If we don't have this, we can't filter properly the results in the UI and main stats gets merged in the rest, completely throwing off the value of these analytics (while it's useful to track those in all branches, the timings and flakiness really matter solely on main).
https://github.com/buildkite/test-collector-javascript/blob/9d29eb099a89bdab9e2b0d3420cc0e111fb0c0bc/util/ci.js#L72-L81 is really imprecise and doesn't account for all variants the value can take. This PR fixes it on our side until they do on their side.
If we don't have this, we can't filter properly the results in the UI and
main
stats gets merged in the rest, completely throwing off the value of these analytics (while it's useful to track those in all branches, the timings and flakiness really matter solely onmain
).Test plan
Tested on this PR + CI