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

Re: [Issue 875] Changed - cvs2svn combines separate commits (?)

From: Greg Stein <gstein_at_lyra.org>
Date: 2002-08-20 02:46:48 CEST

On Mon, Aug 19, 2002 at 10:46:08AM -0500, Karl Fogel wrote:
> Martin Tomes <lists@tomes.org> writes:
> > Scott Lamb wrote:
> > > It definitely is not okay for me. cvs2svn.py is combining commits by
> > > the same author and five-minute span with different commit
> > > messages. In some cases, these are two revisions of the
> > > _same_file_. (Correcting some obvious error in an earlier commit,
> > > for instance.)

That's a bug. Probably crept in when I fixed the script to not combine
commits that were 10 hours apart :-)

> > I use an MD5 checksum of the commit message plus the date of the
> > commit which to identify a set of files commited together. It fails
> > when someone commits the same fix to several branches of the same file
> > and uses the same commit text, but this is a rare occurance and is
> > detectable, I just haven't got round to fixing it.

cvs2svn.py also uses an MD5 checksum. However, it includes the author into
that checksum. We keep the date outside of the checksum since a single
commit can actually span a few seconds, depending on the size of the commit
and the user's bandwidth/latency to the server.

> Just curious: why use an MD5 sum instead of using the message itself?

Fixed length.


Greg Stein, http://www.lyra.org/
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Tue Aug 20 02:42:53 2002

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.