Closed joeyparrish closed 11 months ago
Unfortunately, this isn't triggering the test itself because the new trigger (pull_request_target) runs the workflow that exists on the target (shaka-project), not the fork (joeyparrish). So the effect of this isn't seen until after it is merged.
However, you can see the effect here on a dummy PR on my fork:
https://github.com/joeyparrish/static-ffmpeg-binaries/actions/runs/6341990006/job/17226831542?pr=5
Configure Build Matrix logs:
{
enableDebug: true,
enableSelfHosted: false,
matrix: [
{
os: 'ubuntu-latest',
os_name: 'linux',
target_arch: 'x64',
exe_ext: ''
},
{
os: 'macos-latest',
os_name: 'osx',
target_arch: 'x64',
exe_ext: ''
},
{
os: 'windows-latest',
os_name: 'win',
target_arch: 'x64',
exe_ext: '.exe'
}
]
}
You can see the effect of vars.ENABLE_DEBUG
.
After this is merged, we should see the effect of vars.ENABLE_SELF_HOSTED
on the repo in shaka-project
, after which I can do another manually-triggered build and verify that everything works on linux-arm64.
Above I mentioned that pull_request_target
runs the workflow that exists on the target (shaka-project), not the fork (joeyparrish). So one good side-effect of splitting the build scripts out from the workflow into separate script files (#22) is that this enables pull_request_target
to test changes to those scripts.
I discovered some further issues that impact PR testing, so I've fixed those here and rewritten the PR description to match. PTAL!
This fixes multiple CI issues: