fix: use fallbacks to get PR id for CodeBuild #1079
Conversation
|
So, any thoughts on this one? |
|
The idea seems good, I was just waiting for it to be green 👍🏻 |
|
Ok, that's great, thanks. I don't think that failure is anything to do with this PR, although I could be wrong. Looks like it only happens for one of those Travis tasks, and has happened for a few PRs recently, perhaps caused by some async task continuing after a test has ended 🤔 I will see if I can pin it down later. |
|
So yeah, it seems like there were some slightly dodgy assertions in some unrelated tests. Those tests are basically asserting this rejection. As the Node version isn't pinned in Travis when running the tests it probably only recently started using Node 15, which changes the default behaviour to throw on unhandled rejections. Those Travis tasks are still pending as I write this, but I think they should pass now 🤞 |
|
Cool! Nice one this look good to me 👍🏻 |
As mentioned in #1077
CODEBUILD_SOURCE_VERSIONseems a little unreliable as a source for the PR number.This PR keeps that as the default but falls back to
CODEBUILD_WEBHOOK_TRIGGER, then to calling the various APIs viagetPullRequestIDForBranch(), which fortunately was already there. Still may not be foolproof, but should be an improvement!