svk pull explained (was: Re: [svk-devel] 2.0-pre "svk pull" tries to merge back to mirror source?)

David Glasser glasser at
Wed Oct 11 11:37:00 EDT 2006

On 10/11/06, Michael Brouwer <mb.7766 at> wrote:
> Actually svk pull noticed there was a branch to merge from and did just
> that. "...gnucache//branches/swig-redo" is a branch of
> "../gnucache/trunk", so svk pull will bring the swig-redo branch up to date
> with changes made in trunk since the last copy or smerge from trunk to this
> branch, and then update your working copy.
> If you don't want to pull in the latest changes from trunk to a branch, use
> svk update --sync (or svk up -s for short) instead. This will sync your
> mirror and update your wc, but not merge changes to the branch from it's
> copy-from source.
> Note this is not svk2 specific this is just how svk pull has always worked.
> What it sounds like you are asking for is a version of svk pull that behaves
> like svk update --sync unless the branch you are pulling to is a local
> branch.  There isn't currently any precedent for svk command to work
> differently on remote v.s. local branches, but if you can show why this is
> needed and suggest a new command line syntax for it perhaps it can be added.

I don't believe that this is true, consistently, either.  I'm pretty
sure that "svk up -ms" chooses its merge target differently from "svk
up -m" (and "svk pull" is basically "svk up -ms").


David Glasser | glasser at |

More information about the svk-devel mailing list