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

Re: Quickly showing branch ancestrally?

From: Paul Hammant <paul_at_hammant.org>
Date: Tue, 15 Sep 2015 05:49:29 -0400

If I expand on what you did, I can get the branch mapping like so:

   svn log --verbose --stop-on-copy branchtwo/test.txt | grep " (from "

Gives:

   A /branchtwo (from /branchone:1)

Close enough, but I fear it's not robust.

On Tue, Sep 15, 2015 at 5:21 AM, Stefan Sperling <stsp_at_elego.de> wrote:

> On Tue, Sep 15, 2015 at 05:10:30AM -0400, Paul Hammant wrote:
> > OK, 'svn copy' makes branches. Most would follow the standard
> > trunk/tags/branches model. If you don't though, if you dip your toe into
> > the world of creative/arbitrary branch designs, it is not clear how you
> > catalog your branch mappings. By comparison, Perforce has branch-specs
> > (which are not perfect in themselves).
> >
> > Subversion doesn't make it easy to determine that a branch was formed
> from
> > a certain place. There's a stack overflow article -
> >
> http://stackoverflow.com/questions/4015412/how-to-find-the-common-ancestor-of-two-branches-in-svn
> > - that asks about determining ancestry, with no one-liner answer, and
> > certainly nothing in the subversion command set.
> >
> > Is there no simple "svn show-branch-ancestor url/path" operation in 1.9x?
> >
> > Subversion must know that info, as it is comfortable issuing quick
> > rejections to my commits:
> >
> > svn: E195016: 'svn://127.0.0.1/foo/three/test.txt_at_10' must be
> > ancestrally related to 'svn://127.0.0.1/foo_at_9'
> >
> > Thanks in advance,
>
> svn log --stop-on-copy BRANCH_URL shows the first ancestor of a branch.
>
> That's not a graph of branches, though, and there's currently no interface
> to build one. But it could be built on to of the APIs (i.e. added to the
> svn client or as a standalone tool).
>
> A minor complication being that not all copies are branches.
> But that might not be a very huge issue. Just show all copies.
>
> And it might not perform very well unless the results for the known
> part of history are cached.
>
Received on 2015-09-15 11:49:42 CEST

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