Resume stashing megawars #1992
Resume stashing megawars #1992
Conversation
Jenkinsfile
Outdated
| @@ -58,6 +58,9 @@ stage('prep') { | |||
| } | |||
| } | |||
| } | |||
| lines.each { | |||
| stash name: it, includes: "pct.sh,excludes.txt,target/pct.jar,target/megawar-$it.war" | |||
There was a problem hiding this comment.
99% of the stash is the megawar (now that PCT has been slimmed down considerably), so it did not seem worth the bother to separate the generic from line-specific contents. Anyway on S3 there is some fixed overhead per blob operation, and each Pipeline step adds a bit of overhead as well, so minimizing the count is probably the better choice. Also this is more legible I think.
…a.lang.String war`
|
Passed in 85m, good. @dduportal are we ready to go here? |
There was a problem hiding this comment.
Seems to leave the repository in a messier state than prior to #1955 by not fully reverting all of #1955. The decision about whether to revert or retain a given portion of #1955 should be driven by technical considerations, not one's desire to mess with or not mess with changes one does not (yet?) understand.
|
#1955 made certain simplifications along with extracting pieces of functionality to separate scripts, so I do not see any particular need to revert more. Also it is far from settled whether ultimately we will wish to use stashes or rebuild the megawar in different branches. Since this change is small, I see no reason to delay it; if someone wishes to devise some further simplifications to some scripts after ci.jenkins.io performance is stable, that can be done then. |
There was a problem hiding this comment.
I do not see any particular need to revert more.
You do not see any particular need to revert more because you have not looked. Why would you write something like
changes I am not familiar with and do not wish to mess with
when those changes are relevant to this current PR? Why have you not looked?
#1955 left behind a bit of a mess, including apparently moving pct-work from its expected target/local-test/pct-work to pct-work, a regression that still remains in this PR and has affected others who work on this repository.
I am (again) requesting a systematic evaluation of the revert of #1955, which requires becoming familiar with the subsequent changes to these same files. This is your responsibility as someone who contributes code to a repository that other people also work on. If the portion of #1955 was valuable, it should be retained. If the portion of #1955 is no longer necessary or buggy, it should be removed.
If there is a specific problem it can be addressed. |
|
Note that I asked "Why would you write something like […] when those changes are relevant to this current PR?" and "Why have you not looked?" @jglick did not respond. |
|
Merging to unblock the work on #1969 |
Reverts the essence of #1955, though a literal revert proved impractical since #1955 cleaned up stuff like
MAVEN_ARGSusage, there were some minor subsequent changes, and mainly because #1970 made subtle changes I am not familiar with and do not wish to mess with. Instead keeping the Bash scripts as they were and reintroducing a simplified stash system just for CI. Requested by jenkins-infra/helpdesk#3496 (comment) due to #1969 (comment).