I will most likely need to refactor some of the components & dependencies files to get the desired gneration implementation the way I want.
Specficially I want to be able to eliminate macros I'm using for enums and common patterns for implmeentation of the data structures.
When it comes to the cpp files, I may leave those alone as the macros largely help ith readability.
Replacing those macros is expensive and most likely not worth it.
The macros under consideration with replacing using the library bootstrap are:
* Define_Types
* Define_Operators
* Define_Specifiers
* GEN_Define_Attribute_Tokens
* Define_CodeType
* Using_Code
* Define_TokType
* def_constant_spec ?
* Helper Macros for def_**_body functions ?
* AST unallowed types case macros?
(The last three I'm unsure about as they work fine, and neeeding the debugger steps there is a rare scenario...)
The enums could be manually generated and have its fields derived from a CSV (which is what genc is currently doing).
This would allow the user to specify custom attribute macros as well with greater ease.
I may maually inline ProcessModuleFlags, as its not even necessary as any specific symbol with a module flag will only use the export value. Import is only used on modules themselves (from what I can tell).
The Parser::lex function could be offloaded to its own file in case the user wants to swap the entire thing out.
(Most likely may want to for various purposes)
The problem with extracting any definitions out of a component file currently is that will lead to splintering that componnet to multiple other components.
This is necessary as the proper scanner is not implemented yet (only a reduimentary scan_file proc is made so far).
Interfaces are assumed to have the public access specifier for their content (if its a class definition)
Started to prepare to segement library code into more files (Less scrolling, need for bootstrapping a tailored version + single header support anyway).
- Renamed macro gen_time to GEN_TIME
- Moved scanner and editor to their own headers, I'm going to consider them extensions.
- I'm preparing to setup the library to build on multiple compiler platforms: clang, gcc, msvc.
I'm rewritting it the way I'd like to learn it.
- I want to use csv parsing heavily with the library so I'm just going to add it to the scanner.
- Globaly memory allocator moved to regular gen header/source as its something really just made for the library.
- Some small refactors to macros
- The parser was updated to support tokenizing preprocessor directives.
- The purpose is based off intuition that it will be required for the scanner.
- Added support for anonymous structs.
- Gave Token_Fmt::token_map its own static memory.
- Minor natvis fix for CodeBody
- Renamed ESpecifier::Static_Member to just Static (acts as a general use case) specifier option
- Setup the lex token array with a configurable arena allocator.
Two major things left before V0.3-4:
- Attribute and Module parisng support with base case test
- AST serializtaion strings get a dedicated slag allocator.
- Got rid of dynamic arrays for body entries, were using links only.
- Halfed the size of the ast from 256 to 128 bytes.
- Fields for different ast types are not accessed directly. Each type has a unique filtered AST for ease of use.
Also upated the gencpp.refactor script with almost all relevant symbols.
gen.undef.macros.hpp also filled out
Ready to complete gencpp related todos left in implementation...
- Made a better global allocator for the process.
- Some small fixes to gen.hpp, removed clear_code_memory as I'm designing this library to for now never free any memory.
- Fixes to memory usage for cached strings
- Added missing verification for attributes in some upfront constructors. Added attribute param for def_type procedure.
- Started to use internal and global keywords in gen.cpp for associated definitions
- Progress toward getting the parsing constructors to support operator definitions.
- There was an *attempt* to get parse_type to support parsing function types. Its not tested yet....
- Its not an nice setup, there is no validation of parameters, problably will add that in the future.
- Readonly overcompilcates things for the scope of this project. I'm avoding const correctness to avoid mental overhead.
- Indentation was removed as it still required a formatting pass after, and the only significant thing needed was the newlines.
- Removed some opinionated takes from readme, trying to keep it straight to the point.
- Used def_execution more in array and ring defs (was using untyped_str when could have been using execution...)
token_fmt returns StrC now. Did some fixes so it can properly function now!
Lots of changes to the Array non-parsed implementation.
It shows no errors in the edtitor static analysis, however I haven't tried it yet.
Unfortunately have some friction with how I would like to set it up like with everything in C++...
Can't inherit from ArrayBase to provide the common members to all Array defs, without introducing constructor complexity.
I may be open to using constructors in the future, depends on how I feel when trying out the context pattern used in jai and odin...
Decided not to support the incremental API, its not necessary as the ergonomics are not that big a deal.
Got operators to pass the sanity base cases, which means now all upfront constructors pass the base cases!
Next up is getting it to pass the array container generation.
Adjusted upfront api a bit, def_using no longer handles the namespace case, a sepearate function: def_using_namespace now does.
Mostly fixes to the seralization as I generate more cases to find the bugs.
