Switch to downloading dependencies from Github releases #3236
Conversation
|
Uhh, this wasn't supposed to be merged. I guess it's fine because the sync will override it? |
|
https://github.com/OpenMW/openmw/actions/runs/11778077287 says it didn't work anyway. When I had to test the Actions after making a change, I made a branch on my fork and didn't need to make a PR for it to run. Except in the Actions tab, I can't see an attempted run for this branch in any of the normal places, so I guess that's just what it does when there's a syntax error. |
|
Yeah, I've got them running happily on my own branch now. I think the first one was weird because I didn't push the master branch first (so the Actions tab just said some stuff about adding them at all.) |
|
That run's dying on https://gitlab.com/OpenMW/openmw/-/blob/master/components/config/launchersettings.cpp?ref_type=heads#L295, which I originally did with a default capture, and then had to stop because Clang |
|
Yep https://github.com/OpenMW/openmw/actions/runs/11540499199/job/32121398575#step:12:711 It was failing before that pretty often due to the tests failing. That seemed to start with https://github.com/OpenMW/openmw/actions/runs/10697010351/job/29653330876#step:22:3202 a couple of months ago. |
|
We've at least got stronger evidence that it was caused by https://gitlab.com/OpenMW/openmw/-/merge_requests/4336 (or at least one of the few immediately preceding it) as we've got the log of actions showing the crash wasn't happening, and the first failure was when that MR got merged. It was open for a week, and doesn't look like it was rebased, so with the rate things have been failing since its merge, we'd have expected at least one of the actions runs between it being branched and merged to have failed if it was a different MR that caused the problem. |
|
That does look pretty convincing. Hmm. |
PR for testing actions.
https://gitlab.com/OpenMW/openmw/-/merge_requests/4450