coderjeff
coderjeff
Latest modifications check for malformed values and should finish this PR. Will merge in soon. The fbc build date (what is shown in fbc -version) is ultimately determined by the...
Latest force push only to rebase on to current master.
@dkl, I requested review from you for this change. I know you did a lot of work on classes and operator overloading. This change does not exactly follow g++, but...
The advantages for this PR change are small, but notable, I think: - static, in my opinion, is better than non-static overloaded member, since for binary operators, can make operator...
From previous post: > static overloaded operators produce a similar procedure signature as c++ non-static overloaded operator members; Compatible linkage, though not compatible syntax fbc versus c++ Correction: My own...
@dkl, thank you for your insights. Very helpful to keep this moving forward. I started writing tests for fbc/gcc name mangling of member procs, then realized that don't have `thiscall`...
With a few fixups, I was able to compile and run a simple program on windows using `llc.exe`. I don't know what changes would be needed for `clang.exe` on windows...
I'm working on a fix for escaping the procedure names in `symb-mangling.bas:hMangleProc()`.
I think we need to escape global names for windows targets on LLVM. - For asm we always use the mangled name as-is. - For gcc backend we sort of...
@rversteegen > I'm going to also rework this patch to do a runtime fallback to clang instead of hardcoding on win32. FYI, I was thinking of something kind of like...