From: Steffen P. <ste...@gm...> - 2005-03-14 04:18:08
|
Hi, a few weeks ago I received the patch that adds branch support from Scott from Atlassian (thanks btw!). Unfortunately the merge did not work out too well, since the patch had been diffed against the repository quite a while ago. Part of the model classes had changed and the patch conflicted with the symbolic names support as it naturally touched the same parts of the code. Anyways, I just finished a first working branch support implementation based on the patch. I decided to take a take a slightly different approach than the original patch though. If understood the code correctly, the original patch limits the analysis to a single branch, whereas in my implementation all branches that are contained in the log file are added to the model and can therefore be analysed. Here is an example of the generated output of an analysis of a single branch : http://statcvs-xml.berlios.de/branches/branch_branch-support-1-branch.html. The commit log on the bottom shows that the revisions are indeed on a branch. And here is an example comparing three branches: http://statcvs-xml.berlios.de/branches/branch_stats.html The log file was generated by running 'cvs log -rbranch-support-1-branch,:HEAD,statcvs' on the statcvs-xml cvs. I still need to figure out how to calculate the loc correctly but at least the parser seems to work. If anybody want to look at the code, it is located in the statcvs-xml cvs (http://statcvs-xml.berlios.de/cvs-usage.html) in the 'branch-support-1-branch' branch. I am not quite sure what everybody expects from branch support (just analyse a single branch? or mutliple branches? or include the head branch up to the point where the branch was forked?). If anyone want to discuss this or tweak the code, please get in touch with me. Either by means of this list or private email. Steffen -- Steffen Pingel - ste...@gm... - http://steffenpingel.de |