ls-remote is indeed sufficient to retrieve branches HEADs, and we got such a change implemented in 2.4.0 development build, BUT git plugin do have some legacy logic to only consider tip branches, i.e. exclude branches that are ancestors for other branches. As a result, we ended with infinite build loop. Mark discovered this just before the release so we had to rollback.
That being said, multi-branch definitively is a better approach vs the current buggy multi-branch support in git-plugin, and maven job has lot's of side effects and limitations so that some of us consider it the evil job type. Setting equivalent freestyle job require some extra configuration but is feasible afaik. 2015-08-03 21:05 GMT+02:00 Nigel Magnay <[email protected]>: > I don't see why polling via a workspace should be required. 'git > ls-remote' is sufficient to determine branch names and SHA1s, and thus to > determine if there are any changes to build. And in any case, there will be > no workspace, since all builds on lightweight docker containers, torn down > immediately after every build. > > It's pretty lousy UX that several options seem to imply the need for a > workspace (yes, I understand this may be required to do 'advanced' > processing, but not in my usecase) - which may never ever exist, resulting > in mysterious (because the config page never mentions this) build > everything over and again behaviour. > > On the project type: that's not that simple, as we rely on the artifact > information stored by Jenkins against the build in the maven project type. > But yes, I've never understood why there's a "native/integrated maven > build" *type -* which can invoke maven precisely once- rather than a > "native/integrated maven build *step*". > > > On Mon, Aug 3, 2015 at 5:54 PM, Mark Waite <[email protected]> > wrote: > >> Considering how poorly a number of developers think of the maven project >> type, aren't you at least tempted to try a multi-branch with a maven build >> step, rather than using the maven project type? I think your users will >> thank you. >> >> Mark Waite >> >> On Mon, Aug 3, 2015 at 10:30 AM Nigel Magnay <[email protected]> >> wrote: >> >>> Multi-branch looks useful, but sadly this project uses the maven project >>> type. >>> >>> I did see some reports about it getting confused that the branch name >>> has a '/' in it. If I get a chance I'll try and diagnose. >>> >>> >>> On Mon, Aug 3, 2015 at 5:20 PM, Mark Waite <[email protected]> >>> wrote: >>> >>>> I've recently been using the multi-branch plugin with great success. >>>> It automatically creates a new job for each branch it detects, rather than >>>> "hiding" the detection of a new branch inside the execution history of a >>>> single job. You might experiment with it to see if it meets your needs >>>> better than building different branches in the same job. >>>> >>>> I believe the "nonexisting workspace" message is evidence of a bug in >>>> the plugin. I don't know the depth or breadth of the bug, but I've seen >>>> the same message. If you find a clean way to duplicate it, please submit a >>>> bug report to track it. >>>> >>>> Mark Waite >>>> >>>> On Mon, Aug 3, 2015 at 10:04 AM Nigel Magnay <[email protected]> >>>> wrote: >>>> >>>>> So I'm trying to set up the git plugin to build feature branches. >>>>> >>>>> Features are on branches like feature/foo and feature/blah. This is a >>>>> common pattern as it's derived from git-flow. >>>>> >>>>> I seem to have tried every combination of branch specifiers, and all >>>>> this results in is jenkins building the same SHA1 over, and over again. >>>>> I've tried removing wildcards, and specifying directly two branch >>>>> specifiers in multiple different formats (refs/heads/features/A, >>>>> remotes/origin/features/A) all to no avail. >>>>> >>>>> All of them seem to result in a polling log declaring >>>>> nonexisting_workspace >>>>> which I don't understand since determining if there are changes ought >>>>> not to require a checkout. But it simply re-triggers and fills up my >>>>> disks. >>>>> >>>>> What have I missed? >>>>> >>>>> -- >>>>> You received this message because you are subscribed to the Google >>>>> Groups "Jenkins Developers" group. >>>>> To unsubscribe from this group and stop receiving emails from it, send >>>>> an email to [email protected]. >>>>> To view this discussion on the web visit >>>>> https://groups.google.com/d/msgid/jenkinsci-dev/CAPYP83TA_0H5vhRRh-YxNLjrFqJGnMXtZjHkn5bA83sPzzHyWQ%40mail.gmail.com >>>>> <https://groups.google.com/d/msgid/jenkinsci-dev/CAPYP83TA_0H5vhRRh-YxNLjrFqJGnMXtZjHkn5bA83sPzzHyWQ%40mail.gmail.com?utm_medium=email&utm_source=footer> >>>>> . >>>>> For more options, visit https://groups.google.com/d/optout. >>>>> >>>> -- >>>> You received this message because you are subscribed to the Google >>>> Groups "Jenkins Developers" group. >>>> To unsubscribe from this group and stop receiving emails from it, send >>>> an email to [email protected]. >>>> >>> To view this discussion on the web visit >>>> https://groups.google.com/d/msgid/jenkinsci-dev/CAO49JtFpPSoa0n3hdkshHenkRm%2BSjTn2WqTQiwTOPqThFOfr9Q%40mail.gmail.com >>>> <https://groups.google.com/d/msgid/jenkinsci-dev/CAO49JtFpPSoa0n3hdkshHenkRm%2BSjTn2WqTQiwTOPqThFOfr9Q%40mail.gmail.com?utm_medium=email&utm_source=footer> >>>> . >>> >>> >>>> For more options, visit https://groups.google.com/d/optout. >>>> >>> -- >>> You received this message because you are subscribed to the Google >>> Groups "Jenkins Developers" group. >>> To unsubscribe from this group and stop receiving emails from it, send >>> an email to [email protected]. >>> To view this discussion on the web visit >>> https://groups.google.com/d/msgid/jenkinsci-dev/CAPYP83T%2B8bv6DKueX1%2BGULP9As4sj%3Drgxgiz5mkmxvpCiRgQ%2BA%40mail.gmail.com >>> <https://groups.google.com/d/msgid/jenkinsci-dev/CAPYP83T%2B8bv6DKueX1%2BGULP9As4sj%3Drgxgiz5mkmxvpCiRgQ%2BA%40mail.gmail.com?utm_medium=email&utm_source=footer> >>> . >>> For more options, visit https://groups.google.com/d/optout. >>> >> -- >> You received this message because you are subscribed to the Google Groups >> "Jenkins Developers" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to [email protected]. >> To view this discussion on the web visit >> https://groups.google.com/d/msgid/jenkinsci-dev/CAO49JtEQ-hpsm9Zu9JaKG2hBkr4kVvg7uz83PwYi87Tuix2A9g%40mail.gmail.com >> <https://groups.google.com/d/msgid/jenkinsci-dev/CAO49JtEQ-hpsm9Zu9JaKG2hBkr4kVvg7uz83PwYi87Tuix2A9g%40mail.gmail.com?utm_medium=email&utm_source=footer> >> . >> >> For more options, visit https://groups.google.com/d/optout. >> > > -- > You received this message because you are subscribed to the Google Groups > "Jenkins Developers" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > To view this discussion on the web visit > https://groups.google.com/d/msgid/jenkinsci-dev/CAPYP83TRpCmZ6pHOneq85uie7EU%2BmRWSTPyO13NF5OArk2J-6g%40mail.gmail.com > <https://groups.google.com/d/msgid/jenkinsci-dev/CAPYP83TRpCmZ6pHOneq85uie7EU%2BmRWSTPyO13NF5OArk2J-6g%40mail.gmail.com?utm_medium=email&utm_source=footer> > . > > For more options, visit https://groups.google.com/d/optout. > -- You received this message because you are subscribed to the Google Groups "Jenkins Developers" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/jenkinsci-dev/CANMVJzkV-yZLc5q1e7ZCt0JTFHMUBW%3Dhin8MVrz74vKoA5grPw%40mail.gmail.com. For more options, visit https://groups.google.com/d/optout.
