 // hash_stream_128 will read the stream in chunks and compute a
 // hash from its contents
-hash_stream_128 :: proc(s: io.Stream) -> ([16]byte, bool) {
-    hash: [16]byte
+hash_stream_128 :: proc(s: io.Stream) -> ([DIGEST_SIZE_128]byte, bool) {
+    hash: [DIGEST_SIZE_128]byte
    ctx: _sha3.Sha3_Context
-    ctx.mdlen = 16
+    ctx.mdlen = DIGEST_SIZE_128
    _sha3.init(&ctx)
    buf := make([]byte, 512)
    defer delete(buf)
@@ -62,7 +85,7 @@ hash_stream_128 :: proc(s: io.Stream) -> ([16]byte, bool) {

 // hash_file_128 will read the file provided by the given handle
 // and compute a hash
-hash_file_128 :: proc(hd: os.Handle, load_at_once := false) -> ([16]byte, bool) {
+hash_file_128 :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE_128]byte, bool) {
    if !load_at_once {
        return hash_stream_128(os.stream_from_handle(hd))
    } else {
@@ -70,7 +93,7 @@ hash_file_128 :: proc(hd: os.Handle, load_at_once := false) -> ([16]byte, bool)
            return hash_bytes_128(buf[:]), ok
        }
    }
-    return [16]byte{}, false
+    return [DIGEST_SIZE_128]byte{}, false
 }

 hash_128 :: proc {
@@ -78,20 +101,22 @@ hash_128 :: proc {
    hash_file_128,
    hash_bytes_128,
    hash_string_128,
+    hash_bytes_to_buffer_128,
+    hash_string_to_buffer_128,
 }

 // hash_string_256 will hash the given input and return the
 // computed hash
-hash_string_256 :: proc(data: string) -> [32]byte {
+hash_string_256 :: proc(data: string) -> [DIGEST_SIZE_256]byte {
    return hash_bytes_256(transmute([]byte)(data))
 }

 // hash_bytes_256 will hash the given input and return the
 // computed hash
-hash_bytes_256 :: proc(data: []byte) -> [32]byte {
-    hash: [32]byte
+hash_bytes_256 :: proc(data: []byte) -> [DIGEST_SIZE_256]byte {
+    hash: [DIGEST_SIZE_256]byte
    ctx: _sha3.Sha3_Context
-    ctx.mdlen = 32
+    ctx.mdlen = DIGEST_SIZE_256
    _sha3.init(&ctx)
    _sha3.update(&ctx, data)
    _sha3.shake_xof(&ctx)
@@ -99,12 +124,32 @@ hash_bytes_256 :: proc(data: []byte) -> [32]byte {
    return hash
 }

+// hash_string_to_buffer_256 will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer_256 :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer_256(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer_256 will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer_256 :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE_256, "Size of destination buffer is smaller than the digest size")
+    ctx: _sha3.Sha3_Context
+    ctx.mdlen = DIGEST_SIZE_256
+    _sha3.init(&ctx)
+    _sha3.update(&ctx, data)
+    _sha3.shake_xof(&ctx)
+    _sha3.shake_out(&ctx, hash)
+}
+
 // hash_stream_256 will read the stream in chunks and compute a
 // hash from its contents
-hash_stream_256 :: proc(s: io.Stream) -> ([32]byte, bool) {
-    hash: [32]byte
+hash_stream_256 :: proc(s: io.Stream) -> ([DIGEST_SIZE_256]byte, bool) {
+    hash: [DIGEST_SIZE_256]byte
    ctx: _sha3.Sha3_Context
-    ctx.mdlen = 32
+    ctx.mdlen = DIGEST_SIZE_256
    _sha3.init(&ctx)
    buf := make([]byte, 512)
    defer delete(buf)
@@ -122,7 +167,7 @@ hash_stream_256 :: proc(s: io.Stream) -> ([32]byte, bool) {

 // hash_file_256 will read the file provided by the given handle
 // and compute a hash
-hash_file_256 :: proc(hd: os.Handle, load_at_once := false) -> ([32]byte, bool) {
+hash_file_256 :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE_256]byte, bool) {
    if !load_at_once {
        return hash_stream_256(os.stream_from_handle(hd))
    } else {
@@ -130,7 +175,7 @@ hash_file_256 :: proc(hd: os.Handle, load_at_once := false) -> ([32]byte, bool)
            return hash_bytes_256(buf[:]), ok
        }
    }
-    return [32]byte{}, false
+    return [DIGEST_SIZE_256]byte{}, false
 }

 hash_256 :: proc {
@@ -138,13 +183,15 @@ hash_256 :: proc {
    hash_file_256,
    hash_bytes_256,
    hash_string_256,
+    hash_bytes_to_buffer_256,
+    hash_string_to_buffer_256,
 }

 /*
    Low level API
 */

-Sha3_Context :: _sha3.Sha3_Context
+Shake_Context :: _sha3.Sha3_Context

 init :: proc(ctx: ^_sha3.Sha3_Context) {
    _sha3.init(ctx)
@@ -0,0 +1,335 @@
+package siphash
+
+/*
+    Copyright 2022 zhibog
+    Made available under the BSD-3 license.
+
+    List of contributors:
+        zhibog:  Initial implementation.
+
+    Implementation of the SipHash hashing algorithm, as defined at <https://github.com/veorq/SipHash> and <https://www.aumasson.jp/siphash/siphash.pdf>
+
+    Use the specific procedures for a certain setup. The generic procdedures will default to Siphash 2-4
+*/
+
+import "core:crypto"
+import "core:crypto/util"
+
+/*
+    High level API
+*/
+
+KEY_SIZE    :: 16
+DIGEST_SIZE :: 8
+
+// sum_string_1_3 will hash the given message with the key and return
+// the computed hash as a u64
+sum_string_1_3 :: proc(msg, key: string) -> u64 {
+    return sum_bytes_1_3(transmute([]byte)(msg), transmute([]byte)(key))
+}
+
+// sum_bytes_1_3 will hash the given message with the key and return
+// the computed hash as a u64
+sum_bytes_1_3 :: proc (msg, key: []byte) -> u64 {
+    ctx: Context
+    hash: u64
+    init(&ctx, key, 1, 3)
+    update(&ctx, msg)
+    final(&ctx, &hash)
+    return hash
+}
+
+// sum_string_to_buffer_1_3 will hash the given message with the key and write
+// the computed hash into the provided destination buffer
+sum_string_to_buffer_1_3 :: proc(msg, key: string, dst: []byte) {
+    sum_bytes_to_buffer_1_3(transmute([]byte)(msg), transmute([]byte)(key), dst)
+}
+
+// sum_bytes_to_buffer_1_3 will hash the given message with the key and write
+// the computed hash into the provided destination buffer
+sum_bytes_to_buffer_1_3 :: proc(msg, key, dst: []byte) {
+    assert(len(dst) >= DIGEST_SIZE, "crypto/siphash: Destination buffer needs to be at least of size 8")
+    hash  := sum_bytes_1_3(msg, key)
+    _collect_output(dst[:], hash)
+}
+
+sum_1_3 :: proc {
+    sum_string_1_3,
+    sum_bytes_1_3,
+    sum_string_to_buffer_1_3,
+    sum_bytes_to_buffer_1_3,
+}
+
+// verify_u64_1_3 will check if the supplied tag matches with the output you 
+// will get from the provided message and key
+verify_u64_1_3 :: proc (tag: u64 msg, key: []byte) -> bool {
+    return sum_bytes_1_3(msg, key) == tag
+}
+
+// verify_bytes will check if the supplied tag matches with the output you 
+// will get from the provided message and key
+verify_bytes_1_3 :: proc (tag, msg, key: []byte) -> bool {
+    derived_tag: [8]byte
+    sum_bytes_to_buffer_1_3(msg, key, derived_tag[:])
+    return crypto.compare_constant_time(derived_tag[:], tag) == 1
+}
+
+verify_1_3 :: proc {
+    verify_bytes_1_3,
+    verify_u64_1_3,
+}
+
+// sum_string_2_4 will hash the given message with the key and return
+// the computed hash as a u64
+sum_string_2_4 :: proc(msg, key: string) -> u64 {
+    return sum_bytes_2_4(transmute([]byte)(msg), transmute([]byte)(key))
+}
+
+// sum_bytes_2_4 will hash the given message with the key and return
+// the computed hash as a u64
+sum_bytes_2_4 :: proc (msg, key: []byte) -> u64 {
+    ctx: Context
+    hash: u64
+    init(&ctx, key, 2, 4)
+    update(&ctx, msg)
+    final(&ctx, &hash)
+    return hash
+}
+
+// sum_string_to_buffer_2_4 will hash the given message with the key and write
+// the computed hash into the provided destination buffer
+sum_string_to_buffer_2_4 :: proc(msg, key: string, dst: []byte) {
+    sum_bytes_to_buffer_2_4(transmute([]byte)(msg), transmute([]byte)(key), dst)
+}
+
+// sum_bytes_to_buffer_2_4 will hash the given message with the key and write
+// the computed hash into the provided destination buffer
+sum_bytes_to_buffer_2_4 :: proc(msg, key, dst: []byte) {
+    assert(len(dst) >= DIGEST_SIZE, "crypto/siphash: Destination buffer needs to be at least of size 8")
+    hash  := sum_bytes_2_4(msg, key)
+    _collect_output(dst[:], hash)
+}
+
+sum_2_4 :: proc {
+    sum_string_2_4,
+    sum_bytes_2_4,
+    sum_string_to_buffer_2_4,
+    sum_bytes_to_buffer_2_4,
+}
+
+sum_string           :: sum_string_2_4
+sum_bytes            :: sum_bytes_2_4
+sum_string_to_buffer :: sum_string_to_buffer_2_4
+sum_bytes_to_buffer  :: sum_bytes_to_buffer_2_4
+sum :: proc {
+    sum_string,
+    sum_bytes,
+    sum_string_to_buffer,
+    sum_bytes_to_buffer,
+}
+
+// verify_u64_2_4 will check if the supplied tag matches with the output you 
+// will get from the provided message and key
+verify_u64_2_4 :: proc (tag: u64 msg, key: []byte) -> bool {
+    return sum_bytes_2_4(msg, key) == tag
+}
+
+// verify_bytes will check if the supplied tag matches with the output you 
+// will get from the provided message and key
+verify_bytes_2_4 :: proc (tag, msg, key: []byte) -> bool {
+    derived_tag: [8]byte
+    sum_bytes_to_buffer_2_4(msg, key, derived_tag[:])
+    return crypto.compare_constant_time(derived_tag[:], tag) == 1
+}
+
+verify_2_4 :: proc {
+    verify_bytes_2_4,
+    verify_u64_2_4,
+}
+
+verify_bytes :: verify_bytes_2_4
+verify_u64   :: verify_u64_2_4
+verify :: proc {
+    verify_bytes,
+    verify_u64,
+}
+
+// sum_string_4_8 will hash the given message with the key and return
+// the computed hash as a u64
+sum_string_4_8 :: proc(msg, key: string) -> u64 {
+    return sum_bytes_4_8(transmute([]byte)(msg), transmute([]byte)(key))
+}
+
+// sum_bytes_4_8 will hash the given message with the key and return
+// the computed hash as a u64
+sum_bytes_4_8 :: proc (msg, key: []byte) -> u64 {
+    ctx: Context
+    hash: u64
+    init(&ctx, key, 4, 8)
+    update(&ctx, msg)
+    final(&ctx, &hash)
+    return hash
+}
+
+// sum_string_to_buffer_4_8 will hash the given message with the key and write
+// the computed hash into the provided destination buffer
+sum_string_to_buffer_4_8 :: proc(msg, key: string, dst: []byte) {
+    sum_bytes_to_buffer_4_8(transmute([]byte)(msg), transmute([]byte)(key), dst)
+}
+
+// sum_bytes_to_buffer_4_8 will hash the given message with the key and write
+// the computed hash into the provided destination buffer
+sum_bytes_to_buffer_4_8 :: proc(msg, key, dst: []byte) {
+    assert(len(dst) >= DIGEST_SIZE, "crypto/siphash: Destination buffer needs to be at least of size 8")
+    hash  := sum_bytes_4_8(msg, key)
+    _collect_output(dst[:], hash)
+}
+
+sum_4_8 :: proc {
+    sum_string_4_8,
+    sum_bytes_4_8,
+    sum_string_to_buffer_4_8,
+    sum_bytes_to_buffer_4_8,
+}
+
+// verify_u64_4_8 will check if the supplied tag matches with the output you 
+// will get from the provided message and key
+verify_u64_4_8 :: proc (tag: u64 msg, key: []byte) -> bool {
+    return sum_bytes_4_8(msg, key) == tag
+}
+
+// verify_bytes will check if the supplied tag matches with the output you 
+// will get from the provided message and key
+verify_bytes_4_8 :: proc (tag, msg, key: []byte) -> bool {
+    derived_tag: [8]byte
+    sum_bytes_to_buffer_4_8(msg, key, derived_tag[:])
+    return crypto.compare_constant_time(derived_tag[:], tag) == 1
+}
+
+verify_4_8 :: proc {
+    verify_bytes_4_8,
+    verify_u64_4_8,
+}
+
+/*
+    Low level API
+*/
+
+init :: proc(ctx: ^Context, key: []byte, c_rounds, d_rounds: int) {
+    assert(len(key) == KEY_SIZE, "crypto/siphash: Invalid key size, want 16")
+    ctx.c_rounds = c_rounds
+    ctx.d_rounds = d_rounds
+    is_valid_setting := (ctx.c_rounds == 1 && ctx.d_rounds == 3) ||
+                        (ctx.c_rounds == 2 && ctx.d_rounds == 4) ||
+                        (ctx.c_rounds == 4 && ctx.d_rounds == 8)
+    assert(is_valid_setting, "crypto/siphash: Incorrect rounds set up. Valid pairs are (1,3), (2,4) and (4,8)")
+    ctx.k0 = util.U64_LE(key[:8])
+    ctx.k1 = util.U64_LE(key[8:])
+    ctx.v0 = 0x736f6d6570736575 ~ ctx.k0
+    ctx.v1 = 0x646f72616e646f6d ~ ctx.k1
+    ctx.v2 = 0x6c7967656e657261 ~ ctx.k0
+    ctx.v3 = 0x7465646279746573 ~ ctx.k1
+    ctx.is_initialized = true
+}
+
+update :: proc(ctx: ^Context, data: []byte) {
+    assert(ctx.is_initialized, "crypto/siphash: Context is not initialized")
+    ctx.last_block = len(data) / 8 * 8
+    ctx.buf = data
+    i := 0
+    m: u64
+    for i < ctx.last_block {
+        m = u64(ctx.buf[i] & 0xff)
+        i += 1
+
+        for r in u64(1)..<8 {
+            m |= u64(ctx.buf[i] & 0xff) << (r * 8)
+            i += 1
+        }
+
+        ctx.v3 ~= m
+        for _ in 0..<ctx.c_rounds {
+            _compress(ctx)
+        }
+
+        ctx.v0 ~= m
+    }
+}
+
+final :: proc(ctx: ^Context, dst: ^u64) {
+    m: u64
+    for i := len(ctx.buf) - 1; i >= ctx.last_block; i -= 1 {
+        m <<= 8
+        m |= u64(ctx.buf[i] & 0xff)
+    }
+    m |= u64(len(ctx.buf) << 56)
+
+    ctx.v3 ~= m
+
+    for _ in 0..<ctx.c_rounds {
+        _compress(ctx)
+    }
+
+    ctx.v0 ~= m
+    ctx.v2 ~= 0xff
+
+    for _ in 0..<ctx.d_rounds {
+        _compress(ctx)
+    }
+
+    dst^ = ctx.v0 ~ ctx.v1 ~ ctx.v2 ~ ctx.v3
+
+    reset(ctx)
+}
+
+reset :: proc(ctx: ^Context) {
+    ctx.k0, ctx.k1 = 0, 0
+    ctx.v0, ctx.v1 = 0, 0
+    ctx.v2, ctx.v3 = 0, 0
+    ctx.last_block = 0
+    ctx.c_rounds = 0
+    ctx.d_rounds = 0
+    ctx.is_initialized = false
+}
+
+Context :: struct {
+    v0, v1, v2, v3: u64,    // State values
+    k0, k1:         u64,    // Split key
+    c_rounds:       int,    // Number of message rounds
+    d_rounds:       int,    // Number of finalization rounds
+    buf:            []byte, // Provided data
+    last_block:     int,    // Offset from the last block
+    is_initialized: bool,
+}
+
+_get_byte :: #force_inline proc "contextless" (byte_num: byte, into: u64) -> byte {
+    return byte(into >> (((~byte_num) & (size_of(u64) - 1)) << 3))
+}
+
+_collect_output :: #force_inline proc "contextless" (dst: []byte, hash: u64) {
+    dst[0] = _get_byte(7, hash)
+    dst[1] = _get_byte(6, hash)
+    dst[2] = _get_byte(5, hash)
+    dst[3] = _get_byte(4, hash)
+    dst[4] = _get_byte(3, hash)
+    dst[5] = _get_byte(2, hash)
+    dst[6] = _get_byte(1, hash)
+    dst[7] = _get_byte(0, hash)
+}
+
+_compress :: #force_inline proc "contextless" (ctx: ^Context) {
+    ctx.v0 += ctx.v1
+    ctx.v1  = util.ROTL64(ctx.v1, 13)
+    ctx.v1 ~= ctx.v0
+    ctx.v0  = util.ROTL64(ctx.v0, 32)
+    ctx.v2 += ctx.v3
+    ctx.v3  = util.ROTL64(ctx.v3, 16)
+    ctx.v3 ~= ctx.v2
+    ctx.v0 += ctx.v3
+    ctx.v3  = util.ROTL64(ctx.v3, 21)
+    ctx.v3 ~= ctx.v0
+    ctx.v2 += ctx.v1
+    ctx.v1  = util.ROTL64(ctx.v1, 17)
+    ctx.v1 ~= ctx.v2
+    ctx.v2  = util.ROTL64(ctx.v2, 32)
+}
@@ -15,16 +15,22 @@ import "core:io"

 import "../util"

+/*
+    High level API
+*/
+
+DIGEST_SIZE :: 32
+
 // hash_string will hash the given input and return the
 // computed hash
-hash_string :: proc(data: string) -> [32]byte {
+hash_string :: proc(data: string) -> [DIGEST_SIZE]byte {
    return hash_bytes(transmute([]byte)(data))
 }

 // hash_bytes will hash the given input and return the
 // computed hash
-hash_bytes :: proc(data: []byte) -> [32]byte {
-    hash: [32]byte
+hash_bytes :: proc(data: []byte) -> [DIGEST_SIZE]byte {
+    hash: [DIGEST_SIZE]byte
    ctx: Sm3_Context
    init(&ctx)
    update(&ctx, data)
@@ -32,10 +38,28 @@ hash_bytes :: proc(data: []byte) -> [32]byte {
    return hash
 }

+// hash_string_to_buffer will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE, "Size of destination buffer is smaller than the digest size")
+    ctx: Sm3_Context
+    init(&ctx)
+    update(&ctx, data)
+    final(&ctx, hash)
+}
+
 // hash_stream will read the stream in chunks and compute a
 // hash from its contents
-hash_stream :: proc(s: io.Stream) -> ([32]byte, bool) {
-    hash: [32]byte
+hash_stream :: proc(s: io.Stream) -> ([DIGEST_SIZE]byte, bool) {
+    hash: [DIGEST_SIZE]byte
    ctx: Sm3_Context
    init(&ctx)
    buf := make([]byte, 512)
@@ -53,7 +77,7 @@ hash_stream :: proc(s: io.Stream) -> ([32]byte, bool) {

 // hash_file will read the file provided by the given handle
 // and compute a hash
-hash_file :: proc(hd: os.Handle, load_at_once := false) -> ([32]byte, bool) {
+hash_file :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE]byte, bool) {
    if !load_at_once {
        return hash_stream(os.stream_from_handle(hd))
    } else {
@@ -61,7 +85,7 @@ hash_file :: proc(hd: os.Handle, load_at_once := false) -> ([32]byte, bool) {
            return hash_bytes(buf[:]), ok
        }
    }
-    return [32]byte{}, false
+    return [DIGEST_SIZE]byte{}, false
 }

 hash :: proc {
@@ -69,6 +93,8 @@ hash :: proc {
    hash_file,
    hash_bytes,
    hash_string,
+    hash_bytes_to_buffer,
+    hash_string_to_buffer,
 }

 /*
@@ -146,9 +172,6 @@ Sm3_Context :: struct {
    length:    u64,
 }

-BLOCK_SIZE_IN_BYTES :: 64
-BLOCK_SIZE_IN_32    :: 16
-
 IV := [8]u32 {
    0x7380166f, 0x4914b2b9, 0x172442d7, 0xda8a0600,
    0xa96f30bc, 0x163138aa, 0xe38dee4d, 0xb0fb0e4e,
@@ -19,16 +19,19 @@ import "../util"
    High level API
 */

+DIGEST_SIZE_256 :: 32
+DIGEST_SIZE_512 :: 64
+
 // hash_string_256 will hash the given input and return the
 // computed hash
-hash_string_256 :: proc(data: string) -> [32]byte {
+hash_string_256 :: proc(data: string) -> [DIGEST_SIZE_256]byte {
    return hash_bytes_256(transmute([]byte)(data))
 }

 // hash_bytes_256 will hash the given input and return the
 // computed hash
-hash_bytes_256 :: proc(data: []byte) -> [32]byte {
-    hash: [32]byte
+hash_bytes_256 :: proc(data: []byte) -> [DIGEST_SIZE_256]byte {
+    hash: [DIGEST_SIZE_256]byte
    ctx: Streebog_Context
    ctx.is256 = true
    init(&ctx)
@@ -37,10 +40,29 @@ hash_bytes_256 :: proc(data: []byte) -> [32]byte {
    return hash
 }

+// hash_string_to_buffer_256 will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer_256 :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer_256(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer_256 will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer_256 :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE_256, "Size of destination buffer is smaller than the digest size")
+    ctx: Streebog_Context
+    ctx.is256 = true
+    init(&ctx)
+    update(&ctx, data)
+    final(&ctx, hash[:])
+}
+
 // hash_stream_256 will read the stream in chunks and compute a
 // hash from its contents
-hash_stream_256 :: proc(s: io.Stream) -> ([32]byte, bool) {
-    hash: [32]byte
+hash_stream_256 :: proc(s: io.Stream) -> ([DIGEST_SIZE_256]byte, bool) {
+    hash: [DIGEST_SIZE_256]byte
    ctx: Streebog_Context
    ctx.is256 = true
    init(&ctx)
@@ -59,7 +81,7 @@ hash_stream_256 :: proc(s: io.Stream) -> ([32]byte, bool) {

 // hash_file_256 will read the file provided by the given handle
 // and compute a hash
-hash_file_256 :: proc(hd: os.Handle, load_at_once := false) -> ([32]byte, bool) {
+hash_file_256 :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE_256]byte, bool) {
    if !load_at_once {
        return hash_stream_256(os.stream_from_handle(hd))
    } else {
@@ -67,7 +89,7 @@ hash_file_256 :: proc(hd: os.Handle, load_at_once := false) -> ([32]byte, bool)
            return hash_bytes_256(buf[:]), ok
        }
    }
-    return [32]byte{}, false
+    return [DIGEST_SIZE_256]byte{}, false
 }

 hash_256 :: proc {
@@ -75,18 +97,20 @@ hash_256 :: proc {
    hash_file_256,
    hash_bytes_256,
    hash_string_256,
+	hash_bytes_to_buffer_256,
+    hash_string_to_buffer_256,
 }

 // hash_string_512 will hash the given input and return the
 // computed hash
-hash_string_512 :: proc(data: string) -> [64]byte {
+hash_string_512 :: proc(data: string) -> [DIGEST_SIZE_512]byte {
    return hash_bytes_512(transmute([]byte)(data))
 }

 // hash_bytes_512 will hash the given input and return the
 // computed hash
-hash_bytes_512 :: proc(data: []byte) -> [64]byte {
-    hash: [64]byte
+hash_bytes_512 :: proc(data: []byte) -> [DIGEST_SIZE_512]byte {
+    hash: [DIGEST_SIZE_512]byte
    ctx: Streebog_Context
    init(&ctx)
    update(&ctx, data)
@@ -94,10 +118,28 @@ hash_bytes_512 :: proc(data: []byte) -> [64]byte {
    return hash
 }

+// hash_string_to_buffer_512 will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer_512 :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer_512(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer_512 will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer_512 :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE_512, "Size of destination buffer is smaller than the digest size")
+    ctx: Streebog_Context
+    init(&ctx)
+    update(&ctx, data)
+    final(&ctx, hash[:])
+}
+
 // hash_stream_512 will read the stream in chunks and compute a
 // hash from its contents
-hash_stream_512 :: proc(s: io.Stream) -> ([64]byte, bool) {
-    hash: [64]byte
+hash_stream_512 :: proc(s: io.Stream) -> ([DIGEST_SIZE_512]byte, bool) {
+    hash: [DIGEST_SIZE_512]byte
    ctx: Streebog_Context
    init(&ctx)
    buf := make([]byte, 512)
@@ -115,7 +157,7 @@ hash_stream_512 :: proc(s: io.Stream) -> ([64]byte, bool) {

 // hash_file_512 will read the file provided by the given handle
 // and compute a hash
-hash_file_512 :: proc(hd: os.Handle, load_at_once := false) -> ([64]byte, bool) {
+hash_file_512 :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE_512]byte, bool) {
    if !load_at_once {
        return hash_stream_512(os.stream_from_handle(hd))
    } else {
@@ -123,7 +165,7 @@ hash_file_512 :: proc(hd: os.Handle, load_at_once := false) -> ([64]byte, bool)
            return hash_bytes_512(buf[:]), ok
        }
    }
-    return [64]byte{}, false
+    return [DIGEST_SIZE_512]byte{}, false
 }

 hash_512 :: proc {
@@ -131,6 +173,8 @@ hash_512 :: proc {
    hash_file_512,
    hash_bytes_512,
    hash_string_512,
+    hash_bytes_to_buffer_512,
+    hash_string_to_buffer_512,
 }

 /*
@@ -19,16 +19,20 @@ import "../_tiger"
    High level API
 */

+DIGEST_SIZE_128 :: 16
+DIGEST_SIZE_160 :: 20
+DIGEST_SIZE_192 :: 24
+
 // hash_string_128 will hash the given input and return the
 // computed hash
-hash_string_128 :: proc(data: string) -> [16]byte {
+hash_string_128 :: proc(data: string) -> [DIGEST_SIZE_128]byte {
    return hash_bytes_128(transmute([]byte)(data))
 }

 // hash_bytes_128 will hash the given input and return the
 // computed hash
-hash_bytes_128 :: proc(data: []byte) -> [16]byte {
-    hash: [16]byte
+hash_bytes_128 :: proc(data: []byte) -> [DIGEST_SIZE_128]byte {
+    hash: [DIGEST_SIZE_128]byte
    ctx: _tiger.Tiger_Context
    ctx.ver = 1
    _tiger.init(&ctx)
@@ -37,10 +41,29 @@ hash_bytes_128 :: proc(data: []byte) -> [16]byte {
    return hash
 }

+// hash_string_to_buffer_128 will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer_128 :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer_128(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer_128 will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer_128 :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE_128, "Size of destination buffer is smaller than the digest size")
+    ctx: _tiger.Tiger_Context
+    ctx.ver = 1
+    _tiger.init(&ctx)
+    _tiger.update(&ctx, data)
+    _tiger.final(&ctx, hash)
+}
+
 // hash_stream_128 will read the stream in chunks and compute a
 // hash from its contents
-hash_stream_128 :: proc(s: io.Stream) -> ([16]byte, bool) {
-    hash: [16]byte
+hash_stream_128 :: proc(s: io.Stream) -> ([DIGEST_SIZE_128]byte, bool) {
+    hash: [DIGEST_SIZE_128]byte
    ctx: _tiger.Tiger_Context
    ctx.ver = 1
    _tiger.init(&ctx)
@@ -59,7 +82,7 @@ hash_stream_128 :: proc(s: io.Stream) -> ([16]byte, bool) {

 // hash_file_128 will read the file provided by the given handle
 // and compute a hash
-hash_file_128 :: proc(hd: os.Handle, load_at_once := false) -> ([16]byte, bool) {
+hash_file_128 :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE_128]byte, bool) {
    if !load_at_once {
        return hash_stream_128(os.stream_from_handle(hd))
    } else {
@@ -67,7 +90,7 @@ hash_file_128 :: proc(hd: os.Handle, load_at_once := false) -> ([16]byte, bool)
            return hash_bytes_128(buf[:]), ok
        }
    }
-    return [16]byte{}, false
+    return [DIGEST_SIZE_128]byte{}, false
 }

 hash_128 :: proc {
@@ -75,18 +98,20 @@ hash_128 :: proc {
    hash_file_128,
    hash_bytes_128,
    hash_string_128,
+    hash_bytes_to_buffer_128,
+    hash_string_to_buffer_128,
 }

 // hash_string_160 will hash the given input and return the
 // computed hash
-hash_string_160 :: proc(data: string) -> [20]byte {
+hash_string_160 :: proc(data: string) -> [DIGEST_SIZE_160]byte {
    return hash_bytes_160(transmute([]byte)(data))
 }

 // hash_bytes_160 will hash the given input and return the
 // computed hash
-hash_bytes_160 :: proc(data: []byte) -> [20]byte {
-    hash: [20]byte
+hash_bytes_160 :: proc(data: []byte) -> [DIGEST_SIZE_160]byte {
+    hash: [DIGEST_SIZE_160]byte
    ctx: _tiger.Tiger_Context
    ctx.ver = 1
    _tiger.init(&ctx)
@@ -95,10 +120,29 @@ hash_bytes_160 :: proc(data: []byte) -> [20]byte {
    return hash
 }

+// hash_string_to_buffer_160 will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer_160 :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer_160(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer_160 will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer_160 :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE_160, "Size of destination buffer is smaller than the digest size")
+    ctx: _tiger.Tiger_Context
+    ctx.ver = 1
+    _tiger.init(&ctx)
+    _tiger.update(&ctx, data)
+    _tiger.final(&ctx, hash)
+}
+
 // hash_stream_160 will read the stream in chunks and compute a
 // hash from its contents
-hash_stream_160 :: proc(s: io.Stream) -> ([20]byte, bool) {
-    hash: [20]byte
+hash_stream_160 :: proc(s: io.Stream) -> ([DIGEST_SIZE_160]byte, bool) {
+    hash: [DIGEST_SIZE_160]byte
    ctx: _tiger.Tiger_Context
    ctx.ver = 1
    _tiger.init(&ctx)
@@ -117,7 +161,7 @@ hash_stream_160 :: proc(s: io.Stream) -> ([20]byte, bool) {

 // hash_file_160 will read the file provided by the given handle
 // and compute a hash
-hash_file_160 :: proc(hd: os.Handle, load_at_once := false) -> ([20]byte, bool) {
+hash_file_160 :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE_160]byte, bool) {
    if !load_at_once {
        return hash_stream_160(os.stream_from_handle(hd))
    } else {
@@ -125,7 +169,7 @@ hash_file_160 :: proc(hd: os.Handle, load_at_once := false) -> ([20]byte, bool)
            return hash_bytes_160(buf[:]), ok
        }
    }
-    return [20]byte{}, false
+    return [DIGEST_SIZE_160]byte{}, false
 }

 hash_160 :: proc {
@@ -133,18 +177,20 @@ hash_160 :: proc {
    hash_file_160,
    hash_bytes_160,
    hash_string_160,
+    hash_bytes_to_buffer_160,
+    hash_string_to_buffer_160,
 }

 // hash_string_192 will hash the given input and return the
 // computed hash
-hash_string_192 :: proc(data: string) -> [24]byte {
+hash_string_192 :: proc(data: string) -> [DIGEST_SIZE_192]byte {
    return hash_bytes_192(transmute([]byte)(data))
 }

 // hash_bytes_192 will hash the given input and return the
 // computed hash
-hash_bytes_192 :: proc(data: []byte) -> [24]byte {
-    hash: [24]byte
+hash_bytes_192 :: proc(data: []byte) -> [DIGEST_SIZE_192]byte {
+    hash: [DIGEST_SIZE_192]byte
    ctx: _tiger.Tiger_Context
    ctx.ver = 1
    _tiger.init(&ctx)
@@ -153,10 +199,29 @@ hash_bytes_192 :: proc(data: []byte) -> [24]byte {
    return hash
 }

+// hash_string_to_buffer_192 will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer_192 :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer_192(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer_192 will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer_192 :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE_192, "Size of destination buffer is smaller than the digest size")
+    ctx: _tiger.Tiger_Context
+    ctx.ver = 1
+    _tiger.init(&ctx)
+    _tiger.update(&ctx, data)
+    _tiger.final(&ctx, hash)
+}
+
 // hash_stream_192 will read the stream in chunks and compute a
 // hash from its contents
-hash_stream_192 :: proc(s: io.Stream) -> ([24]byte, bool) {
-    hash: [24]byte
+hash_stream_192 :: proc(s: io.Stream) -> ([DIGEST_SIZE_192]byte, bool) {
+    hash: [DIGEST_SIZE_192]byte
    ctx: _tiger.Tiger_Context
    ctx.ver = 1
    _tiger.init(&ctx)
@@ -175,7 +240,7 @@ hash_stream_192 :: proc(s: io.Stream) -> ([24]byte, bool) {

 // hash_file_192 will read the file provided by the given handle
 // and compute a hash
-hash_file_192 :: proc(hd: os.Handle, load_at_once := false) -> ([24]byte, bool) {
+hash_file_192 :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE_192]byte, bool) {
    if !load_at_once {
        return hash_stream_192(os.stream_from_handle(hd))
    } else {
@@ -183,7 +248,7 @@ hash_file_192 :: proc(hd: os.Handle, load_at_once := false) -> ([24]byte, bool)
            return hash_bytes_192(buf[:]), ok
        }
    }
-    return [24]byte{}, false
+    return [DIGEST_SIZE_192]byte{}, false
 }

 hash_192 :: proc {
@@ -191,6 +256,8 @@ hash_192 :: proc {
    hash_file_192,
    hash_bytes_192,
    hash_string_192,
+    hash_bytes_to_buffer_192,
+    hash_string_to_buffer_192,
 }

 /*
@@ -19,16 +19,20 @@ import "../_tiger"
    High level API
 */

+DIGEST_SIZE_128 :: 16
+DIGEST_SIZE_160 :: 20
+DIGEST_SIZE_192 :: 24
+
 // hash_string_128 will hash the given input and return the
 // computed hash
-hash_string_128 :: proc(data: string) -> [16]byte {
+hash_string_128 :: proc(data: string) -> [DIGEST_SIZE_128]byte {
    return hash_bytes_128(transmute([]byte)(data))
 }

 // hash_bytes_128 will hash the given input and return the
 // computed hash
-hash_bytes_128 :: proc(data: []byte) -> [16]byte {
-    hash: [16]byte
+hash_bytes_128 :: proc(data: []byte) -> [DIGEST_SIZE_128]byte {
+    hash: [DIGEST_SIZE_128]byte
    ctx: _tiger.Tiger_Context
    ctx.ver = 2
    _tiger.init(&ctx)
@@ -37,10 +41,29 @@ hash_bytes_128 :: proc(data: []byte) -> [16]byte {
    return hash
 }

+// hash_string_to_buffer_128 will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer_128 :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer_128(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer_128 will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer_128 :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE_128, "Size of destination buffer is smaller than the digest size")
+    ctx: _tiger.Tiger_Context
+    ctx.ver = 2
+    _tiger.init(&ctx)
+    _tiger.update(&ctx, data)
+    _tiger.final(&ctx, hash)
+}
+
 // hash_stream_128 will read the stream in chunks and compute a
 // hash from its contents
-hash_stream_128 :: proc(s: io.Stream) -> ([16]byte, bool) {
-    hash: [16]byte
+hash_stream_128 :: proc(s: io.Stream) -> ([DIGEST_SIZE_128]byte, bool) {
+    hash: [DIGEST_SIZE_128]byte
    ctx: _tiger.Tiger_Context
    ctx.ver = 2
    _tiger.init(&ctx)
@@ -59,7 +82,7 @@ hash_stream_128 :: proc(s: io.Stream) -> ([16]byte, bool) {

 // hash_file_128 will read the file provided by the given handle
 // and compute a hash
-hash_file_128 :: proc(hd: os.Handle, load_at_once := false) -> ([16]byte, bool) {
+hash_file_128 :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE_128]byte, bool) {
    if !load_at_once {
        return hash_stream_128(os.stream_from_handle(hd))
    } else {
@@ -67,7 +90,7 @@ hash_file_128 :: proc(hd: os.Handle, load_at_once := false) -> ([16]byte, bool)
            return hash_bytes_128(buf[:]), ok
        }
    }
-    return [16]byte{}, false
+    return [DIGEST_SIZE_128]byte{}, false
 }

 hash_128 :: proc {
@@ -75,18 +98,20 @@ hash_128 :: proc {
    hash_file_128,
    hash_bytes_128,
    hash_string_128,
+    hash_bytes_to_buffer_128,
+    hash_string_to_buffer_128,
 }

 // hash_string_160 will hash the given input and return the
 // computed hash
-hash_string_160 :: proc(data: string) -> [20]byte {
+hash_string_160 :: proc(data: string) -> [DIGEST_SIZE_160]byte {
    return hash_bytes_160(transmute([]byte)(data))
 }

 // hash_bytes_160 will hash the given input and return the
 // computed hash
-hash_bytes_160 :: proc(data: []byte) -> [20]byte {
-    hash: [20]byte
+hash_bytes_160 :: proc(data: []byte) -> [DIGEST_SIZE_160]byte {
+    hash: [DIGEST_SIZE_160]byte
    ctx: _tiger.Tiger_Context
    ctx.ver = 2
    _tiger.init(&ctx)
@@ -95,10 +120,29 @@ hash_bytes_160 :: proc(data: []byte) -> [20]byte {
    return hash
 }

+// hash_string_to_buffer_160 will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer_160 :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer_160(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer_160 will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer_160 :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE_160, "Size of destination buffer is smaller than the digest size")
+    ctx: _tiger.Tiger_Context
+    ctx.ver = 2
+    _tiger.init(&ctx)
+    _tiger.update(&ctx, data)
+    _tiger.final(&ctx, hash)
+}
+
 // hash_stream_160 will read the stream in chunks and compute a
 // hash from its contents
-hash_stream_160 :: proc(s: io.Stream) -> ([20]byte, bool) {
-    hash: [20]byte
+hash_stream_160 :: proc(s: io.Stream) -> ([DIGEST_SIZE_160]byte, bool) {
+    hash: [DIGEST_SIZE_160]byte
    ctx: _tiger.Tiger_Context
    ctx.ver = 2
    _tiger.init(&ctx)
@@ -117,7 +161,7 @@ hash_stream_160 :: proc(s: io.Stream) -> ([20]byte, bool) {

 // hash_file_160 will read the file provided by the given handle
 // and compute a hash
-hash_file_160 :: proc(hd: os.Handle, load_at_once := false) -> ([20]byte, bool) {
+hash_file_160 :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE_160]byte, bool) {
    if !load_at_once {
        return hash_stream_160(os.stream_from_handle(hd))
    } else {
@@ -125,7 +169,7 @@ hash_file_160 :: proc(hd: os.Handle, load_at_once := false) -> ([20]byte, bool)
            return hash_bytes_160(buf[:]), ok
        }
    }
-    return [20]byte{}, false
+    return [DIGEST_SIZE_160]byte{}, false
 }

 hash_160 :: proc {
@@ -133,18 +177,20 @@ hash_160 :: proc {
    hash_file_160,
    hash_bytes_160,
    hash_string_160,
+    hash_bytes_to_buffer_160,
+    hash_string_to_buffer_160,
 }

 // hash_string_192 will hash the given input and return the
 // computed hash
-hash_string_192 :: proc(data: string) -> [24]byte {
+hash_string_192 :: proc(data: string) -> [DIGEST_SIZE_192]byte {
    return hash_bytes_192(transmute([]byte)(data))
 }

 // hash_bytes_192 will hash the given input and return the
 // computed hash
-hash_bytes_192 :: proc(data: []byte) -> [24]byte {
-    hash: [24]byte
+hash_bytes_192 :: proc(data: []byte) -> [DIGEST_SIZE_192]byte {
+    hash: [DIGEST_SIZE_192]byte
    ctx: _tiger.Tiger_Context
    ctx.ver = 2
    _tiger.init(&ctx)
@@ -153,10 +199,29 @@ hash_bytes_192 :: proc(data: []byte) -> [24]byte {
    return hash
 }

+// hash_string_to_buffer_192 will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer_192 :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer_192(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer_192 will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer_192 :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE_192, "Size of destination buffer is smaller than the digest size")
+    ctx: _tiger.Tiger_Context
+    ctx.ver = 2
+    _tiger.init(&ctx)
+    _tiger.update(&ctx, data)
+    _tiger.final(&ctx, hash)
+}
+
 // hash_stream_192 will read the stream in chunks and compute a
 // hash from its contents
-hash_stream_192 :: proc(s: io.Stream) -> ([24]byte, bool) {
-    hash: [24]byte
+hash_stream_192 :: proc(s: io.Stream) -> ([DIGEST_SIZE_192]byte, bool) {
+    hash: [DIGEST_SIZE_192]byte
    ctx: _tiger.Tiger_Context
    ctx.ver = 2
    _tiger.init(&ctx)
@@ -175,7 +240,7 @@ hash_stream_192 :: proc(s: io.Stream) -> ([24]byte, bool) {

 // hash_file_192 will read the file provided by the given handle
 // and compute a hash
-hash_file_192 :: proc(hd: os.Handle, load_at_once := false) -> ([24]byte, bool) {
+hash_file_192 :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE_192]byte, bool) {
    if !load_at_once {
        return hash_stream_192(os.stream_from_handle(hd))
    } else {
@@ -183,7 +248,7 @@ hash_file_192 :: proc(hd: os.Handle, load_at_once := false) -> ([24]byte, bool)
            return hash_bytes_192(buf[:]), ok
        }
    }
-    return [24]byte{}, false
+    return [DIGEST_SIZE_192]byte{}, false
 }

 hash_192 :: proc {
@@ -191,6 +256,8 @@ hash_192 :: proc {
    hash_file_192,
    hash_bytes_192,
    hash_string_192,
+    hash_bytes_to_buffer_192,
+    hash_string_to_buffer_192,
 }

 /*
@@ -19,16 +19,18 @@ import "../util"
    High level API
 */

+DIGEST_SIZE :: 64
+
 // hash_string will hash the given input and return the
 // computed hash
-hash_string :: proc(data: string) -> [64]byte {
+hash_string :: proc(data: string) -> [DIGEST_SIZE]byte {
    return hash_bytes(transmute([]byte)(data))
 }

 // hash_bytes will hash the given input and return the
 // computed hash
-hash_bytes :: proc(data: []byte) -> [64]byte {
-	hash: [64]byte
+hash_bytes :: proc(data: []byte) -> [DIGEST_SIZE]byte {
+	hash: [DIGEST_SIZE]byte
 	ctx: Whirlpool_Context
    // init(&ctx) No-op
    update(&ctx, data)
@@ -36,10 +38,28 @@ hash_bytes :: proc(data: []byte) -> [64]byte {
    return hash
 }

+// hash_string_to_buffer will hash the given input and assign the
+// computed hash to the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_string_to_buffer :: proc(data: string, hash: []byte) {
+    hash_bytes_to_buffer(transmute([]byte)(data), hash)
+}
+
+// hash_bytes_to_buffer will hash the given input and write the
+// computed hash into the second parameter.
+// It requires that the destination buffer is at least as big as the digest size
+hash_bytes_to_buffer :: proc(data, hash: []byte) {
+    assert(len(hash) >= DIGEST_SIZE, "Size of destination buffer is smaller than the digest size")
+    ctx: Whirlpool_Context
+    // init(&ctx) No-op
+    update(&ctx, data)
+    final(&ctx, hash)
+}
+
 // hash_stream will read the stream in chunks and compute a
 // hash from its contents
-hash_stream :: proc(s: io.Stream) -> ([64]byte, bool) {
-	hash: [64]byte
+hash_stream :: proc(s: io.Stream) -> ([DIGEST_SIZE]byte, bool) {
+	hash: [DIGEST_SIZE]byte
 	ctx: Whirlpool_Context
 	// init(&ctx) No-op
 	buf := make([]byte, 512)
@@ -57,7 +77,7 @@ hash_stream :: proc(s: io.Stream) -> ([64]byte, bool) {

 // hash_file will read the file provided by the given handle
 // and compute a hash
-hash_file :: proc(hd: os.Handle, load_at_once := false) -> ([64]byte, bool) {
+hash_file :: proc(hd: os.Handle, load_at_once := false) -> ([DIGEST_SIZE]byte, bool) {
 	if !load_at_once {
        return hash_stream(os.stream_from_handle(hd))
    } else {
@@ -65,7 +85,7 @@ hash_file :: proc(hd: os.Handle, load_at_once := false) -> ([64]byte, bool) {
            return hash_bytes(buf[:]), ok
        }
    }
-    return [64]byte{}, false
+    return [DIGEST_SIZE]byte{}, false
 }

 hash :: proc {
@@ -73,6 +93,8 @@ hash :: proc {
    hash_file,
    hash_bytes,
    hash_string,
+    hash_bytes_to_buffer,
+    hash_string_to_buffer,
 }

 /*
@@ -0,0 +1,221 @@
+package debug_pe
+
+PE_SIGNATURE_OFFSET_INDEX_POS :: 0x3c
+PE_SIGNATURE :: u32le(0x0000_4550) // "PE\x00\x00"
+PE_SIGNATURE_STRING :: "PE\x00\x00"
+
+OPTIONAL_HEADER_MAGIC :: enum u16le {
+	PE32      = 0x010b,
+	PE32_PLUS = 0x020b,
+}
+
+Optional_Header_Base :: struct #packed {
+	magic:                          OPTIONAL_HEADER_MAGIC,
+	major_linker_version:           u8,
+	minor_linker_version:           u8,
+	size_of_code:                   u32le,
+	size_of_initialized_data:       u32le,
+	size_of_uninitialized_data:     u32le,
+	address_of_entry_point:         u32le,
+	base_of_code:                   u32le,
+}
+
+File_Header :: struct #packed {
+	machine:                 IMAGE_FILE_MACHINE,
+	number_of_sections:      u16le,
+	time_date_stamp:         u32le,
+	pointer_to_symbol_table: u32le,
+	number_of_symbols:       u32le,
+	size_of_optional_header: u16le,
+	characteristics:         IMAGE_FILE_CHARACTERISTICS,
+}
+
+Data_Directory :: struct #packed {
+	virtual_address: u32le,
+	size:            u32le,
+}
+
+Optional_Header32 :: struct #packed {
+	using base: Optional_Header_Base,
+	base_of_data:                   u32le,
+	image_base:                     u32le,
+	section_alignment:              u32le,
+	file_alignment:                 u32le,
+	major_operating_system_version: u16le,
+	minor_operating_system_version: u16le,
+	major_image_version:            u16le,
+	minor_image_version:            u16le,
+	major_subsystem_version:        u16le,
+	minor_subsystem_version:        u16le,
+	win32_version_value:            u32le,
+	size_of_image:                  u32le,
+	size_of_headers:                u32le,
+	check_sum:                      u32le,
+	subsystem:                      IMAGE_SUBSYSTEM,
+	dll_characteristics:            IMAGE_DLLCHARACTERISTICS,
+	size_of_stack_reserve:          u32le,
+	size_of_stack_commit:           u32le,
+	size_of_heap_reserve:           u32le,
+	size_of_heap_commit:            u32le,
+	loader_flags:                   u32le,
+	number_of_rva_and_sizes:        u32le,
+	data_directory:                 [16]Data_Directory,
+}
+
+Optional_Header64 :: struct #packed {
+	using base: Optional_Header_Base,
+	image_base:                     u64le,
+	section_alignment:              u32le,
+	file_alignment:                 u32le,
+	major_operating_system_version: u16le,
+	minor_operating_system_version: u16le,
+	major_image_version:            u16le,
+	minor_image_version:            u16le,
+	major_subsystem_version:        u16le,
+	minor_subsystem_version:        u16le,
+	win32_version_value:            u32le,
+	size_of_image:                  u32le,
+	size_of_headers:                u32le,
+	check_sum:                      u32le,
+	subsystem:                      IMAGE_SUBSYSTEM,
+	dll_characteristics:            IMAGE_DLLCHARACTERISTICS,
+	size_of_stack_reserve:          u64le,
+	size_of_stack_commit:           u64le,
+	size_of_heap_reserve:           u64le,
+	size_of_heap_commit:            u64le,
+	loader_flags:                   u32le,
+	number_of_rva_and_sizes:        u32le,
+	data_directory:                 [16]Data_Directory,
+}
+
+// .debug section
+Debug_Directory_Entry :: struct {
+	characteristics:     u32le,
+	time_date_stamp:     u32le,
+	major_version:       u16le,
+	minor_version:       u16le,
+	type:                IMAGE_DEBUG_TYPE,
+	size_of_data:        u32le,
+	address_of_raw_data: u32le,
+	pointer_to_raw_data: u32le,
+}
+
+
+IMAGE_FILE_MACHINE :: enum u16le {
+	UNKNOWN     = 0x0,
+	AM33        = 0x1d3,
+	AMD64       = 0x8664,
+	ARM         = 0x1c0,
+	ARMNT       = 0x1c4,
+	ARM64       = 0xaa64,
+	EBC         = 0xebc,
+	I386        = 0x14c,
+	IA64        = 0x200,
+	LOONGARCH32 = 0x6232,
+	LOONGARCH64 = 0x6264,
+	M32R        = 0x9041,
+	MIPS16      = 0x266,
+	MIPSFPU     = 0x366,
+	MIPSFPU16   = 0x466,
+	POWERPC     = 0x1f0,
+	POWERPCFP   = 0x1f1,
+	R4000       = 0x166,
+	SH3         = 0x1a2,
+	SH3DSP      = 0x1a3,
+	SH4         = 0x1a6,
+	SH5         = 0x1a8,
+	THUMB       = 0x1c2,
+	WCEMIPSV2   = 0x169,
+}
+
+// IMAGE_DIRECTORY_ENTRY constants
+IMAGE_DIRECTORY_ENTRY :: enum u8 {
+	EXPORT         = 0,
+	IMPORT         = 1,
+	RESOURCE       = 2,
+	EXCEPTION      = 3,
+	SECURITY       = 4,
+	BASERELOC      = 5,
+	DEBUG          = 6,
+	ARCHITECTURE   = 7, // reserved
+	GLOBALPTR      = 8,
+	TLS            = 9,
+	LOAD_CONFIG    = 10,
+	BOUND_IMPORT   = 11,
+	IAT            = 12,
+	DELAY_IMPORT   = 13,
+	COM_DESCRIPTOR = 14, // DLR Runtime headers
+	_RESERVED      = 15,
+}
+#assert(len(IMAGE_DIRECTORY_ENTRY) == 16)
+
+
+IMAGE_FILE_CHARACTERISTICS :: distinct bit_set[IMAGE_FILE_CHARACTERISTIC; u16le]
+IMAGE_FILE_CHARACTERISTIC :: enum u16le {
+	RELOCS_STRIPPED         = 0,
+	EXECUTABLE_IMAGE        = 1,
+	LINE_NUMS_STRIPPED      = 2,
+	LOCAL_SYMS_STRIPPED     = 3,
+	AGGRESIVE_WS_TRIM       = 4,
+	LARGE_ADDRESS_AWARE     = 5,
+
+	BYTES_REVERSED_LO       = 7,
+	MACHINE_32BIT           = 8, // IMAGE_FILE_32BIT_MACHINE  originally
+	DEBUG_STRIPPED          = 9,
+	REMOVABLE_RUN_FROM_SWAP = 10,
+	NET_RUN_FROM_SWAP       = 11,
+	SYSTEM                  = 12,
+	DLL                     = 13,
+	UP_SYSTEM_ONLY          = 14,
+	BYTES_REVERSED_HI       = 15,
+}
+
+IMAGE_SUBSYSTEM :: enum u16le {
+	UNKNOWN                  = 0,
+	NATIVE                   = 1,
+	WINDOWS_GUI              = 2,
+	WINDOWS_CUI              = 3,
+	OS2_CUI                  = 5,
+	POSIX_CUI                = 7,
+	NATIVE_WINDOWS           = 8,
+	WINDOWS_CE_GUI           = 9,
+	EFI_APPLICATION          = 10,
+	EFI_BOOT_SERVICE_DRIVER  = 11,
+	EFI_RUNTIME_DRIVER       = 12,
+	EFI_ROM                  = 13,
+	XBOX                     = 14,
+	WINDOWS_BOOT_APPLICATION = 16,
+}
+
+IMAGE_DLLCHARACTERISTICS :: distinct bit_set[IMAGE_DLLCHARACTERISTIC; u16le]
+IMAGE_DLLCHARACTERISTIC :: enum u16le {
+	HIGH_ENTROPY_VA       = 5,
+	DYNAMIC_BASE          = 6,
+	FORCE_INTEGRITY       = 7,
+	NX_COMPAT             = 8,
+	NO_ISOLATION          = 9,
+	NO_SEH                = 10,
+	NO_BIND               = 11,
+	APPCONTAINER          = 12,
+	WDM_DRIVER            = 13,
+	GUARD_CF              = 14,
+	TERMINAL_SERVER_AWARE = 15,
+}
+
+IMAGE_DEBUG_TYPE :: enum u32le {
+	UNKNOWN               = 0,  // An unknown value that is ignored by all tools.
+	COFF                  = 1,  // The COFF debug information (line numbers, symbol table, and string table). This type of debug information is also pointed to by fields in the file headers.
+	CODEVIEW              = 2,  // The Visual C++ debug information.
+	FPO                   = 3,  // The frame pointer omission (FPO) information. This information tells the debugger how to interpret nonstandard stack frames, which use the EBP register for a purpose other than as a frame pointer.
+	MISC                  = 4,  // The location of DBG file.
+	EXCEPTION             = 5,  // A copy of .pdata section.
+	FIXUP                 = 6,  // Reserved.
+	OMAP_TO_SRC           = 7,  // The mapping from an RVA in image to an RVA in source image.
+	OMAP_FROM_SRC         = 8,  // The mapping from an RVA in source image to an RVA in image.
+	BORLAND               = 9,  // Reserved for Borland.
+	RESERVED10            = 10, // Reserved.
+	CLSID                 = 11, // Reserved.
+	REPRO                 = 16, // PE determinism or reproducibility.
+	EX_DLLCHARACTERISTICS = 20, // Extended DLL characteristics bits.
+}
+
@@ -0,0 +1,131 @@
+package debug_pe
+
+import "core:runtime"
+import "core:io"
+
+Section_Header32 :: struct {
+	name:                    [8]u8,
+	virtual_size:            u32le,
+	virtual_address:         u32le,
+	size_of_raw_data:        u32le,
+	pointer_to_raw_data:     u32le,
+	pointer_to_relocations:  u32le,
+	pointer_to_line_numbers: u32le,
+	number_of_relocations:   u16le,
+	number_of_line_numbers:  u16le,
+	characteristics:         IMAGE_SCN_CHARACTERISTICS,
+}
+
+Reloc :: struct {
+	virtual_address:    u32le,
+	symbol_table_index: u32le,
+	type:               IMAGE_REL,
+}
+
+IMAGE_SCN_CHARACTERISTICS :: enum u32le {
+	TYPE_NO_PAD            = 0x00000008, // The section should not be padded to the next boundary. This flag is obsolete and is replaced by IMAGE_SCN_ALIGN_1BYTES. This is valid only for object files. = 0x00000010, // Reserved for future use.
+	CNT_CODE               = 0x00000020, // The section contains executable code.
+	CNT_INITIALIZED_DATA   = 0x00000040, // The section contains initialized data.
+	CNT_UNINITIALIZED_DATA = 0x00000080, // The section contains uninitialized data.
+	LNK_OTHER              = 0x00000100, // Reserved for future use.
+	LNK_INFO               = 0x00000200, // The section contains comments or other information. The .drectve section has this type. This is valid for object files only. = 0x00000400, // Reserved for future use.
+	LNK_REMOVE             = 0x00000800, // The section will not become part of the image. This is valid only for object files.
+	LNK_COMDAT             = 0x00001000, // The section contains COMDAT data. For more information, see COMDAT Sections (Object Only). This is valid only for object files.
+	GPREL                  = 0x00008000, // The section contains data referenced through the global pointer (GP).
+	MEM_PURGEABLE          = 0x00020000, // Reserved for future use.
+	MEM_16BIT              = 0x00020000, // Reserved for future use.
+	MEM_LOCKED             = 0x00040000, // Reserved for future use.
+	MEM_PRELOAD            = 0x00080000, // Reserved for future use.
+	ALIGN_1BYTES           = 0x00100000, // Align data on a 1-byte boundary. Valid only for object files.
+	ALIGN_2BYTES           = 0x00200000, // Align data on a 2-byte boundary. Valid only for object files.
+	ALIGN_4BYTES           = 0x00300000, // Align data on a 4-byte boundary. Valid only for object files.
+	ALIGN_8BYTES           = 0x00400000, // Align data on an 8-byte boundary. Valid only for object files.
+	ALIGN_16BYTES          = 0x00500000, // Align data on a 16-byte boundary. Valid only for object files.
+	ALIGN_32BYTES          = 0x00600000, // Align data on a 32-byte boundary. Valid only for object files.
+	ALIGN_64BYTES          = 0x00700000, // Align data on a 64-byte boundary. Valid only for object files.
+	ALIGN_128BYTES         = 0x00800000, // Align data on a 128-byte boundary. Valid only for object files.
+	ALIGN_256BYTES         = 0x00900000, // Align data on a 256-byte boundary. Valid only for object files.
+	ALIGN_512BYTES         = 0x00A00000, // Align data on a 512-byte boundary. Valid only for object files.
+	ALIGN_1024BYTES        = 0x00B00000, // Align data on a 1024-byte boundary. Valid only for object files.
+	ALIGN_2048BYTES        = 0x00C00000, // Align data on a 2048-byte boundary. Valid only for object files.
+	ALIGN_4096BYTES        = 0x00D00000, // Align data on a 4096-byte boundary. Valid only for object files.
+	ALIGN_8192BYTES        = 0x00E00000, // Align data on an 8192-byte boundary. Valid only for object files.
+	LNK_NRELOC_OVFL        = 0x01000000, // The section contains extended relocations.
+	MEM_DISCARDABLE        = 0x02000000, // The section can be discarded as needed.
+	MEM_NOT_CACHED         = 0x04000000, // The section cannot be cached.
+	MEM_NOT_PAGED          = 0x08000000, // The section is not pageable.
+	MEM_SHARED             = 0x10000000, // The section can be shared in memory.
+	MEM_EXECUTE            = 0x20000000, // The section can be executed as code.
+	MEM_READ               = 0x40000000, // The section can be read.
+	MEM_WRITE              = 0x80000000, // The section can be written to.
+}
+
+
+IMAGE_REL :: enum u16le {
+	I386_ABSOLUTE         = 0x0000,
+	I386_DIR16            = 0x0001,
+	I386_REL16            = 0x0002,
+	I386_DIR32            = 0x0006,
+	I386_DIR32NB          = 0x0007,
+	I386_SEG12            = 0x0009,
+	I386_SECTION          = 0x000A,
+	I386_SECREL           = 0x000B,
+	I386_TOKEN            = 0x000C,
+	I386_SECREL7          = 0x000D,
+	I386_REL32            = 0x0014,
+
+	AMD64_ABSOLUTE        = 0x0000,
+	AMD64_ADDR64          = 0x0001,
+	AMD64_ADDR32          = 0x0002,
+	AMD64_ADDR32NB        = 0x0003,
+	AMD64_REL32           = 0x0004,
+	AMD64_REL32_1         = 0x0005,
+	AMD64_REL32_2         = 0x0006,
+	AMD64_REL32_3         = 0x0007,
+	AMD64_REL32_4         = 0x0008,
+	AMD64_REL32_5         = 0x0009,
+	AMD64_SECTION         = 0x000A,
+	AMD64_SECREL          = 0x000B,
+	AMD64_SECREL7         = 0x000C,
+	AMD64_TOKEN           = 0x000D,
+	AMD64_SREL32          = 0x000E,
+	AMD64_PAIR            = 0x000F,
+	AMD64_SSPAN32         = 0x0010,
+
+	ARM_ABSOLUTE          = 0x0000,
+	ARM_ADDR32            = 0x0001,
+	ARM_ADDR32NB          = 0x0002,
+	ARM_BRANCH24          = 0x0003,
+	ARM_BRANCH11          = 0x0004,
+	ARM_SECTION           = 0x000E,
+	ARM_SECREL            = 0x000F,
+	ARM_MOV32             = 0x0010,
+
+	THUMB_MOV32           = 0x0011,
+	THUMB_BRANCH20        = 0x0012,
+	THUMB_BRANCH24        = 0x0014,
+	THUMB_BLX23           = 0x0015,
+
+	ARM_PAIR              = 0x0016,
+
+	ARM64_ABSOLUTE        = 0x0000,
+	ARM64_ADDR32          = 0x0001,
+	ARM64_ADDR32NB        = 0x0002,
+	ARM64_BRANCH26        = 0x0003,
+	ARM64_PAGEBASE_REL21  = 0x0004,
+	ARM64_REL21           = 0x0005,
+	ARM64_PAGEOFFSET_12A  = 0x0006,
+	ARM64_PAGEOFFSET_12L  = 0x0007,
+	ARM64_SECREL          = 0x0008,
+	ARM64_SECREL_LOW12A   = 0x0009,
+	ARM64_SECREL_HIGH12A  = 0x000A,
+	ARM64_SECREL_LOW12L   = 0x000B,
+	ARM64_TOKEN           = 0x000C,
+	ARM64_SECTION         = 0x000D,
+	ARM64_ADDR64          = 0x000E,
+	ARM64_BRANCH19        = 0x000F,
+	ARM64_BRANCH14        = 0x0010,
+	ARM64_REL32           = 0x0011,
+}
+
+PE_CODE_VIEW_SIGNATURE_RSDS :: u32le(0x5344_5352)
@@ -0,0 +1,108 @@
+package debug_pe
+
+COFF_SYMBOL_SIZE :: 18
+
+COFF_Symbol :: struct {
+	name:                  [8]u8,
+	value:                 u32le,
+	section_number:        i16le,
+	type:                  IMAGE_SYM_TYPE,
+	storage_class:         IMAGE_SYM_CLASS,
+	number_of_aux_symbols: u8,
+}
+
+// COFF_Symbol_Aux_Format5 describes the expected form of an aux symbol
+// attached to a section definition symbol. The PE format defines a
+// number of different aux symbol formats: format 1 for function
+// definitions, format 2 for .be and .ef symbols, and so on. Format 5
+// holds extra info associated with a section definition, including
+// number of relocations + line numbers, as well as COMDAT info. See
+// https://docs.microsoft.com/en-us/windows/win32/debug/pe-format#auxiliary-format-5-section-definitions
+// for more on what's going on here.
+COFF_Symbol_Aux_Format5 :: struct {
+	size:             u32le,
+	num_relocs:       u16le,
+	num_line_numbers: u16le,
+	checksum:         u32le,
+	sec_num:          u16le,
+	selection:        IMAGE_COMDAT_SELECT,
+	_:                [3]u8, // padding
+}
+
+IMAGE_COMDAT_SELECT :: enum u8 {
+	NODUPLICATES = 1,
+	ANY          = 2,
+	SAME_SIZE    = 3,
+	EXACT_MATCH  = 4,
+	ASSOCIATIVE  = 5,
+	LARGEST      = 6,
+}
+
+
+// The symbol record is not yet assigned a section. A value of zero indicates
+// that a reference to an external symbol is defined elsewhere. A value of
+// non-zero is a common symbol with a size that is specified by the value.
+IMAGE_SYM_UNDEFINED              :: 0
+// The symbol has an absolute (non-relocatable) value and is not an address.
+IMAGE_SYM_ABSOLUTE               :: -1
+// The symbol provides general type or debugging information but does not
+// correspond to a section. Microsoft tools use this setting along
+// with .file records (storage class FILE).
+IMAGE_SYM_DEBUG                  :: -2
+
+IMAGE_SYM_TYPE :: enum u16le {
+	NULL   = 0,
+	VOID   = 1,
+	CHAR   = 2,
+	SHORT  = 3,
+	INT    = 4,
+	LONG   = 5,
+	FLOAT  = 6,
+	DOUBLE = 7,
+	STRUCT = 8,
+	UNION  = 9,
+	ENUM   = 10,
+	MOE    = 11,
+	BYTE   = 12,
+	WORD   = 13,
+	UINT   = 14,
+	DWORD  = 15,
+	PCODE  = 32768,
+
+	DTYPE_NULL     = 0,
+	DTYPE_POINTER  = 0x10,
+	DTYPE_FUNCTION = 0x20,
+	DTYPE_ARRAY    = 0x30,
+}
+
+IMAGE_SYM_CLASS :: enum u8 {
+	NULL             = 0,
+	AUTOMATIC        = 1,
+	EXTERNAL         = 2,
+	STATIC           = 3,
+	REGISTER         = 4,
+	EXTERNAL_DEF     = 5,
+	LABEL            = 6,
+	UNDEFINED_LABEL  = 7,
+	MEMBER_OF_STRUCT = 8,
+	ARGUMENT         = 9,
+	STRUCT_TAG       = 10,
+	MEMBER_OF_UNION  = 11,
+	UNION_TAG        = 12,
+	TYPE_DEFINITION  = 13,
+	UNDEFINED_STATIC = 14,
+	ENUM_TAG         = 15,
+	MEMBER_OF_ENUM   = 16,
+	REGISTER_PARAM   = 17,
+	BIT_FIELD        = 18,
+	FAR_EXTERNAL     = 68, // Not in PECOFF v8 spec
+	BLOCK            = 100,
+	FUNCTION         = 101,
+	END_OF_STRUCT    = 102,
+	FILE             = 103,
+	SECTION          = 104,
+	WEAK_EXTERNAL    = 105,
+	CLR_TOKEN        = 107,
+
+	END_OF_FUNCTION  = 255,
+}
@@ -1,3 +1,15 @@
 package dynlib

 Library :: distinct rawptr
+
+load_library :: proc(path: string, global_symbols := false) -> (Library, bool) {
+	return _load_library(path, global_symbols)
+}
+
+unload_library :: proc(library: Library) -> bool {
+	return _unload_library(library)
+}
+
+symbol_address :: proc(library: Library, symbol: string) -> (ptr: rawptr, found: bool) #optional_ok {
+	return _symbol_address(library, symbol)
+}
@@ -1,23 +1,24 @@
-// +build linux, darwin, freebsd
+//+build linux, darwin, freebsd, openbsd
+//+private
 package dynlib

 import "core:os"

-load_library :: proc(path: string, global_symbols := false) -> (Library, bool) {
-    flags := os.RTLD_NOW
-    if global_symbols {
-    	flags |= os.RTLD_GLOBAL
-    }
-    lib := os.dlopen(path, flags)
-    return Library(lib), lib != nil
+_load_library :: proc(path: string, global_symbols := false) -> (Library, bool) {
+	flags := os.RTLD_NOW
+	if global_symbols {
+		flags |= os.RTLD_GLOBAL
+	}
+	lib := os.dlopen(path, flags)
+	return Library(lib), lib != nil
 }

-unload_library :: proc(library: Library) {
-    os.dlclose(rawptr(library))
+_unload_library :: proc(library: Library) -> bool {
+	return os.dlclose(rawptr(library))
 }

-symbol_address :: proc(library: Library, symbol: string) -> (ptr: rawptr, found: bool) {
-    ptr = os.dlsym(rawptr(library), symbol)
-    found = ptr != nil
-    return
+_symbol_address :: proc(library: Library, symbol: string) -> (ptr: rawptr, found: bool) {
+	ptr = os.dlsym(rawptr(library), symbol)
+	found = ptr != nil
+	return
 }
@@ -1,10 +1,11 @@
-// +build windows
+//+build windows
+//+private
 package dynlib

 import win32 "core:sys/windows"
 import "core:strings"

-load_library :: proc(path: string, global_symbols := false) -> (Library, bool) {
+_load_library :: proc(path: string, global_symbols := false) -> (Library, bool) {
 	// NOTE(bill): 'global_symbols' is here only for consistency with POSIX which has RTLD_GLOBAL

 	wide_path := win32.utf8_to_wstring(path, context.temp_allocator)
@@ -12,12 +13,12 @@ load_library :: proc(path: string, global_symbols := false) -> (Library, bool) {
 	return handle, handle != nil
 }

-unload_library :: proc(library: Library) -> bool {
+_unload_library :: proc(library: Library) -> bool {
 	ok := win32.FreeLibrary(cast(win32.HMODULE)library)
 	return bool(ok)
 }

-symbol_address :: proc(library: Library, symbol: string) -> (ptr: rawptr, found: bool) {
+_symbol_address :: proc(library: Library, symbol: string) -> (ptr: rawptr, found: bool) {
 	c_str := strings.clone_to_cstring(symbol, context.temp_allocator)
 	ptr = win32.GetProcAddress(cast(win32.HMODULE)library, c_str)
 	found = ptr != nil
@@ -34,6 +34,10 @@ Reader :: struct {
 	// If lazy_quotes is true, a quote may appear in an unquoted field and a non-doubled quote may appear in a quoted field
 	lazy_quotes: bool,

+	// multiline_fields, when set to true, will treat a field starting with a " as a multiline string
+	// therefore, instead of reading until the next \n, it'll read until the next "
+	multiline_fields: bool,
+
 	// reuse_record controls whether calls to 'read' may return a slice using the backing buffer
 	// for performance
 	// By default, each call to 'read' returns a newly allocated slice
@@ -194,32 +198,72 @@ is_valid_delim :: proc(r: rune) -> bool {
@private
 _read_record :: proc(r: ^Reader, dst: ^[dynamic]string, allocator := context.allocator) -> ([]string, Error) {
 	read_line :: proc(r: ^Reader) -> ([]byte, io.Error) {
-		line, err := bufio.reader_read_slice(&r.r, '\n')
-		if err == .Buffer_Full {
-			clear(&r.raw_buffer)
-			append(&r.raw_buffer, ..line)
-			for err == .Buffer_Full {
-				line, err = bufio.reader_read_slice(&r.r, '\n')
+		if !r.multiline_fields {
+			line, err := bufio.reader_read_slice(&r.r, '\n')
+			if err == .Buffer_Full {
+				clear(&r.raw_buffer)
 				append(&r.raw_buffer, ..line)
+				for err == .Buffer_Full {
+					line, err = bufio.reader_read_slice(&r.r, '\n')
+					append(&r.raw_buffer, ..line)
+				}
+				line = r.raw_buffer[:]
 			}
-			line = r.raw_buffer[:]
-		}
-		if len(line) > 0 && err == .EOF {
-			err = nil
-			if line[len(line)-1] == '\r' {
-				line = line[:len(line)-1]
+			if len(line) > 0 && err == .EOF {
+				err = nil
+				if line[len(line)-1] == '\r' {
+					line = line[:len(line)-1]
+				}
 			}
-		}
-		r.line_count += 1
+			r.line_count += 1

-		// normalize \r\n to \n
-		n := len(line)
-		for n >= 2 && string(line[n-2:]) == "\r\n" {
-			line[n-2] = '\n'
-			line = line[:n-1]
-		}
+			// normalize \r\n to \n
+			n := len(line)
+			for n >= 2 && string(line[n-2:]) == "\r\n" {
+				line[n-2] = '\n'
+				line = line[:n-1]
+			}
+			return line, err

-		return line, err
+		} else {
+			// Reading a "line" that can possibly contain multiline fields.
+			// Unfortunately, this means we need to read a character at a time.
+
+			err:       io.Error
+			cur:       rune
+			is_quoted: bool
+
+			field_length := 0
+
+			clear(&r.raw_buffer)
+
+			read_loop: for err == .None {
+				cur, _, err = bufio.reader_read_rune(&r.r)
+
+				if err != .None { break read_loop }
+
+				switch cur {
+				case '"':
+					is_quoted = field_length == 0
+					field_length += 1
+
+				case '\n', '\r':
+					if !is_quoted { break read_loop }
+
+				case r.comma:
+					field_length = 0
+
+				case:
+					field_length += 1
+				}
+
+				rune_buf, rune_len := utf8.encode_rune(cur)
+				append(&r.raw_buffer, ..rune_buf[:rune_len])
+			}
+
+			return r.raw_buffer[:], err
+		}
+		unreachable()
 	}

 	length_newline :: proc(b: []byte) -> int {
@@ -0,0 +1,23 @@
+/*
+    Package endian implements sa simple translation between bytes and numbers with
+    specific endian encodings.
+
+    buf: [100]u8
+    put_u16(buf[:], .Little, 16) or_return
+
+    You may ask yourself, why isn't `byte_order` platform Endianness by default, so we can write:
+    put_u16(buf[:], 16) or_return
+
+    The answer is that very few file formats are written in native/platform endianness. Most of them specify the endianness of
+    each of their fields, or use a header field which specifies it for the entire file.
+
+    e.g. a file which specifies it at the top for all fields could do this:
+    file_order := .Little if buf[0] == 0 else .Big
+    field := get_u16(buf[1:], file_order) or_return
+
+    If on the other hand a field is *always* Big-Endian, you're wise to explicitly state it for the benefit of the reader,
+    be that your future self or someone else.
+
+    field := get_u16(buf[:], .Big) or_return
+*/
+package encoding_endian
@@ -0,0 +1,153 @@
+package encoding_endian
+
+Byte_Order :: enum u8 {
+	Little,
+	Big,
+}
+
+PLATFORM_BYTE_ORDER :: Byte_Order.Little when ODIN_ENDIAN == .Little else Byte_Order.Big
+
+get_u16 :: proc(b: []byte, order: Byte_Order) -> (v: u16, ok: bool) {
+	if len(b) < 2 {
+		return 0, false
+	}
+	#no_bounds_check if order == .Little {
+		v = u16(b[0]) | u16(b[1])<<8
+	} else {
+		v = u16(b[1]) | u16(b[0])<<8
+	}
+	return v, true
+}
+get_u32 :: proc(b: []byte, order: Byte_Order) -> (v: u32, ok: bool) {
+	if len(b) < 4 {
+		return 0, false
+	}
+	#no_bounds_check if order == .Little {
+		v = u32(b[0]) | u32(b[1])<<8 | u32(b[2])<<16 | u32(b[3])<<24
+	} else {
+		v = u32(b[3]) | u32(b[2])<<8 | u32(b[1])<<16 | u32(b[0])<<24
+	}
+	return v, true
+}
+
+get_u64 :: proc(b: []byte, order: Byte_Order) -> (v: u64, ok: bool) {
+	if len(b) < 8 {
+		return 0, false
+	}
+	#no_bounds_check if order == .Little {
+		v = u64(b[0]) | u64(b[1])<<8 | u64(b[2])<<16 | u64(b[3])<<24 |
+		    u64(b[4])<<32 | u64(b[5])<<40 | u64(b[6])<<48 | u64(b[7])<<56
+	} else {
+		v = u64(b[7]) | u64(b[6])<<8 | u64(b[5])<<16 | u64(b[4])<<24 |
+		    u64(b[3])<<32 | u64(b[2])<<40 | u64(b[1])<<48 | u64(b[0])<<56
+	}
+	return v, true
+}
+
+get_i16 :: proc(b: []byte, order: Byte_Order) -> (i16, bool) {
+	v, ok := get_u16(b, order)
+	return i16(v), ok
+}
+get_i32 :: proc(b: []byte, order: Byte_Order) -> (i32, bool) {
+	v, ok := get_u32(b, order)
+	return i32(v), ok
+}
+get_i64 :: proc(b: []byte, order: Byte_Order) -> (i64, bool) {
+	v, ok := get_u64(b, order)
+	return i64(v), ok
+}
+
+get_f16 :: proc(b: []byte, order: Byte_Order) -> (f16, bool) {
+	v, ok := get_u16(b, order)
+	return transmute(f16)v, ok
+}
+get_f32 :: proc(b: []byte, order: Byte_Order) -> (f32, bool) {
+	v, ok := get_u32(b, order)
+	return transmute(f32)v, ok
+}
+get_f64 :: proc(b: []byte, order: Byte_Order) -> (f64, bool) {
+	v, ok := get_u64(b, order)
+	return transmute(f64)v, ok
+}
+
+
+put_u16 :: proc(b: []byte, order: Byte_Order, v: u16) -> bool {
+	if len(b) < 2 {
+		return false
+	}
+	#no_bounds_check if order == .Little {
+		b[0] = byte(v)
+		b[1] = byte(v >> 8)
+	} else {
+		b[0] = byte(v >> 8)
+		b[1] = byte(v)
+	}
+	return true
+}
+put_u32 :: proc(b: []byte, order: Byte_Order, v: u32) -> bool {
+	if len(b) < 4 {
+		return false
+	}
+	#no_bounds_check if order == .Little {
+		b[0] = byte(v)
+		b[1] = byte(v >> 8)
+		b[2] = byte(v >> 16)
+		b[3] = byte(v >> 24)
+	} else {
+		b[0] = byte(v >> 24)
+		b[1] = byte(v >> 16)
+		b[2] = byte(v >> 8)
+		b[3] = byte(v)
+	}
+	return true
+}
+put_u64 :: proc(b: []byte, order: Byte_Order, v: u64) -> bool {
+	if len(b) < 8 {
+		return false
+	}
+	#no_bounds_check if order == .Little {
+		b[0] = byte(v >> 0)
+		b[1] = byte(v >> 8)
+		b[2] = byte(v >> 16)
+		b[3] = byte(v >> 24)
+		b[4] = byte(v >> 32)
+		b[5] = byte(v >> 40)
+		b[6] = byte(v >> 48)
+		b[7] = byte(v >> 56)
+	} else {
+		b[0] = byte(v >> 56)
+		b[1] = byte(v >> 48)
+		b[2] = byte(v >> 40)
+		b[3] = byte(v >> 32)
+		b[4] = byte(v >> 24)
+		b[5] = byte(v >> 16)
+		b[6] = byte(v >> 8)
+		b[7] = byte(v)
+	}
+	return true
+}
+
+put_i16 :: proc(b: []byte, order: Byte_Order, v: i16) -> bool {
+	return put_u16(b, order, u16(v))
+}
+
+put_i32 :: proc(b: []byte, order: Byte_Order, v: i32) -> bool {
+	return put_u32(b, order, u32(v))
+}
+
+put_i64 :: proc(b: []byte, order: Byte_Order, v: i64) -> bool {
+	return put_u64(b, order, u64(v))
+}
+
+
+put_f16 :: proc(b: []byte, order: Byte_Order, v: f16) -> bool {
+	return put_u16(b, order, transmute(u16)v)
+}
+
+put_f32 :: proc(b: []byte, order: Byte_Order, v: f32) -> bool {
+	return put_u32(b, order, transmute(u32)v)
+}
+
+put_f64 :: proc(b: []byte, order: Byte_Order, v: f64) -> bool {
+	return put_u64(b, order, transmute(u64)v)
+}
@@ -0,0 +1,21 @@
+# License
+
+By obtaining, using and/or copying this work, you (the licensee) agree that you have read, understood, and will comply with the following terms and conditions.
+
+Permission to copy, modify, and distribute this software and its documentation, with or without modification, for any purpose and without fee or royalty is hereby granted, provided that you include the following on ALL copies of the software and documentation or portions thereof, including modifications:
+
+The full text of this NOTICE in a location viewable to users of the redistributed or derivative work.
+Any pre-existing intellectual property disclaimers, notices, or terms and conditions. If none exist, the W3C Software Short Notice should be included (hypertext is preferred, text is permitted) within the body of any redistributed or derivative code.
+
+Notice of any changes or modifications to the files, including the date changes were made. (We recommend you provide URIs to the location from which the code is derived.)
+
+# Disclaimers
+
+THIS SOFTWARE AND DOCUMENTATION IS PROVIDED "AS IS," AND COPYRIGHT HOLDERS MAKE NO REPRESENTATIONS OR WARRANTIES, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO, WARRANTIES OF MERCHANTABILITY OR FITNESS FOR ANY PARTICULAR PURPOSE OR THAT THE USE OF THE SOFTWARE OR DOCUMENTATION WILL NOT INFRINGE ANY THIRD PARTY PATENTS, COPYRIGHTS, TRADEMARKS OR OTHER RIGHTS.
+
+COPYRIGHT HOLDERS WILL NOT BE LIABLE FOR ANY DIRECT, INDIRECT, SPECIAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF ANY USE OF THE SOFTWARE OR DOCUMENTATION.
+
+The name and trademarks of copyright holders may NOT be used in advertising or publicity pertaining to the software without specific, written prior permission. Title to copyright in this software and any associated documentation will at all times remain with copyright holders.
+
+# Notes
+This version: http://www.w3.org/Consortium/Legal/2002/copyright-software-20021231
@@ -0,0 +1,374 @@
+package unicode_entity
+/*
+	A unicode entity encoder/decoder
+
+	Copyright 2021 Jeroen van Rijn <nom@duclavier.com>.
+	Made available under Odin's BSD-3 license.
+
+	This code has several procedures to map unicode runes to/from different textual encodings.
+	- SGML/XML/HTML entity
+	-- &#<decimal>;
+	-- &#x<hexadecimal>;
+	-- &<entity name>;   (If the lookup tables are compiled in).
+	Reference: https://www.w3.org/2003/entities/2007xml/unicode.xml	
+
+	- URL encode / decode %hex entity
+	Reference: https://datatracker.ietf.org/doc/html/rfc3986/#section-2.1
+
+	List of contributors:
+		Jeroen van Rijn: Initial implementation.
+*/
+
+import "core:unicode/utf8"
+import "core:unicode"
+import "core:strings"
+
+MAX_RUNE_CODEPOINT :: int(unicode.MAX_RUNE)
+
+write_rune   :: strings.write_rune
+write_string :: strings.write_string
+
+Error :: enum u8 {
+	None = 0,
+	Tokenizer_Is_Nil,
+
+	Illegal_NUL_Character,
+	Illegal_UTF_Encoding,
+	Illegal_BOM,
+
+	CDATA_Not_Terminated,
+	Comment_Not_Terminated,
+	Invalid_Entity_Encoding,
+}
+
+Tokenizer :: struct {
+	r:           rune,
+	w:           int,
+
+	src:         string,
+	offset:      int,
+	read_offset: int,
+}
+
+CDATA_START   :: "<![CDATA["
+CDATA_END     :: "]]>"
+
+COMMENT_START :: "<!--"
+COMMENT_END   :: "-->"
+
+/*
+	Default: CDATA and comments are passed through unchanged.
+*/
+XML_Decode_Option :: enum u8 {
+	/*
+		Do not decode & entities. It decodes by default.
+		If given, overrides `Decode_CDATA`.
+	*/
+	No_Entity_Decode,
+
+	/*
+		CDATA is unboxed.
+	*/
+	Unbox_CDATA,
+
+	/*
+		Unboxed CDATA is decoded as well.
+		Ignored if `.Unbox_CDATA` is not given.
+	*/
+	Decode_CDATA,
+
+	/*
+		Comments are stripped.
+	*/
+	Comment_Strip,
+}
+XML_Decode_Options :: bit_set[XML_Decode_Option; u8]
+
+/*
+	Decode a string that may include SGML/XML/HTML entities.
+	The caller has to free the result.
+*/
+decode_xml :: proc(input: string, options := XML_Decode_Options{}, allocator := context.allocator) -> (decoded: string, err: Error) {
+	context.allocator = allocator
+
+	l := len(input)
+	if l == 0 { return "", .None }
+
+	builder := strings.builder_make()
+	defer strings.builder_destroy(&builder)
+
+	t := Tokenizer{src=input}
+	in_data := false
+
+	loop: for {
+		advance(&t) or_return
+		if t.r < 0 { break loop }
+
+		/*
+			Below here we're never inside a CDATA tag.
+			At most we'll see the start of one, but that doesn't affect the logic.
+		*/
+		switch t.r {
+		case '<':
+			/*
+				Might be the start of a CDATA tag or comment.
+
+				We don't need to check if we need to write a `<`, because if it isn't CDATA or a comment,
+				it couldn't have been part of an XML tag body to be decoded here.
+
+				Keep in mind that we could already *be* inside a CDATA tag.
+				If so, write `>` as a literal and continue.
+			*/
+			if in_data {
+				write_rune(&builder, '<')
+				continue
+			}
+			in_data = _handle_xml_special(&t, &builder, options) or_return
+
+		case ']':
+			/*
+				If we're unboxing _and_ decoding CDATA, we'll have to check for the end tag.
+			*/
+			if in_data {
+				if t.read_offset + len(CDATA_END) < len(t.src) {
+					if string(t.src[t.offset:][:len(CDATA_END)]) == CDATA_END {
+						in_data = false
+						t.read_offset += len(CDATA_END) - 1
+					}
+				}
+				continue
+			} else {
+				write_rune(&builder, ']')
+			}
+
+		case:
+			if in_data && .Decode_CDATA not_in options {
+				/*
+					Unboxed, but undecoded.
+				*/
+				write_rune(&builder, t.r)
+				continue
+			}
+
+			if t.r == '&' {
+				if entity, entity_err := _extract_xml_entity(&t); entity_err != .None {
+					/*
+						We read to the end of the string without closing the entity.
+						Pass through as-is.
+					*/
+					write_string(&builder, entity)
+				} else {
+
+					if .No_Entity_Decode not_in options {
+						if decoded, ok := xml_decode_entity(entity); ok {
+							write_rune(&builder, decoded)
+							continue
+						}
+					}
+
+					/*
+						Literal passthrough because the decode failed or we want entities not decoded.
+					*/
+					write_string(&builder, "&")
+					write_string(&builder, entity)
+					write_string(&builder, ";")
+				}
+			} else {
+				write_rune(&builder, t.r)
+			}
+		}
+	}
+
+	return strings.clone(strings.to_string(builder), allocator), err
+}
+
+advance :: proc(t: ^Tokenizer) -> (err: Error) {
+	if t == nil { return .Tokenizer_Is_Nil }
+	using t
+
+	#no_bounds_check {
+		if read_offset < len(src) {
+			offset = read_offset
+			r, w   = rune(src[read_offset]), 1
+			switch {
+			case r == 0:
+				return .Illegal_NUL_Character
+			case r >= utf8.RUNE_SELF:
+				r, w = utf8.decode_rune_in_string(src[read_offset:])
+				if r == utf8.RUNE_ERROR && w == 1 {
+					return .Illegal_UTF_Encoding
+				} else if r == utf8.RUNE_BOM && offset > 0 {
+					return .Illegal_BOM
+				}
+			}
+			read_offset += w
+			return .None
+		} else {
+			offset = len(src)
+			r = -1
+			return
+		}
+	}
+}
+
+xml_decode_entity :: proc(entity: string) -> (decoded: rune, ok: bool) {
+	entity := entity
+	if len(entity) == 0 { return -1, false }
+
+	switch entity[0] {
+	case '#':
+		base  := 10
+		val   := 0
+		entity = entity[1:]
+
+		if len(entity) == 0 { return -1, false }
+
+		if entity[0] == 'x' || entity[0] == 'X' {
+			base = 16
+			entity = entity[1:]
+		}
+
+		for len(entity) > 0 {
+			r := entity[0]
+			switch r {
+			case '0'..='9':
+				val *= base
+				val += int(r - '0')
+
+			case 'a'..='f':
+				if base == 10 { return -1, false }
+				val *= base
+				val += int(r - 'a' + 10)
+
+			case 'A'..='F':
+				if base == 10 { return -1, false }
+				val *= base
+				val += int(r - 'A' + 10)
+
+			case:
+				return -1, false
+			}
+
+			if val > MAX_RUNE_CODEPOINT { return -1, false }
+			entity = entity[1:]
+		}
+		return rune(val), true
+
+	case:
+		/*
+			Named entity.
+		*/
+		return named_xml_entity_to_rune(entity)
+	}
+}
+
+/*
+	Private XML helper to extract `&<stuff>;` entity.
+*/
+@(private="file")
+_extract_xml_entity :: proc(t: ^Tokenizer) -> (entity: string, err: Error) {
+	assert(t != nil && t.r == '&')
+
+	/*
+		All of these would be in the ASCII range.
+		Even if one is not, it doesn't matter. All characters we need to compare to extract are.
+	*/
+	using t
+
+	length := len(t.src)
+	found  := false
+
+	#no_bounds_check {
+		for read_offset < length {
+			if src[read_offset] == ';' {
+				found = true
+				read_offset += 1
+				break
+			}
+			read_offset += 1
+		}
+	}
+
+	if found {
+		return string(src[offset + 1 : read_offset - 1]), .None
+	}
+	return string(src[offset : read_offset]), .Invalid_Entity_Encoding
+}
+
+/*
+	Private XML helper for CDATA and comments.
+*/
+@(private="file")
+_handle_xml_special :: proc(t: ^Tokenizer, builder: ^strings.Builder, options: XML_Decode_Options) -> (in_data: bool, err: Error) {
+	assert(t != nil && t.r == '<')
+	if t.read_offset + len(CDATA_START) >= len(t.src) { return false, .None }
+
+	if string(t.src[t.offset:][:len(CDATA_START)]) == CDATA_START {
+		t.read_offset += len(CDATA_START) - 1
+
+		if .Unbox_CDATA in options && .Decode_CDATA in options {
+			/*
+				We're unboxing _and_ decoding CDATA
+			*/
+			return true, .None
+		}
+
+		/*
+			CDATA is passed through.
+		*/
+		offset := t.offset
+
+		/*
+			Scan until end of CDATA.
+		*/
+		for {
+			advance(t) or_return
+			if t.r < 0 { return true, .CDATA_Not_Terminated }
+
+			if t.read_offset + len(CDATA_END) < len(t.src) {
+				if string(t.src[t.offset:][:len(CDATA_END)]) == CDATA_END {
+					t.read_offset += len(CDATA_END) - 1
+
+					cdata := string(t.src[offset : t.read_offset])
+	
+					if .Unbox_CDATA in options {
+						cdata = cdata[len(CDATA_START):]
+						cdata = cdata[:len(cdata) - len(CDATA_END)]
+					}
+
+					write_string(builder, cdata)
+					return false, .None
+				}
+			}
+		}
+
+	} else if string(t.src[t.offset:][:len(COMMENT_START)]) == COMMENT_START {
+		t.read_offset += len(COMMENT_START)
+		/*
+			Comment is passed through by default.
+		*/
+		offset := t.offset
+
+		/*
+			Scan until end of Comment.
+		*/
+		for {
+			advance(t) or_return
+			if t.r < 0 { return true, .Comment_Not_Terminated }
+
+			if t.read_offset + len(COMMENT_END) < len(t.src) {
+				if string(t.src[t.offset:][:len(COMMENT_END)]) == COMMENT_END {
+					t.read_offset += len(COMMENT_END) - 1
+
+					if .Comment_Strip not_in options {
+						comment := string(t.src[offset : t.read_offset])
+						write_string(builder, comment)
+					}
+					return false, .None
+				}
+			}
+		}
+
+	}
+	return false, .None
+}
@@ -0,0 +1,76 @@
+package unicode_entity_example
+
+import "core:encoding/xml"
+import "core:strings"
+import "core:mem"
+import "core:fmt"
+import "core:time"
+
+doc_print :: proc(doc: ^xml.Document) {
+	buf: strings.Builder
+	defer strings.builder_destroy(&buf)
+	w := strings.to_writer(&buf)
+
+	xml.print(w, doc)
+	fmt.println(strings.to_string(buf))
+}
+
+_entities :: proc() {
+	doc: ^xml.Document
+	err: xml.Error
+
+	DOC :: #load("../../../../tests/core/assets/XML/unicode.xml")
+
+	OPTIONS  :: xml.Options{
+		flags            = {
+			.Ignore_Unsupported, .Intern_Comments,
+		},
+		expected_doctype = "",
+	}
+
+	parse_duration: time.Duration
+
+	{
+		time.SCOPED_TICK_DURATION(&parse_duration)
+		doc, err = xml.parse(DOC, OPTIONS)
+	}
+	defer xml.destroy(doc)
+
+	doc_print(doc)
+
+	ms := time.duration_milliseconds(parse_duration)
+
+	speed := (f64(1000.0) / ms) * f64(len(DOC)) / 1_024.0 / 1_024.0
+
+	fmt.printf("Parse time: %.2f ms (%.2f MiB/s).\n", ms, speed)
+	fmt.printf("Error: %v\n", err)
+}
+
+_main :: proc() {
+	using fmt
+
+	options := xml.Options{ flags = { .Ignore_Unsupported, .Intern_Comments, .Unbox_CDATA, .Decode_SGML_Entities }}
+
+	doc, _ := xml.parse(#load("test.html"), options)
+
+	defer xml.destroy(doc)
+	doc_print(doc)
+}
+
+main :: proc() {
+	using fmt
+
+	track: mem.Tracking_Allocator
+	mem.tracking_allocator_init(&track, context.allocator)
+	context.allocator = mem.tracking_allocator(&track)
+
+	// _main()
+	_entities()
+
+	if len(track.allocation_map) > 0 {
+		println()
+		for _, v in track.allocation_map {
+			printf("%v Leaked %v bytes.\n", v.location, v.size)
+		}
+	}	
+}
--- a/Show More
+++ b/Show More