On Fri, Feb 05, 2010 at 11:19:09AM -0500, Mark Phippard wrote:
> On Fri, Feb 5, 2010 at 11:06 AM, Todd, David <dtodd_at_irobot.com> wrote:
> > I'm looking for a pointer to a technical explanation of what happens when
> > one uses --reintegrate.
> >
> > A number of my users are of the opinion that it's a fragile mode, due to it
> > complaining about missing ranges in merginfo. It's making it hard to sell
> > use of that facility.
>
> The command is explained here:
>
> http://blogs.open.collab.net/svn/2008/07/subversion-merg.html
>
> There is nothing fragile about it. It is ultimately just a normal SVN
> merge command that has existed since 1.0. The --reintegrate option
> simplifies the syntax you need to provide and more importantly it does
> a bunch of checks before starting the merge to prevent problems. As
> of the latest 1.5.x and 1.6.x releases those checks are all solid. In
> older 1.5.x releases they did have some false positives that would
> prevent the command from running. But now the warnings it reports are
> all legitimate.
Yes.
Most users I've talked to who had trouble using reintegrate didn't understand
how to use it properly. If you use it properly, it works fine.
It's not intuitive at first, however. It requires a bit of thinking to
understand why it's needed and what it really does. In fact, questions
about the reintegrate merge frequently come up on this list. So you're
not alone.
Unfortunately, you did not elaborate on how you are doing reintegrate
merges, so it's hard to give you suitably directed advice.
But many people run into problems when they try to reintegrate the same
branch more than once. If this is what you are trying to do, in addition
to the blog post Mark linked to (and yes, you should definitely read that!),
you can also refer to the following post. It gives a high-level explanation
of what you need to do to reintegrate a branch multiple times, and why:
http://svn.haxx.se/users/archive-2009-12/0377.shtml
Stefan
Received on 2010-02-05 19:04:35 CET