Sequential parallel branches appear incorrectly while build in-progress #1961
Conversation
…nformation into 2 copies one for stages one for steps, so we can fix them one at a time
… its own file as well
…tests that failed with the existing code, and then removed that code to fail some more tests, ready for new impl
… if I bollox this up
… node ids can change sometimes while in-progress due to backwards nature of the walk. Need to work around this
… working. TODO cleanup and more testing
…ph node comparison
…ph node comparison
sophistifunk
requested review from
halkeye,
imeredith,
kzantow,
NicuPascu and
scherler
as code owners
|
Example pipeline that can be used to verify before/after changes: https://gist.github.com/sophistifunk/7b75a86658b80e82715d213820b673c7 |
|
This doesn't address the shortcomings in the front-end code that are mainly responsible for JENKINS-55605 but some occurences of that problem will likely go away due to this fix. |
…r with JENKINS-53816
|
Build 5 was green, but I renamed some stuff so I'll try and get another green before I merge. You know, and a review. |
| } while (run2.isBuilding()); | ||
|
|
||
| // Sanity check, make sure we're *actually* checking stuff. | ||
| assertTrue("Checked multiple times while building", loopCount > 5); |
There was a problem hiding this comment.
I wonder if maybe we could have lower numbers on very slow/not powerful machines?
There was a problem hiding this comment.
I don't think it's a good idea. What use is a green test that's not really testing anything? I'd like to be able to use some other mechanism than time delays in the pipeline, something deterministic, but I don't know what that could really be. FWIW, the count running on my machine was > 90, so 5 seems pretty low.
Description
See JENKINS-53816.
Submitter checklist
Reviewer checklist