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

Re: Backing out changes: the prefered method?

From: Karl Fogel <kfogel_at_newton.ch.collab.net>
Date: 2002-01-31 00:10:34 CET

Kevin Pilch-Bisson <kevin@pilch-bisson.net> writes:
> All well and good, but I wanted to bring up something that seems to have been
> forgotten. How will svn blame deal with reverted code. It would be super
> nifty there was a way to make it go back to the way it was before, instead of
> making it appear that you made a change, when really you just put it back to
> the way someone else made it.
>
> I guess it's not the end of the world if it doesn't happen that way, but it
> would definitely be cool.

I think this can be done the same way we're planning to track
previously merged changes (i.e., "ancestry sets", see the design docs)
for "svn merge". A node's ancestry set is:

   1. all the changes implied by its node revision number
   2. plus any other changes added explicitly in the ancestry property
   3. minues any changes the ancestry prop *subtracts*

Step 3 is not fundamentally different from step 2, and we'd already
been planning step 2. Of course, this brings us a little closer to
changesets, which are a sort of reverse singularity for Subversion --
each step closer to real changeset support is harder than the previous
step. But that's no reason not to do what we can.

-K

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sat Oct 21 14:37:01 2006

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.