[clang] [llvm] [ci] New script to generate test reports as Buildkite Annotations (PR #113447)

[David Spickett via cfe-commits]

DavidSpickett wrote:

> Basically, the sort of "plain text" report that gets added is the difference. Is that correct?

Yes, there is also a Linux report there that shows that it succeeded, but you have to select "All" on the "All | Failed" button. Which is maybe redundant but folks like seeing green boxes, I know I do.

This overall report is much more useful than seeing only the last `check-whatever` that was run at the end of the log. Which might not be the one that failed.

> Are there other differences (such as how test results get reported to Github actions)?

The result sent to GitHub is unchanged. For example the result on this PR is a failure because one test failed on Windows, because of the deliberate failure I put in.

If we wanted to send the report as a comment on GitHub, we can certainly look into that as mentioned here https://discourse.llvm.org/t/using-plugins-in-buildkite-ci-that-require-docker/82701/6?u=davidspickett. The Buildkite specific parts of the script are easy to remove.

https://github.com/llvm/llvm-project/pull/113447