ed/gencpp
1
0
Fork 0
mirror of https://github.com/Ed94/gencpp.git synced 2024-11-10 02:54:53 -08:00
Code Issues Packages Projects Releases Wiki Activity

Docs update, parser impl design changes, AST::add fleshed out.

Browse Source
This commit is contained in:
Edward R. Gonzalez 2023-04-22 22:24:55 -04:00
parent 4c4fe98e7e
commit 7ba474069c
9 changed files with 1330 additions and 1637 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

168
Readme.md
View File

@ -2,7 +2,10 @@
An attempt at simple staged metaprogramming for c/c++. An attempt at simple staged metaprogramming for c/c++.
This library is intended for small-to midsized projects. The library is mostly a compositon of code element constructors.
These build up a code AST to then serialize with a file builder.
Intended for small-to midsized projects.
### TOC ### TOC
@ -15,7 +18,7 @@ This library is intended for small-to midsized projects.
* [The four constructors](#there-are-four-sets-of-interfaces-for-code-ast-generation-the-library-provides) * [The four constructors](#there-are-four-sets-of-interfaces-for-code-ast-generation-the-library-provides)
* [Code generation and modification](#code-generation-and-modification) * [Code generation and modification](#code-generation-and-modification)
* [On multithreading](#on-multi-threading) * [On multithreading](#on-multi-threading)
* [On extending with whatever features you want](#on-extending-with-whatever-features-you-want) * [Extending the library](#extending-the-library)
* [Why](#why) * [Why](#why)
* [TODO](#todo) * [TODO](#todo)
@ -27,7 +30,7 @@ Version 1 will have C and a subset of C++ features available to it.
I will generate with this library a C99 or 11 variant when Version 1 is complete. I will generate with this library a C99 or 11 variant when Version 1 is complete.
A single-header version will also be generated. A single-header version will also be generated.
The size target of this library is to stay under 5000 sloc (data & interface code). The size target of this library is to stay under 5-6k sloc (data & interface code).
With the dependency code being under 10000 sloc. (Containers, Memory, String handling, Language bloat) With the dependency code being under 10000 sloc. (Containers, Memory, String handling, Language bloat)
Any dependencies from the zpl library will be exposed manually with using declarations into global scope. Any dependencies from the zpl library will be exposed manually with using declarations into global scope.
@ -70,7 +73,8 @@ The user is given `Code` typed objects that are used to build up the AST.
Example using each construction interface: Example using each construction interface:
#### Upfront ### Upfront
```cpp ```cpp
Code t_uw = def_type( name(uw) ); Code t_uw = def_type( name(uw) );
@ -88,7 +92,7 @@ Code header;
} }
``` ```
#### Incremental ### Incremental
```cpp ```cpp
// Types are done the same with upfront. Incremental does not have a full interface replacment. // Types are done the same with upfront. Incremental does not have a full interface replacment.
@ -106,7 +110,7 @@ Code header = make_struct( name(ArrayHeader) );
} }
``` ```
#### Parse ### Parse
```cpp ```cpp
Code header = parse_struct( code( Code header = parse_struct( code(
@ -122,20 +126,17 @@ Code header = parse_struct( code(
Parse will automatically generate any types that have not been used previously. Parse will automatically generate any types that have not been used previously.
#### Undtyped ### Undtyped
```cpp ```cpp
Code header = untyped_str( Code header = untyped_str( R("
R("struct ArrayHeader struct ArrayHeader
{ {
#define Using_ArrayHeader_Data \ uw Num;
uw Num; \ uw Capacity;
uw Capacity; \
allocator Allocator; allocator Allocator;
Using_ArrayHeader_Data };
};)" )");
);
``` ```
`name` is a helper macro for providing a string literal with its size, intended for the name paraemter of functions. `name` is a helper macro for providing a string literal with its size, intended for the name paraemter of functions.
@ -213,38 +214,42 @@ If in your use case, decide to have exclusive separation or partial separation o
* Macro or template generation : This library is to avoid those, adding support for them adds unnecessary complexity. * Macro or template generation : This library is to avoid those, adding support for them adds unnecessary complexity.
* Vendor provided dynamic dispatch (virtuals) : Roll your own, this library might roll its own vtable/interface generation helpers in the future. * Vendor provided dynamic dispatch (virtuals) : Roll your own, this library might roll its own vtable/interface generation helpers in the future.
* RTTI : This is kinda covered with the last point, but just wanted to emphasize. * RTTI
* Exceptions : Most fo the * Exceptions
* Execution statment validation : Execution expressions are defined using the untyped string API. * Execution statement validation : Execution expressions are defined using the untyped string API.
Keywords in from "Modern C++": Keywords in from "Modern C++":
* constexpr : Great to store compile-time constants, (easier to garanteed when emitted from gentime) * constexpr : Great to store compile-time constants, (easier to garantee when emitted from gentime)
* consteval : Technically fine so long as templates are not used. Need to make sure to execute in moderation. * consteval : Technically fine so long as templates are not used. Need to make sure to execute in moderation.
* constinit : Better than constexpr at doing its job, however, its only c++ 20. * constinit : Better than constexpr at doing its job, however, its only c++ 20.
* export : Useful if c++ modules ever come around to actually being usable. * export : Useful if c++ modules ever come around to actually being usable.
* import : ^^ * import : ^^
* module : ^^ * module : ^^
These features are in as they either are not horrible when used conservatively or are a performance benefit (modules). These features are not horrible when used conservatively, or are a performance benefit (modules).
When it comes to excution statements: When it comes to expressions:
There is no parse API for validating excution statements (possibly will add in the future, but very limited in what it can do).
This reason there isn't one: thats where the can of worms open for parsing validation. There is no support for validating expressions.
For most metaprogramming (espcially for c/c++), expression validation is not necessary, it can be done by the compiler for the runtime program. The reason: thats where the can of worms open for parsing validation. This library would most likey more than double in size with that addition alone.
For most metaprogramming (espcially for C/C++), expression validation is not necessary for metaprogramming, it can be done by the compiler for the runtime program.
Most of the time, the critical complex metaprogramming conundrums are actaully producing the frame of abstractions around the expressions. Most of the time, the critical complex metaprogramming conundrums are actaully producing the frame of abstractions around the expressions.
Thus its not very much a priority to add such a level of complexity to the library when there isn't a high reward or need for it. Thus its not very much a priority to add such a level of complexity to the library when there isn't a high reward or need for it.
To further this point, lets say you do have an error with an execution statment. It will either be caught by the c++ compiler when compiling the target program, or at runtime for the program. To further this point, lets say you do have an error with an expressions composition.
It will either be caught by the c++ compiler when compiling the target program, or at runtime for the program.
* If its not caught by the compiler, the only downside is the error appers on the generated function. Those with knowledge of how that definition was generated know where to find the code that inlined that expression in that file for that definition. * If its not caught by the compiler, the only downside is the error appers on the generated function.
* If its caught at runtime. The expression will be shown in a stack trace if debug symbols are enabled in the generated function body. Yet again those with knowledge of how that definition was generated know where to find the code that inlined that expression. Those with knowledge of how that definition was generated know where to find the code that inlined that expression in that file for that definition.
* If its caught at runtime. The expression will be shown in a stack trace if debug symbols are enabled in the generated function body.
Yet again those with knowledge of how that definition was generated know where to find the code that inlined that expression.
In both these cases will get objectively better debug information than you would normally get on most c++ compilers with complex macros or templates. In both these cases the user will get objectively better debug information than you would normally get on most c++ compilers/editors using complex macros or templates.
### The Data & Interface ### The Data & Interface
As mentioned in [Usage](#Usage), the user is provided Code objects by calling the constructor functions to generate them or find existing matches. As mentioned in [Usage](#usage), the user is provided Code objects by calling the constructor functions to generate them or find existing matches.
The AST is managed by the library and provided the user via its interface prodedures. The AST is managed by the library and provided the user via its interface prodedures.
However, the user may specificy memory configuration. However, the user may specificy memory configuration.
@ -252,22 +257,44 @@ However, the user may specificy memory configuration.
Data layout of AST struct: Data layout of AST struct:
```cpp ```cpp
AST* Parent;
CachedString Name;
CachedString Comment;
union { union {
array(AST*) Entries; AST* ArrStatic[AST::ArrS_Cap];
CachedString Content; Array(AST*) Entries;
StringCached Content;
SpecifierT ArrSpecs[AST::ArrSpecs_Cap];
}; };
AST* Parent;
StringCached Name;
StringCached Comment;
CodeT Type; CodeT Type;
OperatorT Op; OperatorT Op;
bool Readonly; bool Readonly;
u8 _64_Align[23]; bool DynamicEntries;
u8 StaticIndex;
u8 _Align_Pad[6];
``` ```
*`CodeT` is a typedef for `ECode::Type` which has an underlying type of `u32`* *`CodeT` is a typedef for `ECode::Type` which has an underlying type of `u32`*
*`OperatorT` is a typedef for `EOperator::Type` which has an underlying type of `u32`* *`OperatorT` is a typedef for `EOperator::Type` which has an underlying type of `u32`*
AST widths are setup to be AST_POD_Size.
The width dictates how much the static array can hold before it must give way to using an allocated array:
```cpp
constexpr static
uw ArrS_Cap =
( AST_POD_Size
- sizeof(AST*)
- sizeof(StringCached) * 2
- sizeof(CodeT)
- sizeof(OperatorT)
- sizeof(bool) * 2
- sizeof(u8) * 7 )
/ sizeof(AST*);
```
*Ex: If the AST_POD_Size is 256 the capacity of the static array is 26.*
ASTs can be set to readonly by calling Code's lock() member function. ASTs can be set to readonly by calling Code's lock() member function.
Adding comments is always available even if the AST is set to readonly. Adding comments is always available even if the AST is set to readonly.
@ -278,10 +305,7 @@ Data Notes:
* ASTs are wrapped for the user in a Code struct which essentially a warpper for a AST* type. * ASTs are wrapped for the user in a Code struct which essentially a warpper for a AST* type.
* Both AST and Code have member symbols but their data layout is enforced to be POD types. * Both AST and Code have member symbols but their data layout is enforced to be POD types.
* This library treats memory failures as fatal. * This library treats memory failures as fatal.
* The default setup assumes large definition sets may be provided to bodies so AST::Entires are dynamic arrays. * Entires start as a static array, however if it goes over capacity a dynamic array is allocated for the entires.
* They're allocated to arenas currently and are pretty wasteful if they go over their reserve size (its never recycled).
* Most likely will need to implement a dynamic-sized bucket allocation strategy for the entry arrays if memory is getting stressed.
* Otherwise if you are using fixed size entries and your definitions are under 128~512 entries for the body, you may be better of with a fixed-sized array.
* Strings are stored in their own set of arenas. AST constructors use cached strings for names, and content. * Strings are stored in their own set of arenas. AST constructors use cached strings for names, and content.
## There are four sets of interfaces for Code AST generation the library provides ## There are four sets of interfaces for Code AST generation the library provides
@ -300,7 +324,7 @@ Interface :
* def_class * def_class
* def_enum * def_enum
* def_enum_class * def_execution NOTE: This is equivalent to untyped_str, except that its intended for use only in execution scopes.
* def_friend * def_friend
* def_function * def_function
* def_namespace * def_namespace
@ -314,6 +338,7 @@ Interface :
* def_type * def_type
* def_typedef * def_typedef
* def_using * def_using
* def_class_body * def_class_body
* def_enum_body * def_enum_body
* def_function_body NOTE: Use this for operator bodies as well. * def_function_body NOTE: Use this for operator bodies as well.
@ -323,7 +348,7 @@ Interface :
Usage: Usage:
```c++ ```cpp
<name> = def_<function type>( ... ); <name> = def_<function type>( ... );
Code <name> Code <name>
@ -385,6 +410,7 @@ Interface :
* parse_type * parse_type
* parse_typedef * parse_typedef
* parse_using * parse_using
* parse_classes * parse_classes
* parse_enums * parse_enums
* parse_functions * parse_functions
@ -394,10 +420,16 @@ Interface :
* parse_typedefs * parse_typedefs
* parse_usings * parse_usings
The parse API treats any execution scope definitions with no validation and are turned into untyped Code ASTs.
This includes the assignmetn of variables; due to the library not yet supporting c/c++ expression parsing.
The plural variants provide an array of codes, its up to the user to add them to a body AST
(they are not auto-added to a body)
Usage: Usage:
```cpp ```cpp
Code <name> = parse_<function name>( <string with code> ); Code <name> = parse_<function name>( string with code );
Code <name> = def_<function name>( ..., parse_<function name>( Code <name> = def_<function name>( ..., parse_<function name>(
<string with code> <string with code>
@ -411,9 +443,6 @@ Code <name> = make_<function name>( ... )
} }
``` ```
The parse API treats any execution scope definitions with no validation and are turned into untyped Code ASTs.
This includes the assignmetn of variables; due to the library not yet supporting c/c++ expression parsing.
### Untyped constructions ### Untyped constructions
Code ASTs are constructed using unvalidated strings. Code ASTs are constructed using unvalidated strings.
@ -447,17 +476,28 @@ Template metaprogramming in the traditional sense becomes possible with the use
```cpp ```cpp
char const* token_key, token_value, ...; char const* token_key, token_value, ...;
char const* template = txt( char const* template_str = txt(
Code with {key value} to replace with token_values Code with {key} to replace with token_values
... ...
); );
char const* gen_code_str = token_fmt( template, num_tokens, token, ... ); char const* gen_code_str = token_fmt( template, num_tokens, { token_key, token_value }, ... );
Code <name> = parse_<function name>( gen_code_str ); Code <name> = parse_<function name>( gen_code_str );
``` ```
## Extent of operator overload validation
The AST and constructors will be able to validate that the arguments provided for the operator type match the expected form:
* If return type must match a parameter
* If number of parameters is correct
* If added as a member symbol to a class or struct, that operator matches the requirements for the class (types match up)
The user is responsible for making sure the code types provided are correct
and have the desired specifiers assigned to them beforehand.
## Code generation and modification ## Code generation and modification
There are three provided interfaces: There are three provided file interfaces:
* Builder * Builder
* Editor * Editor
@ -483,9 +523,9 @@ Editor and Scanner are disabled by default, use `GEN_FEATURE_EDITOR` and `GEN_FE
All three have the same parameters with exception to remove which only has SymbolInfo and Policy: All three have the same parameters with exception to remove which only has SymbolInfo and Policy:
* SymbolInfo: * SymbolInfo:
* File : The file the symbol resides in. Leave null to indicate to search all files. * File : The file the symbol resides in. Leave null to indicate to search all files. Leave null to indicated all-file search.
* Marker : #define symbol that indicates a location or following signature is valid to manipulate. Leave null to indicate that the signature should only be used. * Marker : #define symbol that indicates a location or following signature is valid to manipulate. Leave null to indicate the signature should only be used.
* Signature : Use a Code symbol to find a valid location to manipulate, can be further filtered with the marker. Leave null to indicate that the marker should only be used. * Signature : Use a Code symbol to find a valid location to manipulate, can be further filtered with the marker. Leave null to indicate the marker should only be used.
* Policy : Additional policy info for completing the request (empty for now) * Policy : Additional policy info for completing the request (empty for now)
* Code : Code to inject if adding, or replace existing code with. * Code : Code to inject if adding, or replace existing code with.
@ -515,24 +555,19 @@ Request queue in both Editor and Scanner are cleared once process_requests compl
## On multi-threading ## On multi-threading
Its intended eventually for this library to support multi-threading at some point, Currently unsupported. The following changes would have to be made:
however for now it does not.
The following changes would have to be made: * Setup static data accesss with fences if more than one thread will generate ASTs ( or keep a different set for each thread)
* Setup static data accesss with fences if more than one thread will generate ASTs
* Make sure local peristent data of functions are also thread local. * Make sure local peristent data of functions are also thread local.
* The builder should be done on a per-thread basis. * The builder should be done on a per-thread basis.
* Due to the design of the editor and scanner, it will most likely * Due to the design of the editor and scanner, it will most likely be best to make each file a job to process request entries on. Receipts should have an an array to store per thread. They can be combined to the final reciepts array when all files have been processed.
be best to make each file a job to process request entries on.
Receipts should have an an array to store per thread.
They can be combined to the final reciepts array when all files have been processed.
For now single-threaded should be pretty quick even without heavy optimizations. For now single-threaded has a bunch of optimization that most likely have done to it and will be more than capable
for the majority of projects this thing is intended for. (IF you use this on Unreal... well your asking for it...)
## On extending with whatever features you want ## Extending the library
This library is relatively very small, and you can easily extend it. This library is relatively very small, and can be extended without much hassle.
The untyped codes and builder/editor/scanner can be technically be used to circumvent The untyped codes and builder/editor/scanner can be technically be used to circumvent
any sort of constrictions the library has with: modern c++, templates, macros, etc. any sort of constrictions the library has with: modern c++, templates, macros, etc.
@ -560,12 +595,13 @@ Thus a rule of thumb is if its a simple definition you can get away with just th
However, if: However, if:
* Your compile time complexity becomes large. * Compile time complexity becomes large.
* You enjoy actually *seeing* the generated code instead of just the error symbols or the pdb symbols. * You enjoy actually *seeing* the generated code instead of just the error symbols or the pdb symbols.
* You value your debugging expereince, and would like to debug your metaprogram, without having to step through the debug version of the compiler (if you even can) * You value your debugging expereince, and would like to debug your metaprogram, without having to step through the debug version of the compiler (if you even can)
* Want to roll your own reflection system * Want to roll your own reflection system
* Want to maintain a series of libraries for internal use, but don't want to deal with manual merging as often when they update. * Want to maintain a series of libraries for internal use, but don't want to deal with manual merging as often when they update.
* Want to create tailored headers for your code or for your libraries since you usually don't need the majority of the code within them. * Want to create tailored headers for your code or for your libraries since you usually don't need the majority of the code within them.
* You just dislike metaprogramming with template expansion
Then this might help you boostrap a toolset todo so. Then this might help you boostrap a toolset todo so.

94
project/Bloat.cpp
View File

@ -2,11 +2,6 @@
#include "Bloat.hpp" #include "Bloat.hpp"
namespace Global
{
bool ShouldShowDebug = false;
}
namespace Memory namespace Memory
{ {
using namespace zpl; using namespace zpl;
@ -41,92 +36,3 @@ namespace Memory
arena_free( & Global_Arena); arena_free( & Global_Arena);
} }
} }
struct TokEntry
{
char const* Str;
sw Length;
};
ZPL_TABLE( static, TokMap, tokmap_, TokEntry )
sw token_fmt_va( char* buf, uw buf_size, char const* fmt, s32 num_tokens, va_list va )
{
char const* buf_begin = buf;
sw remaining = buf_size;
TokMap tok_map;
{
tokmap_init( & tok_map, g_allocator );
s32 left = num_tokens;
while ( left-- )
{
char const* token = va_arg( va, char const* );
char const* value = va_arg( va, char const* );
TokEntry entry
{
value,
str_len(value, (sw)128)
};
u32 key = crc32( token, str_len(token, 32) );
tokmap_set( & tok_map, key, entry );
}
}
sw result = 0;
char current = *fmt;
while ( current )
{
sw len = 0;
while ( current && current != '{' && remaining )
{
*buf = *fmt;
buf++;
fmt++;
current = *fmt;
}
if ( current == '{' )
{
char const* scanner = fmt;
s32 tok_len = 0;
while ( *scanner != '}' )
{
tok_len++;
scanner++;
}
char const* token = fmt;
u32 key = crc32( token, tok_len );
TokEntry value = * tokmap_get( & tok_map, key );
sw left = value.Length;
while ( left-- )
{
*buf = *value.Str;
buf++;
value.Str++;
}
scanner++;
fmt = scanner;
current = *fmt;
}
}
tokmap_clear( & tok_map );
return result;
}

35
project/Bloat.hpp
View File

@ -45,6 +45,8 @@ using zpl::u32;
using zpl::u64; using zpl::u64;
using zpl::uw; using zpl::uw;
using zpl::sw; using zpl::sw;
using zpl::sptr;
using zpl::uptr;
using zpl::Arena; using zpl::Arena;
using zpl::AllocatorInfo; using zpl::AllocatorInfo;
@ -62,6 +64,7 @@ using zpl::arena_allocator;
using zpl::arena_init_from_memory; using zpl::arena_init_from_memory;
using zpl::arena_init_from_allocator; using zpl::arena_init_from_allocator;
using zpl::arena_free; using zpl::arena_free;
using zpl::assert_crash;
using zpl::str_fmt_buf; using zpl::str_fmt_buf;
using zpl::char_is_alpha; using zpl::char_is_alpha;
using zpl::char_is_alphanumeric; using zpl::char_is_alphanumeric;
@ -102,7 +105,7 @@ using zpl::str_len;
#endif #endif
#if defined(__GNUC__) || defined(__clang__) || true #if defined(__GNUC__) || defined(__clang__)
// Supports 0-10 arguments // Supports 0-10 arguments
#define macro_num_args_impl( _0, \ #define macro_num_args_impl( _0, \
_1, _2, _3, _4, _5, _6, _7, _8, _9, _10, \ _1, _2, _3, _4, _5, _6, _7, _8, _9, _10, \
@ -142,7 +145,6 @@ using zpl::str_len;
#define bit( Value_ ) ( 1 << Value_ ) #define bit( Value_ ) ( 1 << Value_ )
#define bitfield_is_equal( Field_, Mask_ ) ( ( (Mask_) & (Field_) ) == (Mask_) ) #define bitfield_is_equal( Field_, Mask_ ) ( ( (Mask_) & (Field_) ) == (Mask_) )
#define ct constexpr
#define forceinline ZPL_ALWAYS_INLINE #define forceinline ZPL_ALWAYS_INLINE
#define print_nl( _) zpl_printf("\n") #define print_nl( _) zpl_printf("\n")
#define ccast( Type_, Value_ ) * const_cast< Type_* >( & (Value_) ) #define ccast( Type_, Value_ ) * const_cast< Type_* >( & (Value_) )
@ -176,16 +178,12 @@ do \
} \ } \
while(0); while(0);
ct char const* Msg_Invalid_Value = "INVALID VALUE PROVIDED"; constexpr
char const* Msg_Invalid_Value = "INVALID VALUE PROVIDED";
namespace Global
{
extern bool ShouldShowDebug;
}
namespace Memory namespace Memory
{ {
ct uw Initial_Reserve = megabytes(10); constexpr uw Initial_Reserve = megabytes(10);
extern Arena Global_Arena; extern Arena Global_Arena;
// #define g_allocator arena_allocator( & Memory::Global_Arena) // #define g_allocator arena_allocator( & Memory::Global_Arena)
@ -198,28 +196,9 @@ namespace Memory
void cleanup(); void cleanup();
} }
sw token_fmt_va( char* buf, uw buf_size, char const* fmt, s32 num_tokens, va_list va );
inline
char const* token_fmt( char const* fmt, sw num_tokens, ... )
{
local_persist thread_local
char buf[ZPL_PRINTF_MAXLEN] = { 0 };
va_list va;
va_start(va, fmt);
token_fmt_va(buf, ZPL_PRINTF_MAXLEN, fmt, num_tokens, va);
va_end(va);
return buf;
}
inline inline
sw log_fmt(char const *fmt, ...) sw log_fmt(char const *fmt, ...)
{ {
if ( Global::ShouldShowDebug == false )
return 0;
sw res; sw res;
va_list va; va_list va;

4
project/Bloat.redef.hpp
View File

@ -1,4 +0,0 @@
/*
Redefine any macros.
Used if you have some code thats getting parsed and don't want it ot get preprocessed.
*/

18
project/Bloat.undef.hpp
View File

@ -1,18 +0,0 @@
/*
Remvoe any macro definitions related to the Bloat header.
*/
#undef bit
#undef bitfield_is_equal
#undef ct
#undef forceinline
#undef print_nl
#undef scast
#undef rcast
#undef pcast
#undef txt
#undef do_once
#undef do_once_start
#undef do_once_end

87
project/Readme.md Normal file
View File

@ -0,0 +1,87 @@
# Documentation
This library is currently in a bootstrapping phase.
Eventually it will have zero dependencies and have its code size severely constricted.
All dependencies are currently held within `Bloat.hpp` and `Bloat.cpp`
All the library code is contained in two files: `gen.hpp` and `gen.cpp`
## Bloat.hpp/cpp
Currently acts as the isolation header for thridparty dependencies along with code not directly related to the library.
Organization:
* ZPL inclusion and selective symbol exposure to global scope.
* Utility macro definitions used throughout the library.
* Global memory arena definition
* Token string formatter
* Formatted and Fatal Logs
The cpp contains the implementation of the global memory arena and the token formmatter.
Any global symbol pollution will be removed when dependencies are intergrated properly into the library.
## gen.hpp
While getting fleshed out, all feature macros are defined on the top of the header.
These macros are:
* `GEN_DEFINE_DSL` : Define the preprocessor DSL for using the library interface
* `GEN_DEFINE_LIBRARY_CORE_CONSTANTS` : Optional typename codes as they are non-standard to C/C++ and not necessary to library usage
* `GEN_ENCORCE_READONLY_AST` : Defines checks in Code when accessing the AST to make sure readonly marked ASTs are not mutated
* `GEN_FEATURE_INCREMENTAL` : Defines the incremental constructors
* `GEN_FEATURE_PARSING` : Defines the parse constructors
* `GEN_FEATURE_EDITOR` : Defines the file editing features for changing definitions based on ASTs
* `GEN_FEATURE_SCANNER` : Defines the file scanning features for generating ASTs
Due to the design of `gen.hpp` to support being written alongside runtime intended code (in the same file), all the code is wrapped in a `gen_time` `#ifdef` and then wrapped further in a `gen` namespace to avoid pollution of the global scope.
*Note: Its possible with the scanner feature to support parsing runtime files that use "generic" macros or identifiers with certain patterns.
This can be used to auto-queue generation of dependent definitions for the symbols used.*
### Organization
log_failure definition : based on whether to always use fatal on all errors
Major enum definitions and their associated functions used with the AST data
* `ECode` : Used to tag ASTs by their type
* `EOperator` : Used to tag operator overloads with thier op type
* `ESpecifier` : Used with specifier ASTs for all specifiers the user may tag an associated AST with.
#### Data Structures
`StringTable` : Hash table for cached strings. (`StringCached` typedef used to denote strings managed by it)
`AST` : The node data strucuture for the code.
`Code` : Wrapper for `AST` with functionality for handling it appropriately.
`TypeTable` : Hash table for cached typename ASTs.
#### Gen Interface
First set of fowards are either backend functions used for various aspects of AST generation or configurating allocators used for different containers.
Interface fowards defined in order of: Upfront, Incremental, Parsing, Untyped.
From there forwards for the File handlers are defined: Builder, Editor, Scanner.
#### Macros
General helper macros are defined along with the optional DSL macros.
#### Constants
Constants including optional ones are defined.
#### Inlines
Inlined functions related to the AST datatype that required forwards for gen interface functions are defined.
## gen.cpp

1114
project/gen.cpp
View File

File diff suppressed because it is too large Load Diff

535
project/gen.hpp
View File

@ -1,384 +1,16 @@
/* /*
gencpp: A simple staged metaprogramming library for C++. gencpp: An attempt at simple staged metaprogramming for c/c++.
The library is mostly a compositon of code element constructors. See Readme.md for more information from the project repository.
These build up a code AST to then serialize with a file builder.
This library is intended for small-to midsize projects. Public Address:
https://github.com/Ed94/gencpp
AST type checking supports only a small subset of c++.
See the 'ECode' namespace and 'gen API' region to see what is supported.
### *WHAT IS NOT PROVIDED*
This library aims to be used in a "orthodox" or minimal C++ workspace.
* Macro or template generation : This library is to avoid those, adding support for them adds unnecessary complexity.
* Vendor provided dynamic dispatch (virtuals) : Roll your own, this library might roll its own vtable/interface generation helpers in the future.
* RTTI
* Exceptions
* Execution statement validation : Execution expressions are defined using the untyped string API.
Keywords in from "Modern C++":
Specifiers:
* constexpr : Great to store compile-time constants, (easier to garanteed when emitted from gentime)
* consteval : Technically fine so long as templates are not used. Need to make sure to execute in moderation.
* constinit : Better than constexpr at doing its job, however, its only c++ 20.
* export : Useful if c++ modules ever come around to actually being usable.
* import : ^^
* module : ^^
These features are in as they either are not horrible when used conservatively or are a performance benefit (modules).
#### When it comes to expressions:
There is no support for validating expressions.
The reason: thats where the can of worms open for parsing validation. This library would most likey more than double in size with that addition alone.
For most metaprogramming (espcially for C/C++), expression validation is not necessary for metaprogramming, it can be done by the compiler for the runtime program.
Most of the time, the critical complex metaprogramming conundrums are actaully producing the frame of abstractions around the expressions.
Thus its not very much a priority to add such a level of complexity to the library when there isn't a high reward or need for it.
To further this point, lets say you do have an error with an expressions composition.
It will either be caught by the c++ compiler when compiling the target program, or at runtime for the program.
* If its not caught by the compiler, the only downside is the error appers on the generated function.
Those with knowledge of how that definition was generated know where to find the code that inlined that expression in that file for that definition.
* If its caught at runtime. The expression will be shown in a stack trace if debug symbols are enabled in the generated function body.
Yet again those with knowledge of how that definition was generated know where to find the code that inlined that expression.
In both these cases the user will get objectively better debug information than you would normally get on most c++ compilers/editors using complex macros or templates.
### The Data & Interface:
The AST is managed by the library and provided the user via its interface prodedures.
Notes:
* The allocator definitions used are exposed to the user incase they want to dictate memory usage
* You'll find the memory handling in `init`, `gen_string_allocator`, `get_cached_string`, `make_code`, and `make_code_entries`.
* ASTs are wrapped for the user in a Code struct which essentially a warpper for a AST* type.
* Both AST and Code have member symbols but their data layout is enforced to be POD types.
* This library treats memory failures as fatal.
* The AST's data union is can either hold a static array of AST*'s, a dynmaic array if the the static capacity is not enough, or a cached string content.
* The dynamic array is allocated to arenas currently and are pretty wasteful if they go over their reserve size (its never recycled).
Data layout of AST struct:
union {
AST* ArrStatic[ArrS_Cap];
Array(AST*) Entries;
StringCached Content;
};
AST* Parent;
StringCached Name;
StringCached Comment;
CodeT Type;
OperatorT Op;
bool Readonly;
bool DynamicEntries;
u8 StaticIndex;
u8 _Align_Pad[6];
*`CodeT` is a typedef for `ECode::Type` which is the type of the enum.*
AST widths are setup to be AST_POD_Size.
The width dictates how much the static array can hold before it must give way to using an allocated array:
constexpr static
u32 ArrS_Cap =
( AST_POD_Size
- sizeof(AST*)
- sizeof(StringCached) * 2
- sizeof(CodeT)
- sizeof(OperatorT)
- sizeof(bool) * 2
- sizeof(u8) * 7 )
/ sizeof(AST*);
Ex: If the AST_POD_Size is 256 the capacity of the static array is 26.
ASTs can be set to readonly by calling Code's lock() member function.
Adding comments is always available even if the AST is set to readonly.
#### Misc
* The allocator definitions used are exposed to the user incase they want to dictate memory usage
* You'll find the memory handling in `init`, `gen_string_allocator`, `get_cached_string`, `make_code`, and `make_code_entries`.
* ASTs are wrapped for the user in a Code struct which essentially a warpper for a AST* type.
* Both AST and Code have member symbols but their data layout is enforced to be POD types.
* This library treats memory failures as fatal.
* The default setup assumes large definition sets may be provided to bodies so AST::Entires are dynamic arrays.
* They're allocated to arenas currently and are pretty wasteful if they go over their reserve size (its never recycled).
* Most likely will need to implement a dynamic-sized bucket allocation strategy for the entry arrays if memory is getting stressed.
* Otherwise if you are using fixed size entries and your definitions are under 128~512 entries for the body, you may be better of with a fixed-sized array.
* Strings are stored in their own set of arenas. AST constructors use cached strings for names, and content.
### There are four sets of interfaces for Code AST generation the library provides
* Upfront
* Incremental
* Parsing
* Untyped
### Upfront Construction
All component ASTs must be previously constructed, and provided on creation of the code AST.
The construction will fail and return InvalidCode otherwise.
Interface :
* def_class
* def_enum
* def_execution NOTE: This is equivalent to untyped_str, except that its intended for use only in execution scopes.
* def_friend
* def_function
* def_namespace
* def_operator
* def_param
* def_params
* def_specifier
* def_specifiers
* def_struct
* def_variable
* def_type
* def_typedef
* def_using
* def_class_body
* def_enum_body
* def_function_body NOTE: Use this for operator bodies as well.
* def_global_body
* def_namespace_body
* def_struct_body
Usage:
Code <name> = def_<function type>( ... );
Code <name>
{
...
<name> = def_<function name>( ... );
}
### Incremental construction
A Code ast is provided but only completed upfront if all components are provided.
Components are then added using the AST API for adding ASTs:
* code.add( AST* ) // Adds AST with validation.
* code.add_entry( AST* ) // Adds AST entry without validation.
Code ASTs may be explictly validated at anytime using Code's check() member function.
Interface :
* make_class
* make_enum
* make_enum_class
* make_function
* make_global_body
* make_namespace
* make_operator
* make_params
* make_specifiers
* make_struct
Usage:
Code <name> = make_<function name>( ... )
{
<name>->add( ... );
...
}
### Parse construction
A string provided to the API is parsed for the intended language construct.
Interface :
* parse_class
* parse_enum
* parse_friend
* parse_function
* parse_global_body
* parse_namespace
* parse_operator
* parse_struct
* parse_strucs
* parse_variable
* parse_type
* parse_typedef
* parse_using
* parse_classes
* parse_enums
* parse_functions
* parse_namespaces
* parse_operators
* parse_variables
* parse_typedefs
* parse_usings
The parse API treats any execution scope definitions with no validation and are turned into untyped Code ASTs.
This includes the assignmetn of variables; due to the library not yet supporting c/c++ expression parsing.
The pluralvariants provide an array of codes, its up to the user to add them to a body AST
(they are not auto-added to a body)
Usage:
Code <name> = parse_<function name>( string with code );
Code <name> = def_<function name>( ..., parse_<function name>(
<string with code>
));
Code <name> = make_<function name>( ... )
{
<name>->add( parse_<function name>(
<string with code>
));
}
### Untyped constructions
Code ASTs are constructed using unvalidated strings.
Interface :
* token_fmt
* untyped_str
* untyped_fmt
* untyped_token_fmt
During serialization any untyped Code AST is has its string value directly injected inline of
whatever context the content existed as an entry within.
Even though thesee are not validated from somewhat correct c/c++ syntax or components, it doesn't mean that
Untyped code can be added as any component of a Code AST:
* Untyped code cannot have children, thus there cannot be recursive injection this way.
* Untyped code can only be a child of a parent of body AST, or for values of an assignment (ex: variable assignment).
These restrictions help prevent abuse of untyped code to some extent.
Usage Conventions:
```
Code <name> = def_varaible( <type>, <name>, untyped_<function name>(
<string with code>
));
```
Template metaprogramming in the traditional sense becomes possible with the use of `token_fmt` and parse constructors:
```
char const* token_key, token_value;
char const* template = txt(
Code with {key value} to replace with token_values
...
);
char const* gen_code_str = token_fmt( template, num_tokens, token, ... );
Code <name> = parse_<function name>( gen_code_str );
```
## Extent of operator overload validation:
The AST and constructors will be able to validate that the arguments provided for the operator type match the expected form:
* If return type must match a parameter
* If number of parameters is correct
* If added as a member symbol to a class or struct, that operator matches the requirements for the class (types match up)
The user is responsible for making sure the code types provided are correct
and have the desired specifiers assigned to them beforehand.
## Code generation and modification
There are three provided interfaces:
* Builder
* Editor
* Scanner
Editor and Scanner are disabled by default, use GEN_FEATURE_EDITOR and GEN_FEATURE_SCANNER to enable them.
### Builder is a similar object to the jai language's string_builder.
* The purpose of it is to generate a file.
* A file is specified and opened for writting using the open( file_path) ) fucntion.
* The code is provided via print( code ) function will be seralized to its buffer.
* When all seralization is finished, use the write() comamnd to write the buffer to the file.
### Editor is for editing a series of files based on a set of requests provided to it.
* The purpose is to overrite a specific file, it places its contents in a buffer to scan.
* Requests are populated using the following interface:
* add : Add code.
* remove : Remove code.
* replace: Replace code.
All three have the same parameters with exception to remove which only has SymbolInfo and Policy:
* SymbolInfo:
* File : The file the symbol resides in.
Leave null to indicate to search all files.
* Marker : #define symbol that indicates a location or following signature is valid to manipulate.
Leave null to indicate that the signature should only be used.
* Signature : Use a Code symbol to find a valid location to manipulate, can be further filtered with the marker.
Leave null to indicate that the marker should only be used.
* Policy : Additional policy info for completing the request (empty for now)
* Code : Code to inject if adding, or replace existing code with.
Additionally if GEN_FEATURE_EDITOR_REFACTOR is defined, refactor( file_path, specification_path ) wil be made available.
Refactor is based of the refactor library and uses its interface.
It will on call add a request to the queue to run the refactor script on the file.
### Scanner allows the user to generate Code ASTs by reading files.
* The purpose is to grab definitions to generate metadata or generate new code from these definitions.
* Requests are populated using the add( SymbolInfo, Policy ) function.
The symbol info is the same as the one used for the editor. So is the case with Policy.
The file will only be read from, no writting supported.
One great use case is for example: generating the single-header library for gencpp!
### Additional Info (Editor and Scanner)
When all requests have been populated, call process_requests().
It will provide an output of receipt data of the results when it completes.
Files may be added to the Editor and Scanner additionally with add_files( num, files ).
This is intended for when you have requests that are for multiple files.
Request queue in both Editor and Scanner are cleared once process_requests completes.
### On multi-threading:
Currently supported but want to. The following changes would have to be made:
* Setup static data accesss with fences if more than one thread will generate ASTs ( or keep a different set for each thread)
* Make sure local peristent data of functions are also thread local.
* The builder should be done on a per-thread basis.
* Due to the design of the editor and scanner, it will most likely
be best to make each file a job to process request entries on.
Receipts should have an an array to store per thread.
They can be combined to the final reciepts array when all files have been processed.
For now single-threaded has a bunch of optimization that most likely have done to it and will be more than capable
for the majority of projects this thing is intended for. (IF you use this on Unreal... well your asking for it...)
### Notes on extending with whatever features you want
This library is very small, and you can easily extend it.
The untyped codes and builder/editor/scanner can be technically be used to circumvent
any sort of constrictions the library has with: modern c++, templates, macros, etc.
Typical use case is for getting define constants an old C/C++ library with the scanner:
Code parse_defines() can emit a custom code AST with Macro_Constant type.
Another would be getting preprocessor or template metaprogramming Codes from Unreal Engine definitions.
*/ */
#pragma once #pragma once
#include "Bloat.hpp" #include "Bloat.hpp"
// Temporarily here for debugging purposes. // Temporarily here for debugging purposes.
#define GEN_BAN_CPP_TEMPLATES
#define GEN_DEFINE_DSL #define GEN_DEFINE_DSL
#define GEN_DEFINE_LIBRARY_CODE_CONSTANTS #define GEN_DEFINE_LIBRARY_CODE_CONSTANTS
// #define GEN_DONT_USE_FATAL // #define GEN_DONT_USE_FATAL
@ -395,21 +27,17 @@ namespace gen
{ {
using LogFailType = sw(*)(char const*, ...); using LogFailType = sw(*)(char const*, ...);
# ifdef GEN_BAN_CPP_TEMPLATES
# define template static_assert("Templates are banned within gen_time scope blocks")
# endif
// By default this library will either crash or exit if an error is detected while generating codes. // By default this library will either crash or exit if an error is detected while generating codes.
// Even if set to not use fatal, fatal will still be used for memory failures as the library is unusable when they occur. // Even if set to not use fatal, fatal will still be used for memory failures as the library is unusable when they occur.
# ifdef GEN_DONT_USE_FATAL # ifdef GEN_DONT_USE_FATAL
ct LogFailType log_failure = log_fmt; constexpr LogFailType log_failure = log_fmt;
# else # else
ct LogFailType log_failure = fatal; constexpr LogFailType log_failure = fatal;
# endif # endif
namespace ECode namespace ECode
{ {
# define Define_Types \ # define Define_Types \
Entry( Untyped ) \ Entry( Untyped ) \
Entry( Access_Public ) \ Entry( Access_Public ) \
Entry( Access_Protected ) \ Entry( Access_Protected ) \
@ -448,9 +76,9 @@ namespace gen
enum Type : u32 enum Type : u32
{ {
# define Entry( Type ) Type, # define Entry( Type ) Type,
Define_Types Define_Types
# undef Entry # undef Entry
Num_Types, Num_Types,
Invalid Invalid
@ -461,15 +89,15 @@ namespace gen
{ {
static static
char const* lookup[Num_Types] = { char const* lookup[Num_Types] = {
# define Entry( Type ) txt( Type ), # define Entry( Type ) txt( Type ),
Define_Types Define_Types
# undef Entry # undef Entry
}; };
return lookup[ type ]; return lookup[ type ];
} }
# undef Define_Types #undef Define_Types
} }
using CodeT = ECode::Type; using CodeT = ECode::Type;
@ -480,8 +108,8 @@ namespace gen
Class Class
}; };
ct EnumT EnumClass = EnumT::Class; constexpr EnumT EnumClass = EnumT::Class;
ct EnumT EnumRegular = EnumT::Regular; constexpr EnumT EnumRegular = EnumT::Regular;
enum class UsingT : u8 enum class UsingT : u8
{ {
@ -489,12 +117,12 @@ namespace gen
Namespace Namespace
}; };
ct UsingT UsingRegular = UsingT::Regular; constexpr UsingT UsingRegular = UsingT::Regular;
ct UsingT UsingNamespace = UsingT::Namespace; constexpr UsingT UsingNamespace = UsingT::Namespace;
namespace EOperator namespace EOperator
{ {
# define Define_Operators \ # define Define_Operators \
Entry( Assign, = ) \ Entry( Assign, = ) \
Entry( Assign_Add, += ) \ Entry( Assign_Add, += ) \
Entry( Assign_Subtract, -= ) \ Entry( Assign_Subtract, -= ) \
@ -539,9 +167,9 @@ namespace gen
enum Type : u32 enum Type : u32
{ {
# define Entry( Type_, Token_ ) Type_, # define Entry( Type_, Token_ ) Type_,
Define_Operators Define_Operators
# undef Entry # undef Entry
Comma, Comma,
Num_Ops, Num_Ops,
@ -556,32 +184,32 @@ namespace gen
local_persist local_persist
char const* lookup[ Num_Ops ] = { char const* lookup[ Num_Ops ] = {
# define Entry( Type_, Token_ ) txt(Token_), # define Entry( Type_, Token_ ) txt(Token_),
Define_Operators Define_Operators
# undef Entry # undef Entry
"," ","
}; };
return lookup[ op ]; return lookup[ op ];
} }
# undef Define_Operators #undef Define_Operators
} }
using OperatorT = EOperator::Type; using OperatorT = EOperator::Type;
namespace ESpecifier namespace ESpecifier
{ {
# if defined(ZPL_SYSTEM_WINDOWS) #if defined(ZPL_SYSTEM_WINDOWS)
# define API_Export_Code __declspec(dllexport) # define API_Export_Code __declspec(dllexport)
# define API_Import_Code __declspec(dllimport) # define API_Import_Code __declspec(dllimport)
# define API_Keyword __declspec # define API_Keyword __declspec
# elif defined(ZPL_SYSTEM_MACOS) #elif defined(ZPL_SYSTEM_MACOS)
# define API_Export_Code __attribute__ ((visibility ("default"))) # define API_Export_Code __attribute__ ((visibility ("default")))
# define API_Import_Code __attribute__ ((visibility ("default"))) # define API_Import_Code __attribute__ ((visibility ("default")))
# define API_Keyword __attribute__ # define API_Keyword __attribute__
# endif #endif
# define Define_Specifiers \ #define Define_Specifiers \
Entry( API_Import, API_Export_Code ) \ Entry( API_Import, API_Export_Code ) \
Entry( API_Export, API_Import_Code ) \ Entry( API_Export, API_Import_Code ) \
Entry( Attribute, "You cannot stringize an attribute this way" ) \ Entry( Attribute, "You cannot stringize an attribute this way" ) \
@ -614,9 +242,9 @@ namespace gen
enum Type : u32 enum Type : u32
{ {
# define Entry( Specifier, Code ) Specifier, # define Entry( Specifier, Code ) Specifier,
Define_Specifiers Define_Specifiers
# undef Entry # undef Entry
Num_Specifiers, Num_Specifiers,
Invalid, Invalid,
@ -628,9 +256,9 @@ namespace gen
{ {
local_persist local_persist
char const* lookup[ Num_Specifiers ] = { char const* lookup[ Num_Specifiers ] = {
# define Entry( Spec_, Code_ ) txt(Code_), # define Entry( Spec_, Code_ ) txt(Code_),
Define_Specifiers Define_Specifiers
# undef Entry # undef Entry
}; };
return lookup[ specifier ]; return lookup[ specifier ];
@ -661,7 +289,7 @@ namespace gen
return Invalid; return Invalid;
} }
# undef Define_Specifiers #undef Define_Specifiers
} }
using SpecifierT = ESpecifier::Type; using SpecifierT = ESpecifier::Type;
@ -674,7 +302,7 @@ namespace gen
using StringCached = char const*; using StringCached = char const*;
// Desired width of the AST data structure. // Desired width of the AST data structure.
ct u32 AST_POD_Size = 256; constexpr u32 AST_POD_Size = 256;
// TODO: If perf needs it, convert layout an SOA format. // TODO: If perf needs it, convert layout an SOA format.
/* /*
@ -688,7 +316,7 @@ namespace gen
*/ */
struct AST struct AST
{ {
# pragma region Member Functions #pragma region Member Functions
// Used with incremental constructors // Used with incremental constructors
// Adds and checks entries to see if they are valid additions the type of ast. // Adds and checks entries to see if they are valid additions the type of ast.
@ -702,7 +330,6 @@ namespace gen
return Entries[0]; return Entries[0];
} }
forceinline
bool check(); bool check();
AST* duplicate(); AST* duplicate();
@ -784,13 +411,13 @@ namespace gen
forceinline forceinline
bool typename_is_ptr() bool typename_is_ptr()
{ {
zpl::assert_crash("not implemented"); assert_crash("not implemented");
} }
forceinline forceinline
bool typename_is_ref() bool typename_is_ref()
{ {
zpl::assert_crash("not implemented"); assert_crash("not implemented");
} }
forceinline forceinline
@ -833,7 +460,8 @@ namespace gen
} }
String to_string() const; String to_string() const;
# pragma endregion Member Functions #pragma endregion Member Functions
constexpr static constexpr static
uw ArrS_Cap = uw ArrS_Cap =
( AST_POD_Size ( AST_POD_Size
@ -848,7 +476,7 @@ namespace gen
constexpr static constexpr static
uw ArrSpecs_Cap = ArrS_Cap * (sizeof(AST*) / sizeof(SpecifierT)); uw ArrSpecs_Cap = ArrS_Cap * (sizeof(AST*) / sizeof(SpecifierT));
# define Using_Code_POD \ # define Using_Code_POD \
union { \ union { \
AST* ArrStatic[AST::ArrS_Cap]; \ AST* ArrStatic[AST::ArrS_Cap]; \
Array(AST*) Entries; \ Array(AST*) Entries; \
@ -871,11 +499,11 @@ namespace gen
struct CodePOD struct CodePOD
{ {
Using_Code_POD Using_Code_POD
# undef Using_CodePOD; # undef Using_CodePOD;
}; };
ct sw size_AST = sizeof(AST); constexpr sw size_AST = sizeof(AST);
ct sw size_POD = sizeof(CodePOD); constexpr sw size_POD = sizeof(CodePOD);
// Its intended for the AST to have equivalent size to its POD. // Its intended for the AST to have equivalent size to its POD.
// All extra functionality within the AST namespace should just be syntatic sugar. // All extra functionality within the AST namespace should just be syntatic sugar.
@ -892,12 +520,13 @@ namespace gen
*/ */
struct Code struct Code
{ {
# pragma region Statics #pragma region Statics
// Used internally for the most part to identify invaidly generated code. // Used internally for the most part to identify invaidly generated code.
static Code Invalid; static Code Invalid;
# pragma endregion Statics #pragma endregion Statics
# pragma region Member Functions #pragma region Member Functions
inline
Code body() Code body()
{ {
if ( ast == nullptr ) if ( ast == nullptr )
@ -1006,14 +635,14 @@ namespace gen
return ast; return ast;
} }
# pragma endregion Member Functions #pragma endregion Member Functions
AST* ast; AST* ast;
}; };
static_assert( sizeof(Code) == sizeof(AST*), "ERROR: Code is not POD" ); static_assert( sizeof(Code) == sizeof(AST*), "ERROR: Code is not POD" );
// Used when the its desired when omission is allowed in a definition. // Used when the its desired when omission is allowed in a definition.
ct Code NoCode = { nullptr }; constexpr Code NoCode = { nullptr };
// extern const Code InvalidCode; // extern const Code InvalidCode;
/* /*
@ -1117,8 +746,8 @@ namespace gen
# endif # endif
# pragma endregion Incremental # pragma endregion Incremental
# pragma region Parsing #pragma region Parsing
# ifdef GEN_FEATURE_PARSING #ifdef GEN_FEATURE_PARSING
Code parse_class ( s32 length, char const* class_def ); Code parse_class ( s32 length, char const* class_def );
Code parse_enum ( s32 length, char const* enum_def ); Code parse_enum ( s32 length, char const* enum_def );
Code parse_execution ( s32 length, char const* exec_def ); Code parse_execution ( s32 length, char const* exec_def );
@ -1143,14 +772,30 @@ namespace gen
s32 parse_variables ( s32 length, char const* vars_def, Code* out_var_codes ); s32 parse_variables ( s32 length, char const* vars_def, Code* out_var_codes );
s32 parse_typedefs ( s32 length, char const* typedef_def, Code* out_typedef_codes ); s32 parse_typedefs ( s32 length, char const* typedef_def, Code* out_typedef_codes );
s32 parse_usings ( s32 length, char const* usings_def, Code* out_using_codes ); s32 parse_usings ( s32 length, char const* usings_def, Code* out_using_codes );
# endif #endif
# pragma endregion Parsing #pragma endregion Parsing
#pragma region Untyped text
sw token_fmt_va( char* buf, uw buf_size, char const* fmt, s32 num_tokens, va_list va );
inline
char const* token_fmt( char const* fmt, sw num_tokens, ... )
{
local_persist thread_local
char buf[ZPL_PRINTF_MAXLEN] = { 0 };
va_list va;
va_start(va, fmt);
token_fmt_va(buf, ZPL_PRINTF_MAXLEN, fmt, num_tokens, va);
va_end(va);
return buf;
}
# pragma region Untyped text
Code untyped_str ( s32 length, char const* str); Code untyped_str ( s32 length, char const* str);
Code untyped_fmt ( char const* fmt, ... ); Code untyped_fmt ( char const* fmt, ... );
Code untyped_token_fmt( char const* fmt, s32 num_tokens, ... ); Code untyped_token_fmt( char const* fmt, s32 num_tokens, ... );
# pragma endregion Untyped text #pragma endregion Untyped text
struct Builder struct Builder
{ {
@ -1237,7 +882,7 @@ namespace gen
}; };
#endif #endif
#ifdef GEN_FEATURE_SCANNER # ifdef GEN_FEATURE_SCANNER
struct Scanner struct Scanner
{ {
struct RequestEntry struct RequestEntry
@ -1266,7 +911,7 @@ namespace gen
bool process_requests( Array(Receipt) out_receipts ); bool process_requests( Array(Receipt) out_receipts );
}; };
#endif # endif
#pragma endregion Gen Interface #pragma endregion Gen Interface
} }
@ -1432,18 +1077,18 @@ namespace gen
// or string handling while constructing or serializing. // or string handling while constructing or serializing.
// Change them to suit your needs. // Change them to suit your needs.
ct s32 InitSize_DataArrays = 16; constexpr s32 InitSize_DataArrays = 16;
ct s32 InitSize_StringTable = megabytes(4); constexpr s32 InitSize_StringTable = megabytes(4);
ct s32 InitSize_TypeTable = megabytes(4); constexpr s32 InitSize_TypeTable = megabytes(4);
ct s32 CodePool_NumBlocks = 4096; constexpr s32 CodePool_NumBlocks = 4096;
ct s32 InitSize_CodeEntiresArray = 512; constexpr s32 InitSize_CodeEntiresArray = 512;
ct s32 SizePer_CodeEntriresArena = megabytes(16); constexpr s32 SizePer_CodeEntriresArena = megabytes(16);
ct s32 SizePer_StringArena = megabytes(32); constexpr s32 SizePer_StringArena = megabytes(32);
ct s32 MaxNameLength = 128; constexpr s32 MaxNameLength = 128;
ct s32 MaxUntypedStrLength = kilobytes(640); constexpr s32 MaxUntypedStrLength = kilobytes(640);
ct s32 StringTable_MaxHashLength = kilobytes(1); constexpr s32 StringTable_MaxHashLength = kilobytes(1);
// Predefined Codes. Are set to readonly and are setup during gen::init() // Predefined Codes. Are set to readonly and are setup during gen::init()
@ -1505,6 +1150,12 @@ namespace gen
inline inline
bool AST::add_param( AST* type, s32 length, char const* name ) bool AST::add_param( AST* type, s32 length, char const* name )
{ {
if ( Type != ECode::Function )
{
log_failure( "gen::AST::add_param: this AST is not a function - %s", debug_str() );
return Code::Invalid;
}
if ( length <= 0 ) if ( length <= 0 )
{ {
log_failure( "gen::AST::add_param: Invalid name length provided - %d", length ); log_failure( "gen::AST::add_param: Invalid name length provided - %d", length );

0
project/gen.singleheader.cpp
View File