Description
It's been over 2 years since the last PyPi release of pysdl2, and there are a few important bugfixes that have been added to the GitHub version since the last one (importantly for me, the patches that allow pysdl2-dll to work on macOS). As such, I figure it's worth discussing what (if anything) should be done/fixed before the next PyPi version of pysdl2 is released. Off the top of my head, here are a few odds and ends:
- Syncing the package with the latest SDL2 2.0.10 changes (looking at the changelog, I don't think there's much needed here)
- Fixing Appveyor/Travis CI so it's passing again (I'm willing to tackle the bulk of this, pysdl2-dll should simplify the pipeline quite a bit for macOS and Windows)
- Adding official support for pysdl2-dll (essentially just adding a "try: import sdl2dll, except ImportError: pass" to the top of dll.py, so you don't need to add sdl2dll to your scripts to be able to use it).
- Closing some of the remaining open issues that seem easy and important.
Re: pysdl2-dll support, I've got everything working so that it's generating pip-installable wheels with the full suite of relevant macOS/Win32/Win64 SDL2 binaries included, and am getting it onto PyPi shortly, so we should finally be able to close #78!
Also @marcusva since I don't have access to the PyPi repository, you'll have to upload the new release when you find time. What do you think is needed before a 0.9.7 can/should be released?
EDIT: Okay, here's a checklist of what should get done before the next release:
For sure:
- Fix failing Linux tests on TravisCI
- Fix failing macOS tests on TravisCI
- Fix failing Windows tests on Appveyor
- Sync with SDL2 2.0.10 headers
- Sync with SDL2_mixer 2.0.4 headers
- Sync with SDL2_ttf 2.0.15 headers
- Sync with SDL2_image 2.0.5 headers
- Sync with SDL2_gfx 1.0.4 headers
- Resolve Python 3.8 fails to load SDL2.dll #145
- Update README
Maybe:
- Fix failing PyPy tests on TravisCI
- Add sdl2dll support
- Resolve Add tutorial for playing music and sound #3
- Resolve rwops.py's write method returns wrong value #126
- Resolve SDL_RWclose() doesn't seem to like whatever comes out of SDL_RWFromMem() #130
- Resolve SDL_GetPrefPath should return a c_char_p #135
- Resolve Window not shown unless get_surface is called. #139
- Resolve pixels2d doesn't work with PyPy #132
- Resolve SDL_SysWMmsg structure is undefined, even though the definition seems to already have been written #136
- Resolve SDL_JoystickGetGUIDString crashes CPython interperter but not PyPy #75