-
Bug
-
Resolution: Fixed
-
Critical
-
* {{jenkins}}: {{1.652}}
* {{git-plugin}}: {{2.4.2}}
* {{git-client-plugin}}: {{1.19.6}}
the latest git + git-client combination that worked is:
| plugin | version |
|---|---|
| git | 2.4.1 |
| git-client | 1.19.2 |
I have upgraded from the above combination to 2.4.2 + 1.19.6, and then I've started experiencing the following problem for pull requests that had a "catchup" (one or more merges FROM destination branch before the pull request):
20:22:31 java.io.IOException: Failed to parse changelog 20:22:31 at hudson.model.AbstractBuild$AbstractBuildExecution.defaultCheckout(AbstractBuild.java:625) 20:22:31 at jenkins.scm.SCMCheckoutStrategy.checkout(SCMCheckoutStrategy.java:86) 20:22:31 at hudson.model.AbstractBuild$AbstractBuildExecution.run(AbstractBuild.java:529) 20:22:31 at hudson.model.Run.execute(Run.java:1738) 20:22:31 at hudson.model.FreeStyleBuild.run(FreeStyleBuild.java:43) 20:22:31 at hudson.model.ResourceController.execute(ResourceController.java:98) 20:22:31 at hudson.model.Executor.run(Executor.java:410) 20:22:31 Caused by: java.lang.NullPointerException 20:22:31 at com.google.common.base.Preconditions.checkNotNull(Preconditions.java:187) 20:22:31 at com.google.jenkins.plugins.metadata.scm.SourceMetadata.<init>(SourceMetadata.java:62) 20:22:31 at com.google.jenkins.plugins.metadata.scm.SourceMetadataExtractor$FromGitBuildActions.getSourceMetadata(SourceMetadataExtractor.java:242) 20:22:31 at com.google.jenkins.plugins.metadata.scm.SourceMetadataExtractor$FromBuildActions.extract(SourceMetadataExtractor.java:139) 20:22:31 at com.google.jenkins.plugins.metadata.scm.SourceMetadataExtractor.extract(SourceMetadataExtractor.java:101) 20:22:31 at com.google.jenkins.plugins.metadata.scm.SourceMetadataBuildListener.onChangeLogParsed(SourceMetadataBuildListener.java:50) 20:22:31 at hudson.model.listeners.SCMListener.onChangeLogParsed(SCMListener.java:120) 20:22:31 at hudson.model.AbstractBuild$AbstractBuildExecution.defaultCheckout(AbstractBuild.java:623) 20:22:31 ... 6 more
- duplicates
-
-
- Fixed but Unreleased
-
[JENKINS-33433] jobs fail to "parse changelog" for pull requests with "catch-up" merges from destination branch
| Assignee | Original: Mark Waite [ markewaite ] |
hi, markewaite
The problem reproduces on github enterprise and on github.com.
I suspect it has nothing to do with github, though I don't have the time to dig into this.
catchup
I am calling "catchup" a merge from target branch into the feature branch before the merge of feature branch into the target branch:
i.e. if you're branching out from master, and working in your branch feature_coolness, do 3 commits, and in the meanwhile 5 commites were pushed into master, and one of them prevents you from merging, or you just want to make sure you have the latest changes from master for another reason.
So, let's call the above "catchup" of a feature branch from target branch.
A feature branch can have 0 or MORE catchups.
The issue reproduces upon pull request that has 1 or MORE catchups.
The issue does not reproduce on pull requests with 0 catchups.
Is this clearer ?
| Workflow | Original: JNJira [ 169350 ] | New: JNJira + In-Review [ 183460 ] |
+1
but with svn:
11:17:18 No changes for http://svn.dsgfdfqsg/dsfgdfgdfgdfs/pkg since the previous build 11:17:18 java.io.IOException: Failed to parse changelog 11:17:18 at hudson.model.AbstractBuild$AbstractBuildExecution.defaultCheckout(AbstractBuild.java:622) 11:17:18 at jenkins.scm.SCMCheckoutStrategy.checkout(SCMCheckoutStrategy.java:86) 11:17:18 at hudson.model.AbstractBuild$AbstractBuildExecution.run(AbstractBuild.java:529) 11:17:18 at hudson.model.Run.execute(Run.java:1720) 11:17:18 at hudson.model.FreeStyleBuild.run(FreeStyleBuild.java:43) 11:17:18 at hudson.model.ResourceController.execute(ResourceController.java:98) 11:17:18 at hudson.model.Executor.run(Executor.java:404) 11:17:18 Caused by: java.lang.NullPointerException 11:17:18 at java.lang.String.indexOf(String.java:1718) 11:17:18 at java.lang.String.indexOf(String.java:1698) 11:17:18 at com.hp.application.automation.tools.octane.model.processors.scm.SvnSCMProcessor.getRevisionIdFromBuild(SvnSCMProcessor.java:153) 11:17:18 at com.hp.application.automation.tools.octane.model.processors.scm.SvnSCMProcessor.getBuiltRevId(SvnSCMProcessor.java:132) 11:17:18 at com.hp.application.automation.tools.octane.model.processors.scm.SvnSCMProcessor.getSCMData(SvnSCMProcessor.java:67) 11:17:18 at com.hp.application.automation.tools.octane.events.SCMListenerImpl.createSCMData(SCMListenerImpl.java:103) 11:17:18 at com.hp.application.automation.tools.octane.events.SCMListenerImpl.onChangeLogParsed(SCMListenerImpl.java:71) 11:17:18 at hudson.model.listeners.SCMListener.onChangeLogParsed(SCMListener.java:120) 11:17:18 at hudson.model.AbstractBuild$AbstractBuildExecution.defaultCheckout(AbstractBuild.java:620) 11:17:18 ... 6 more
after having restored these plugins, it's ok:
- scm-api
- credentials
- github-api
- git
- subversion
- hp-application-automation-tools-plugin
| Priority | Original: Minor [ 4 ] | New: Critical [ 2 ] |
| Component/s | New: scm-api-plugin [ 18054 ] |
| Component/s | New: subversion-plugin [ 15485 ] |
The stack trace seems to indicate that a null pointer exception is happening in the google-source-plugin. It does not seem willing to accept that sometimes the value of branch may be null.
I don't know what would have changed between the previous plugin versions and the new plugin versions which would cause that. Can you provide more information about your configuration and the job which is showing that null pointer exception? As far as I can tell, the google source plugin has not changed since its original 0.1 release in 2014, so this must be due to one of the changes in either the git plugin or the git client plugin.
I don't understand what you mean by a "pull request that has a catch-up". Can you explain further which source control system you're using (github, bitbucket, gitlab, something else) and how the pull request is represented which is causing the null pointer exception?