- fix a bug that bool is captured as int64_t in inliner.
- fix a bug of additional vars in inliner.
- add compiled conversion to greatly reduce launch overhead of inline function. currently only inline cuda/metal kernels in cumm support this.
- fix a bug in simple type analysis
- add metal constant filter in type analysis
- add numpy scalar type support in type analysis
- add attributes and
code_before_func_def
support for apple metal
- add support for param class reload by check id in
sys.module
- fix wrong version.txt
- fix wrong enum typing
- fix missing includes and alias when use impl-only dep with header only
- add extern c to cuda globals
- Add function for inliner code inspect
- add a option to faster check whether perform reload
- fix pre-capture small bug
- Add support for complex inliner and cpu inliner
- Add better capture support for inliner.
- Add PCCM_DISABLE_CODE_CHANGE to disable auto code override for debugging.
- Add include path for pybind only code
- Fix some compiler don't support '~'.
- Fix dynamic decl problem.
- fix small bug in annotation gen.
- add gen_cmake.
- BREAK CHANGE: Change build_pybind api.
- fix some bugs.