Details
-
Bug
-
Status: Open (View Workflow)
-
Major
-
Resolution: Unresolved
-
jenkins master 2.88 - openSUSE Leap 42.3
bibucket branch source plugin 2.2.7
bitbucket plugin 1.1.8
Description
scope: declarative pipeline + bitbucket multi branch
for optimization purpose (big repo), i chose to skip the default checkout.
options {
disableConcurrentBuilds()
skipDefaultCheckout(env.PR_NUM != null)
}
i also implemented my own checkout, depending on branch's type (feature, PR-,...)
pipeline triggers jobs on every commit i do, but somehow the "changes" are not caught
when i do NOT skip the default checkout, everything works as expected.
Question: is this as "per-design", how can i get rid of this issue?
Attachments
Issue Links
- duplicates
-
-
- Open
-
Activity
kgiloo
added a comment - vivek YES indeed: with version 2.2.3 pipeline shows the changes as expected.
kgiloo
added a comment - vivek YES indeed: with version 2.2.3 pipeline shows the changes as expected. kgiloo
added a comment - vivek Any chance to have this fixed in a near future? kgiloo
added a comment - vivek Please can you update me on this bug? it has occurred since version 2.2.3
kgiloo
added a comment - vivek Please can you update me on this bug? it has occurred since version 2.2.3 stephenconnolly abayer With lightweight checkout, full checkout doesn't happen, somehow its interfering with detecting changes with
skipDefaultCheckout(env.PR_NUM != null)
?
Well, if you've got skipDefaultCheckout(true), there's no automatic checkout of the full repo happening on the agent, so that'd basically be the same as if you just had a Pipeline like
echo "hi"
So if you don't see changes with that Scripted Pipeline example either, then the problem is definitely that light checkouts aren't updating the changes. Then again, I'm not 100% sure that's a problem vs just how it's supposed to work... stephenconnolly would know best on that behavior, since it's branch source level, not Declarative level.
Possibly dupe of JENKINS-46431.
(Why would you implement your own checkout? checkout scm already handles every case.)
kgiloo
added a comment - jglick we need to extract the history from our workspace and as far as i know the 'checkout scm' per design does not handle the clone configuration... or am i wrong?
kgiloo
added a comment - jglick we need to extract the history from our workspace and as far as i know the 'checkout scm' per design does not handle the clone configuration... or am i wrong? 
kgiloo Did this start failing since bitbucket-branch-source 2.2.7? Do you see it working with 2.2.3? M making sure it was not a regression caused due to light weight checkout feature added in 2.2.7.