|
I'm also seeing these errors:
2005-06-23 12:11:53,512 WARN [sf.statcvs.input.Builder] Revision 1.27.2.5 has no main branch - ignoring revision. and from the revision numbers I don't think that this is related to importing files. Looking at the statcvs and Jira source, it could be that Jira is ignoring files that only exist on branches, not on the mainline? If the warning message displayed the filename, it would be easier to track down what is happening (that's a minor request for a future release) The problem with this message is that it makes people think that there is corruption in their CVS repositories. If it is an expected occurence, then it shouldn't be displayed. ~Matt I believe I have a small test case for this. I modified the statcvs-patched jar file to display comments, so I could actually
work out which files were causing this (very useful). Then I created a minimal cvs.log file (attached) and when Jira started up, I saw: 2005-06-23 16:53:43,932 WARN [sf.statcvs.input.Builder] Revision 1.27.2.5, comment: fix for bug 1392. use a connection timeout for TCP connection: has no main branch - ignoring revision. The source file in question displays fine with ViewCVS, and I can check out each of the mentioned revisions, so I suspect that the modifications made by Atlassian to the statcvs source code may be making some invalid assumptions. I don't think that the attached cvs log should generate any such warnings. ~Matt Aha! I withdraw that suspicion of Jira's modifications to statcvs. I was in fact mistaken to say that this file looked ok in ViewCVS. It turned out that someone had messed with the branch numbers in the past (cvs admin is a risky command). So in this case statcvs was correctly noting real problems in oour CVS repository, and the warnings should not be disabled.
In the attached cvs log file, the tag that created the branch TAG_2_1_1_1_BRANCH: 1.27.0.4 should have been: TAG_2_1_1_1_BRANCH: 1.27.0.2 (http://durak.org/cvswebsites/doc/cvs_58.php#SEC58 ~Matt So - can we close this issue?
Please close. We were able to resolve by modifying the log4j.xml on the server.
Yes, ignoring the 1.1.1.1 warnings seems like the right thing to do. However, other warnings from statcvs are important, since they indicate real problems with the user's repository.
How about having Jira filter out the 1.1.1.1 warnings itself? ~Matt |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
Improvement
Open
Minor
Change log level of "Revision 1.1.1.1 has no main branch" messages in statcvs jar to debug
This warning is printed as JIRA does not read revisions that signify the initial import of the module to CVS.
There messages can be safely ignored. Does the CVS integration work as expected?
If you would like these log messages to be supressed please add teh following 2 lines to the WEB-INF/classes/log4j.properties file under the JIRA web application and restart JIRA:
log4j.category.net.sf.statcvs.input = ERROR, console
log4j.additivity.net.sf.statcvs.input = false
I hope this helps. Please let us know if you have any other questions or would like more information. Otherwise please resolve this issue.
Thanks,
Anton