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

RE: merging a patch

From: Dale Worley <dworley_at_pingtel.com>
Date: 2005-05-03 16:20:47 CEST

> From: Ben Collins-Sussman [mailto:sussman@collab.net]
>
> 'svn merge' isn't "just a diff", it's quite sensitive to history
> ("ancestry"). Read about it here:

OK, I forgot that.

> So when the merge command wants to add a file, it doesn't
> just schedule
> it for addition; it schedules it as a *copy*, i.e.
> addition-with-history. If the history comes from a completely
> different repository, everything breaks.

But that's not entirely true, either. As long as the repositories have
similar file trees, and the changes are relatively small changes in files,
"svn merge" works just fine. (And that has proved useful to us here.)

Indeed, a few weeks ago I filed a bug report that "svn merge" when
scheduling an add, should check whether the source repository and
destination repository are the same or not, and do addition-with-history or
addition-without-history correspondingly.

Maybe this isn't what you've intended for "svn merge" to be able to do, but
in many cases it works OK, and it's useful, and there's no conceptual reason
it shouldn't work.

Dale

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Tue May 3 16:28:10 2005

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

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