timja
commented
8 years ago I wouldn't say it's harmless as the side effect seems to be that in the email output test results from imported xml files are not seen. In /var/lib/jenkins/email-templates we have a file called groov.groovy with the following
<% def junitResultList = it.JUnitTestResult
if (junitResultList.size() > 0) { %>
| "bg1" colspan="2">Tests |
| "bg2" colspan="2"> Name: ${packageResult.getName()} Failed: ${packageResult.getFailCount()} test(s), Passed: ${packageResult.getPassCount()} test(s), Skipped: ${packageResult.getSkipCount()} test(s), Total: ${packageResult.getPassCount()+packageResult.getFailCount()+packageResult.getSkipCount()} test(s) |
| "test_failed" colspan="2"> |
When the full list of tests from two xml files is not displayed then this error is always shown referencing the exact job.
Your Jenkins log can fill up with long stack traces like
h.ExpressionFactory2$JexlExpression#evaluate: Caught exception evaluating: browser.getChangeSetLink(cs) in /jenkins/job/.../1234/. Reason: java.net.MalformedURLException: no protocol: java.net.MalformedURLException: no protocol: at java.net.URL.(URL.java:585)
at java.net.URL.(URL.java:482)
at java.net.URL.(URL.java:431)
at hudson.plugins.git.browser.GitRepositoryBrowser.getUrl(GitRepositoryBrowser.java:47)
at hudson.plugins.git.browser.GithubWeb.getChangeSetLink(GithubWeb.java:40)
at ...
when this is a harmless condition: just an unconfigured git repository browser base URL.
Probably RepositoryBrowser should get a new method that calls getChangeSetLink but logs exceptions at FINE and returns null, and then Jelly views (both in core and plugins) should be updated to use it. (Falling back to displaying an unlinked changeset, just as if getChangeSetLink returned null.)
Originally reported by jglick, imported from: RepositoryBrowser.getChangeSetLink exceptions from Jelly pollute log