ci: enable tests

[madebr]

[madebr]

I see only 2 (!!!) test issues:

• [ ] testautomation --filter audio_resampleLoss

  ERROR: 07/19/24 19:09:00: Assert 'Expected output length 9600000, got 9600004.': Failed
  

  Is the test too strict? Or is this actual audio distortion?
• [x] The tests are timing out on the msys2 runners and nothing is actually run. I will look into this. (EDIT: this is fixed. msys2 did not pick up SDL3's PATH set by setup-sdl, so it's now set by cmake)

[slouken]

Is the test too strict? Or is this actual audio distortion?

I think this is actual audio distortion, I was waiting for @icculus to look into this.

[sezero]

Suggest rebasing and retesting especially after eae0b80c095fd6171e524bcca86c4627f547c6c2 which fixed incompatibility between dynapi and main body .

[madebr]

I see a segfault on macos. Perhaps it's related to https://github.com/libsdl-org/sdl2-compat/issues/206?

[slouken]

I see a segfault on macos. Perhaps it's related to #206?

I'm sure it is.

[madebr]

There is still some segfault on macos This is using SDL3 hash fe5dcf79be0c00d7220ea0a8fdecf1306d2c90d9 -> https://github.com/libsdl-org/SDL/commit/fe5dcf79be0c00d7220ea0a8fdecf1306d2c90d9

[slouken]

There is still some segfault on macos This is using SDL3 hash fe5dcf79be0c00d7220ea0a8fdecf1306d2c90d9 -> libsdl-org/SDL@fe5dcf7

Fixed in a81922a. @icculus, can you double check that this matches the original SDL2 behavior on macOS?

[madebr]

Woohoo! Awesome job, SDL team !

[madebr]

@slouken you might want to disable the "Create a merge commit" option for PRs for this repo, and other satellite repos, as well

[slouken]

@slouken you might want to disable the "Create a merge commit" option for PRs for this repo, and other satellite repos, as well

Done!