Looks like a tag got misparsed here? Not important, but thought it might be useful.
Screenshot of log attached. The first event is an expected response to me pushing a 0.1.2
tag for which the corresponding commit had not been pushed. Later, I pushed the commit, and then pushed tags again (to catch the later one I was actually trying to work with). In addition to the correct jobs I was trying to start, the second tag push created the .1.2
event.
All of this was in my test repo and incidental to something else, so I don't need any help on this one, it just looks fishy and I thought you might want to know about it.
For easier copy/paste, the first webhook is 52460d88-be00-11e9-8428-994441ce371d
and the second one is 9e8078dc-be14-11e9-9866-89bc6e323437
.
Comments are currently closed for this discussion. You can start a new one.
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
1 Posted by Ilya Finkelshte... on 14 Aug, 2019 06:46 PM
Thank you! Can you please also send a link to the the build (with correct jobs) which was started when you pushed tag second time?
2 Posted by finnre on 14 Aug, 2019 07:05 PM
It's a little hard to tell because the timing isn't exactly the same in the log, but I think it was this one: https://ci.appveyor.com/project/relsqui/matrix-repro/builds/26678322
Might technically have been this one, the branch build for the same commit as the above tag build: https://ci.appveyor.com/project/relsqui/matrix-repro/builds/26678301
(My notes from when I was working at the time are here, they say basically the same as the above in more words: https://github.com/relsqui/matrix-repro/blob/master/02-multiple-ext... and in general feel free to poke around that repo/build history if it's useful, they're public on purpose.)
3 Posted by Ilya Finkelshte... on 15 Aug, 2019 05:00 AM
Here is what I see.
First webhook part:
Second webhook part:
Please note the
"base_ref": "refs/tags/0.1.2"
in the second webhook content. When AppVeyor sees thatbase_ref
us notnull
, we assume it containsrefs/heads/branch_name
and substring it accordingly. Hense0
was lost.This is probably a bug, I just need to understand the repro (will read your lab description later for that).
Ilya Finkelshteyn closed this discussion on 15 Oct, 2019 09:01 PM.