Just out of curiousity, why is a working copy required to generate the statistics?
Both svn log and svn diff allow URLs as input, such that a working copy is not required. svn ls will provide the contents of a folder if given a URL, if the working copy is required just to get an idea of what the folders/files look like.
I guess if the working copy were not being used, you'd need to provide the URL of the top folder for which stats will be generated.
I'm guessing it's mainly due to its origins (StatCVS). That being cvs would need a local working copy to retrieve this information.
We use it to:
- determine the LOC in the last revision, inferring previous versions from this seed. We could use svn cat but that would be slow.
- determine some stuff relating to deleted files, binary files
Other things too but they don't come to mind right away.
Technically, it is feasible to implement a patch that would query everything on the server, but that would be slower.
> Technically, it is feasible to implement a patch that would query everything on the server, but that would be slower.
You could make it as an optional configuration paramter?!
Yes, see existing feature request 1687161.
Sign up for the SourceForge newsletter:
You seem to have CSS turned off.
Please don't fill out this field.