[skip ci] not working

dane's Avatar

dane

15 Jul, 2014 07:36 PM

I've noticed that builds with `[skip ci]` in the comment message still trigger an appveyor run: https://ci.appveyor.com/project/Mapbox/node-sqlite3-163/build/1.0.43. This seems to conflict with the docs: http://www.appveyor.com/docs/how-to/skip-build.

Also, it would be great to support another term to be able to skip that is different than travis. I often want to skip travis builds but not appveyor builds and vice versa. What about `[skip appveyor]`?

  1. Support Staff 1 Posted by Feodor Fitsner on 15 Jul, 2014 08:34 PM

    Feodor Fitsner's Avatar

    Will take a look. [skip appveyor] is nice idea, thanks.

  2. Support Staff 2 Posted by Feodor Fitsner on 16 Jul, 2014 06:00 AM

    Feodor Fitsner's Avatar

    Trying to reproduce by pushing changes with [skip ci] in comments... How did you run that build - using "New build" button?

  3. 3 Posted by dane on 16 Jul, 2014 06:38 AM

    dane's Avatar

    No. It was a single commit if I recall correctly.

  4. Support Staff 4 Posted by Feodor Fitsner on 16 Jul, 2014 06:41 AM

    Feodor Fitsner's Avatar

    Hm, that's strange... Could you try reproducing it again?

  5. 5 Posted by dane on 16 Jul, 2014 04:37 PM

    dane's Avatar

    I can not replicate using the same commit message as before. So perhaps this was just a one time hiccup?

  6. Support Staff 6 Posted by Feodor Fitsner on 16 Jul, 2014 04:47 PM

    Feodor Fitsner's Avatar

    Anyway, I've checked-in [skip appveyor] support - will deploy tomorrow.

  7. Support Staff 7 Posted by Feodor Fitsner on 17 Jul, 2014 08:39 AM

    Feodor Fitsner's Avatar

    Update with [skip appveyor] has been deployed.

  8. 8 Posted by acgetchell on 28 Oct, 2017 05:21 AM

    acgetchell's Avatar

    [skip ci] still doesn't work for AppVeyor.

    In this instance, I was doing just documentation updates and wanted to skip both Travis and AppVeyor.

    https://ci.appveyor.com/project/acgetchell/causal-sets-explorer/build/1.0.70

  9. Support Staff 9 Posted by Ilya Finkelshte... on 29 Oct, 2017 07:52 PM

    Ilya Finkelshteyn's Avatar

    @acgetchell we recently changed behavior to look for [skip ci] (and other skip strings in commit messages) only in the text before first empty line. We did that because a lot of confusion happens then people merge and squash multiple commits and single older commit prevents whole merge commit to be built. You can move [skip ci] to the first line, or make it in the second line, but without empty line before it, or you can skip commit with .md files changed as described here

  10. 10 Posted by Jesse Rosenberg... on 04 Nov, 2017 08:08 PM

    Jesse Rosenberger's Avatar

    It's not clear to me why this commit would have ran on AppVeyor, even given your new criteria:

    https://ci.appveyor.com/project/meteor/meteor/build/83/job/lvsqxs0n57j8oy82

  11. 11 Posted by Jesse Rosenberg... on 04 Nov, 2017 08:11 PM

    Jesse Rosenberger's Avatar

    Also, it seems [the documentation](https://www.appveyor.com/docs/how-to/filtering-commits/#skip-directive-in-commit-message) linked to above may need to be updated if the commit message skip directives are (now) only observed under particular circumstances, since it still dictates "anywhere to commit message".

  12. 12 Posted by acgetchell on 04 Nov, 2017 09:12 PM

    acgetchell's Avatar

    Yes, it still ignores [skip ci] when it's the first line of the commit message.

    https://ci.appveyor.com/project/acgetchell/cdt-plusplus/build/1.0.207

  13. Support Staff 13 Posted by Ilya Finkelshte... on 09 Nov, 2017 03:11 AM

    Ilya Finkelshteyn's Avatar

    Documentation is updated. Commits referenced both here and here do not not have [skip ci] in the first line of commit message. They both are after first empty line. You can open GitHub webhook content and see it yourself: part with [skip ci] will be after \n\n. Probably [skip ci] was added into description part in GitHub (or GitHub for desktop) UI, which is separated with empty line behind the scenes.

  14. 14 Posted by Jesse Rosenberg... on 09 Nov, 2017 12:04 PM

    Jesse Rosenberger's Avatar

    I wrote that commit message with vi and used the same format that I believe is the standard usage case:

    Commit Subject
    
    Commit Body/Description
    

    This would be the equivalent of Commit Subject\n\nCommit Description.

    The default commit message format produced by git is in that same format when doing squashes, merge commits, etc., so I don't see how the format that I'm using is wrong. It's also the "best practice" format which is recommended by most editors, and as you pointed out, the way that GitHub expects it to be.

    Are you asking developers to use a different format? Like this:

    Commit Subject
    [skip ci]
    
    Commit Body/Description
    

    If not, can you give an example of a [skip ci] commit message, which follows standard best practices, which would actually skip the CI?

  15. Support Staff 15 Posted by Ilya Finkelshte... on 09 Nov, 2017 06:05 PM

    Ilya Finkelshteyn's Avatar

    Jesse, thank you for clarifying. I would say like this for example:

    Commit Subject
    Update user documentation [skip ci]
    
    Commit Body/Description
    Explained how to do something
    

    Sorry for the trouble. We tried to eliminate one confusion but it seems that brought another.

Reply to this discussion

Internal reply

Formatting help / Preview (switch to plain text) No formatting (switch to Markdown)

Attaching KB article:

»

Attached Files

You can attach files up to 10MB

If you don't have an account yet, we need to confirm you're human and not a machine trying to post spam.

Keyboard shortcuts

Generic

? Show this help
ESC Blurs the current field

Comment Form

r Focus the comment reply box
^ + ↩ Submit the comment

You can use Command ⌘ instead of Control ^ on Mac