Page 1 of 1
Posted: Fri Aug 17, 2007 1:18 pm
Synchronization of a project with ~500 files is really slow. Can it be that the synchronisation is a serial process? May be it would help to make the requests parallel.
Posted: Thu Aug 23, 2007 1:50 pm
First of all the Synchronization process speed depends on the connection speed. Connection speed is limited by latency and throughput. If the problem with synchronization speed relates to wery high connection latency, parallelizing of queries will not help because only one query is performed in order to access the repository and streamed data will be transferred as fast as possible. In case if throughput is low parallelizing also will not help due to channel limitation. At the same time each query contains lot of excess information which increases channel load. So, I think the best way in your case is scheduling of the synchronization process using Eclipse IDE features. In that case synchronization will run periodically in background and this solution should save you time.
Posted: Thu Aug 23, 2007 2:06 pm
we used the editor CodeWright before we changed to Eclipse. Here it was possible to use a plugin from PushOk for the subversion integration. The synchronisation was done in a few seconds. Eclipse needs some minutes...
But the scheduler is i nice idea...
Posted: Wed Mar 12, 2008 3:46 pm
personally i think the "Fetching Changes" for each changed file is what makes teh Syncronize very slow.
Why has subversive to fetch the changes when i press the syncronize button? Wouldn't it be enough if it fetches the changes if i double click / Open in Compare Editor a changed file?
What i really miss is something as fast as svn status -u with a console client.
Or is there already an option for disabling fetching the changes for all files on syncronize, and i just haven't found it yet?