Syntax error sanity tests.

build.bat

@@ -12,7 +12,6 @@ cl %compile_flags% ..\samples\static_site_generator\static_site_generator.c
 cl %compile_flags% ..\samples\output_parse\output_parse.c
 cl %compile_flags% ..\samples\c_code_generation.c
 cl %compile_flags% ..\samples\node_errors\node_errors.c
-cl %compile_flags% ..\samples\syntax_errors\syntax_errors.c
 echo.
 echo ~~~ Build All Tests ~~~
 cl %compile_flags% ..\tests\sanity_tests.c
@@ -64,9 +63,3 @@ echo ~~~ Running Error Generation Sample ~~~
 pushd build
 node_errors.exe %~dp0\samples\node_errors\node_errors.md
 popd
-
-echo.
-echo ~~~ Running Syntax Error Detection and Reporting Sample ~~~
-pushd build
-for /f "tokens=*" %%f in ('dir /s/b ..\samples\syntax_errors\*.md ^| sort') do syntax_errors.exe %%f
-popd

build.sh

@@ -24,7 +24,6 @@ $CC $compile_flags ../samples/static_site_generator/static_site_generator.c -o s
 $CC $compile_flags ../samples/output_parse/output_parse.c -o output_parse
 $CC $compile_flags ../samples/c_code_generation.c -o c_code_generation
 $CC $compile_flags ../samples/node_errors/node_errors.c -o node_errors
-$CC $compile_flags ../samples/syntax_errors/syntax_errors.c -o syntax_errors
 echo
 echo ~~~ Build All Tests ~~~
 $CC $compile_flags ../tests/sanity_tests.c -o sanity_tests
@@ -64,10 +63,3 @@ echo ~~~ Running Error Generation Sample ~~~
 pushd build
 ./node_errors ../samples/node_errors/node_errors.md
 popd
-
-echo
-echo ~~~ Running Syntax Error Detection and Reporting Sample ~~~
-pushd build
-./syntax_errors ../samples/syntax_errors/*.md
-popd
-

build_with_clang.bat

@@ -14,7 +14,6 @@ clang %compile_flags% ..\samples\static_site_generator\static_site_generator.c -
 clang %compile_flags% ..\samples\output_parse\output_parse.c -o output_parse.exe
 clang %compile_flags% ..\samples\c_code_generation.c -o c_code_generation.exe
 clang %compile_flags% ..\samples\node_errors\node_errors.c -o node_errors.exe
-clang %compile_flags% ..\samples\syntax_errors\syntax_errors.c -o syntax_errors.exe
 echo.
 echo ~~~ Build All Tests ~~~
 clang %compile_flags% ..\tests\sanity_tests.c -o sanity_tests.exe
@@ -66,9 +65,3 @@ echo ~~~ Running Error Generation Sample ~~~
 pushd build
 node_errors.exe %~dp0\samples\node_errors\node_errors.md
 popd
-
-echo.
-echo ~~~ Running Syntax Error Detection and Reporting Sample ~~~
-pushd build
-for /f "tokens=*" %%f in ('dir /s/b ..\samples\syntax_errors\*.md ^| sort') do syntax_errors.exe %%f
-popd

samples/syntax_errors/00.md

@@ -1 +0,0 @@
-{

samples/syntax_errors/01.md

@@ -1 +0,0 @@
-(

samples/syntax_errors/02.md

@@ -1 +0,0 @@
-}

samples/syntax_errors/03.md

@@ -1 +0,0 @@
-'

samples/syntax_errors/04.md

@@ -1,2 +0,0 @@
-a:'''
-multi-line text literal

samples/syntax_errors/05.md

@@ -1 +0,0 @@
-/* asd 

samples/syntax_errors/06.md

@@ -1 +0,0 @@
-/* /* asd */

samples/syntax_errors/07.md

@@ -1 +0,0 @@
-label:@tag {1, 2, 3} /* unterminated comment

samples/syntax_errors/08.md

@@ -1 +0,0 @@
-#include <stdio.h>

samples/syntax_errors/09.md

@@ -1 +0,0 @@
-@"tag" node

samples/syntax_errors/10.md

@@ -1,2 +0,0 @@
-#namespace foo
-{a,#b}

samples/syntax_errors/syntax_errors.c

@@ -1,25 +0,0 @@
-// Sample code to demonstrate syntax error detection and reporting
-
-#include "md.h"
-#include "md.c"
-
-int main(int argument_count, char **arguments)
-{
-    for(int i = 1; i < argument_count; i += 1)
-    {
-        MD_ParseResult parse = MD_ParseWholeFile(MD_S8CString(arguments[i]));
-        if(parse.first_error)
-        {
-            for(MD_Error *error = parse.first_error; error; error = error->next)
-            {
-                MD_OutputError(stderr, error);
-            }
-        }
-        else
-        {
-            printf("No error in file %s\n", arguments[i]);
-        }
-    }
-    
-    return 0;
-}

source/md_impl.c

@@ -1106,7 +1106,7 @@ _MD_ComputeTextLiteralChop(MD_u32 bytes_to_skip, MD_u8 *beg, MD_u8 *end)
     // NOTE(mal): By counting the bytes to chop off text literals and encoding exact token.string and
     //            token.outer_string we delegate the responsibility of generating errors to the parser
     //            That way, we can avoid generating errors during peeks and save them for actual token
-    //            consumption. That allows predictable error ordering.
+    //            consumption. That allows more predictable error ordering.
     MD_u32 result = 0;
     MD_u8 *skip_end = beg + bytes_to_skip;
     MD_u8 *chop_beg = end - bytes_to_skip;
@@ -1178,7 +1178,7 @@ MD_Parse_LexNext(MD_ParseCtx *ctx)
                     {
                         at += 2;
                         token.kind = MD_TokenKind_Comment;
-                        skip_n = chop_n = 2;
+                        skip_n = 2;
                         int counter = 1;
                         for (;at < one_past_last && counter > 0; at += 1)
                         {
@@ -1196,6 +1196,10 @@ MD_Parse_LexNext(MD_ParseCtx *ctx)
                                 }
                             }
                         }
+                        if(counter == 0)
+                        {
+                            chop_n = 2;
+                        }
                     }
                 }
                 if (token.kind == MD_TokenKind_Nil) goto symbol_lex;
@@ -1438,7 +1442,7 @@ _MD_Error(MD_ParseCtx *ctx, MD_Node *node, MD_u8 *at, MD_b32 catastrophic, char
     if(!ctx->catastrophic_error || !prev_error || prev_error->catastrophic == 0)
     {
         MD_Error *error = _MD_PushArray(_MD_GetCtx(), MD_Error, 1);
-        error->node = node;
+        error->node = node ? node : MD_NilNode();
         error->catastrophic = catastrophic;
         error->location = error_loc;
         error->filename = ctx->filename;
@@ -1547,8 +1551,7 @@ _MD_CommentIsSyntacticallyCorrect(MD_Token comment_token)
     MD_String8 inner = comment_token.string;
     MD_String8 outer = comment_token.outer_string;
     MD_b32 incorrect = (MD_StringMatch(MD_StringPrefix(outer, 2), MD_S8Lit("/*"), 0) &&   // C-style comment
-                        (!MD_StringMatch(MD_StringSuffix(outer, 2), MD_S8Lit("*/"), 0) || // Unfinished
+                        (inner.str != outer.str + 2 || inner.str + inner.size != outer.str + outer.size - 2)); // Internally unbalanced
-                         inner.str + inner.size +2 != outer.str + outer.size));           // Internally unbalanced
     result = !incorrect;
     return result;
 }
@@ -1561,6 +1564,8 @@ _MD_ParseOneNode(MD_ParseCtx *ctx)
     MD_ParseResult result = MD_ZERO_STRUCT;
     result.node = MD_NilNode();
 
+    MD_Error *ctx_last_error = ctx->last_error;
+    
     MD_Token token;
     _MD_MemoryZero(&token, sizeof(token));
     
@@ -1722,7 +1727,7 @@ _MD_ParseOneNode(MD_ParseCtx *ctx)
     }
     
     result.bytes_parsed = (MD_u64)(ctx->at - at_first);
-
+    result.first_error = ctx_last_error ? ctx_last_error->next : 0;
     if(!MD_NodeIsNil(result.node))
     {
         result.node->first_tag = first_tag;

tests/sanity_tests.c

@@ -444,7 +444,43 @@ int main(void)
                            MD_StringMatch(parse.node->comment_after, MD_S8Lit(""), 0));
             }
         }
+    }
 
+    Test("Syntax Errors")
+    {
+        struct { char *s; int columns[2]; } tests[] = {
+            {"{", {1}},
+            {"}", {1}},
+            {"'", {1}},
+            {"a:'''\nmulti-line text literal", {3}},
+            {"/* foo", {1}},
+            {"label:@tag {1, 2, 3} /* /* unterminated comment */", {8, 22}},
+            {"#include <stdio.h>", {2}},
+            {"@\"tag\" node", {2}},
+            {"{a,,#b}", {4, 5}},
+        };
+
+        int max_error_count = MD_ArrayCount(tests[0].columns);
+
+        for(int i_test = 0; i_test < MD_ArrayCount(tests); ++i_test)
+        {
+            MD_ParseResult parse = MD_ParseWholeString(MD_S8Lit("test.md"), MD_S8CString(tests[i_test].s));
+
+            MD_b32 columns_match = 1;
+            {
+                MD_Error *e = parse.first_error;
+                for(int i_error = 0; i_error < max_error_count && tests[i_test].columns[i_error]; ++i_error)
+                {
+                    if(!e || e->location.column != tests[i_test].columns[i_error])
+                    {
+                        columns_match = 0;
+                        break;
+                    }
+                    e = e->next;
+                }
+            }
+            TestResult(columns_match);
+        }
     }
 
     return 0;