[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: binary performance (was Re: Subversion & Word)

From: Greg Hudson <ghudson_at_MIT.EDU>
Date: 2003-03-05 20:56:37 CET

On Wed, 2003-03-05 at 11:51, Brian Denny wrote:
> we have some largeish (~5MB each) binaries in our CVS repository. commits
> and updates go pretty quick on new entries, but after a while they slow
> down (we were seeing times of two minutes per file at the tip of a long
> branch).

At a minimum, CVS needs to read the entire RCS file in order to produce
any version. For binary files, since there is no effective diffing
going on, that means if you have N versions of average size S, it will
take O(N*S) time to check out the file.

(My understanding may be out of date now that CVS includes its RCS code
internally. It may stop reading the archive when it finishes reading
the version it wants, in which case recent versions would still be
accessed efficiently. But versions on branches would be slow to
access.)

Subversion has a binary diff algorithm which may or may not be effective
on Word files. But even if it isn't, Subversion doesn't have to scan
through lots of versions of a file in order to get to the one it wants.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Wed Mar 5 20:57:27 2003

This is an archived mail posted to the Subversion Dev mailing list.

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.