Comment From: wilkinsona
I think this was opened in error. The CI logged the following:
parse error: Invalid numeric literal at line 2, column 0
That looks like a temporary glitch as the other Java 14 version checks all ran successfully and didn't find anything to upgrade. We should improve the error handling to avoid creating the issue.
Comment From: dreis2211
What about just adding set -e on top? That should probably exit the script before it tries to create the issue. What do you think? I can provide a PR for that.
Comment From: wilkinsona
That would certainly be better than how it behaves at the moment. I was wondering if we should output the response that couldn't be parsed so that we can be sure that it was just a glitch rather than a breaking change in the format of the response.
Comment From: dreis2211
I'm currently playing around with more incremental steps and additional checks. Since I'm hitting that at the moment and I can't see the logs of the build unfortunately to see where it exactly failed: maybe it's related to rate-limiting? This produces another result which can't be parsed.
Comment From: wilkinsona
Closing in favour of #22110.