inconsistent branching (svn cp).

[Paweł Sikora <pl...@agmk.net>]

Hi,

i'm wondering why the svn branching produces different history for local vs. remote copy?
please check the attached test.sh and final svn logs from http://pastebin.com/UNKXJCcH.
revision 3 and 4 do the same thing (branch via directory copy) but produce different logs.

an old svn manual http://svnbook.red-bean.com/en/1.0/ch04s02.html#svn-ch-4-sect-2.1
shows 'Adding...' only for top-level branch directory during local copy which looks correct.

is it a bug in recent svn-1.7?

BR,
Paweł.

please CC me on reply.

Re: inconsistent branching (svn cp).

[Paweł Sikora <pl...@agmk.net>]

On Monday 09 of January 2012 10:17:09 Philip Martin wrote:
> Paweł Sikora <pl...@agmk.net> writes:
> 
> > i'm wondering why the svn branching produces different history for
> > local vs. remote copy?  please check the attached test.sh and final
> > svn logs from http://pastebin.com/UNKXJCcH.  revision 3 and 4 do the
> > same thing (branch via directory copy) but produce different logs.
> 
> No, they don't produce the same thing. The wc-to-repository copy is
> from a mixed revision wc, so the commit has to do multiple server-side
> copies to reproduce the wc in the reposiitory.

and this recent behaviour exposes a regression in 'merge --reintegrate'.
i've reported this issue on dev@ in 'svn branching/merging regression' thread.

Re: inconsistent branching (svn cp).

[Philip Martin <ph...@wandisco.com>]

Paweł Sikora <pl...@agmk.net> writes:

> i'm wondering why the svn branching produces different history for
> local vs. remote copy?  please check the attached test.sh and final
> svn logs from http://pastebin.com/UNKXJCcH.  revision 3 and 4 do the
> same thing (branch via directory copy) but produce different logs.

No, they don't produce the same thing. The wc-to-repository copy is
from a mixed revision wc, so the commit has to do multiple server-side
copies to reproduce the wc in the reposiitory.

-- 
Philip