Closed falkTX closed 2 years ago
Thanks, but, y'know, we'd both be a lot less annoyed by a lot of things if you'd just build the damned library normally and not #include the implementation in a different bloody programming language for no particular reason ;)
One reason is for having a safe namespace where I can use 2 or more pugl versions in the same process. I am using pugl on the host side for a few little things, which can directly conflict with pugl from plugins. Gets specially annoying in debug builds where symbol conflicts can easily get in the way.
Fair enough, although I feel like if that's an issue, it means plugins have a sloppy public symbol table that should be fixed anyway...
developers can be lazy bastards though ;) since my usecase is for a plugin framework, I go a bit out of the way to ensure proper namespacing. gets quite annoying on macOS though...
I guess if you mostly live in C++ land, that's more or less viable. In C land, violating the integrity/privacy of compilation units in particular will end very not-well (in a sense it's removing proper namespacing). This is why I don't support that for any other library, but Pugl is a bit of a weird case.
Anyway!
Merged as decde99, thanks.
Similar to #87 but applies to win32 code now. Warnings reported by mingw:
maybe the macro should go into a common header?