New version check: error on null/empty #2358
Conversation
|
@networkException see my comment above, this doesn't work... |
| @@ -23,9 +23,9 @@ jobs: | |||
| - name: Get the latest Chromium version | |||
| id: latest-version | |||
| run: | | |||
There was a problem hiding this comment.
set -e needs to be set before running jq commands, otherwise workflow will fail only if the last jq statement fails which might not be the case.
There was a problem hiding this comment.
From what I can tell from their documentation it seems that -e is already set by default.
There was a problem hiding this comment.
That's only for when you pass a single command, if it's multiple lines it doesn't.
There was a problem hiding this comment.
Ok then, I've added that to the PR.
|
gah, I didn't submit my comment which is why you didn't see it! it was "pending" |
|
@PF4Public P.S. you haven't requested any changes |
|
Sorry everyone for stopping this PR. The issue is that with this workflow we sometimes run into secondary limits for some reason, which results in GitHub spamming failed workflow notifications. I remember this was happening even on 2 hour intervals. I kinda got used to them (albeit it is a bit annoying, since I care for failed workflows in my repository). My concern was that if Ahrotahn merges this PR, this could affect where GitHub sends the notifications. But it turned to be a no issue as Ahrotahn kindly browsed the documentation and found out that only changing cron schedule should divert notifications. Not an issue with this PR per se, but sorry @rany2 if you took it personally. Now the problem itself could be solved in the action we are using. There is already an issue filled: dblock/create-a-github-issue#39. Solving it probably requires some experience with Node.js and JavaScript (of which I don't have enough). I'd be grateful if someone could solve it :) |
No no, I didn't take anything personally. It's just that you said you'll "explain in a moment" but didn't for a couple of hours. So I assumed you did a review like me but forgot to mark it as complete :) |
This PR fixes a bug with the new version check action mentioned in #2347 using the fix suggested by @rany2.