Parsing svn log file

[GoogleCodeExporter]

I noticed that the svnlog parser is rather slow on my multigigabite 5k revision 
repository (it's running for more than 12 hours now). I wrote a quick hack to 
fill in the sqlite table from the xml verbose log (svn log -v --xml [REPOS]). 

Attached is the code I use to create the records. I am not sure if this is 
useful for you, so please have a look at the attached source. It runs in 2 
seconds for the 5k revisions a 300k line log file.  

I stripped out the loading of the sqlite database because I think you'd want to 
implement it into a svnlogiter class. I didn't implement the copyfrompath and 
copyfromrev. This approach doesn't give line counts.

Please let me know if you think a svn log based approach could be included.

Original issue reported on code.google.com by f.ba...@gmail.com on 2 Jun 2011 at 12:10

Attachments:

• svnxmllog2sqlite.py

[GoogleCodeExporter]

It's meant as an enhancement, not a defect, but I can't seem to change that.

Original comment by f.ba...@gmail.com on 2 Jun 2011 at 12:11

[GoogleCodeExporter]

Thanks. I will add it to svnplot. 

svnplot takes a lot of time for linecount analysis. If you disable the 
linecount, it goes lot faster. Have you tried running svnlog2sqlite.py without 
-l option ?

However, it does query one revision at a time. So it can still somewhat slow.

Original comment by nitinbh...@gmail.com on 3 Jun 2011 at 11:34

• Added labels: Type-Enhancement
• Removed labels: Type-Defect