Hi Brad!
> Greetings,
>
> We are in the early stages of a trial transition to Subversion from
> ClearCase and CVS. In the past I have relied heavily on the
> cvs-status emacs interface.
>
> I have noticed that working directory trees of even 1000 files cause
> svn-status and commits (which apparently trigger a status refresh) to
> be unbearably slow. I'm guessing much of this time is consumed by the
> sort of the status results. Operations that took 10 seconds with CVS
> now take many minutes with subversion. I see that this slowness has
> been reported before.
>
> Is there a workaround or any effort in the works to improve this
> situation? Would it be difficult to at least avoid rebuilding the
> entire status buffer on single file commits? Or perhaps sorting the
> status results on Linux/Unix systems via a pipe to the sort command
> rather than within emacs would help.
As a short term solution: There is the variable
svn-status-sort-status-buffer to switch of the sorting completly.
However, I guess, the sorting will be (perhaps too) wrong in that case.
I have a nice idea to speed up the commit process. Expect it in the
next few days/weeks.
> In additional feedback, I would like to point out that while the tree
> indent structured status buffer looks nice, it actually is less useful
> in some cases than the CVS version because searching within the buffer
> for a specific path is more difficult in the case of large trees with
> replicated directory and file names in different parts of the tree.
Thanks for the feedback.
Stefan.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Mon Oct 4 11:21:25 2004