Also worth noting that the current trunk behaves the same way 1.7.x does.
On Sep 27, 2011 11:50 AM, "Arwin Arni" <arwin_at_collab.net> wrote:
>
> Hi All,
>
> Here is the issue:
>
> Repo contains:
>
> .
> `-- A
> `-- B
>
> This was committed by the user 'admin'.
>
> Authz rules are:
> [/]
> admin=rw
> arwin=rw
> [/A/B]
> admin=rw
> arwin=
>
> ---------------------------------------------------------------
> Using 1.6.x:
>
> $svn co http://r52/svn/cp_fail_repo --username=arwin wc_fail
> A wc_fail/A
> Checked out revision 1.
>
> $cd wc_fail
> $svn cp A C
> svn: Directory 'C/B' is missing
> svn: Directory 'C/B' is missing
>
> $svn st
> ? C
>
> $svn info C
> Path: C
> URL: http://r52/svn/cp_fail_repo/A
> Repository Root: http://r52/svn/cp_fail_repo
> Repository UUID: 0a0eb43e-3f08-4eaa-9b65-b9cc7530273a
> Revision: 1
> Node Kind: directory
> Schedule: normal
> Last Changed Author: admin
> Last Changed Rev: 1
> Last Changed Date: 2011-09-27 11:22:49 +0530 (Tue, 27 Sep 2011)
>
> ---------------------------------------------------------------
> Using 1.7.x:
>
> $svn co http://r52/svn/cp_fail_repo --username=arwin wc_fail
> A wc_fail/A
> Checked out revision 1.
>
> $cd wc_fail
> $svn cp A C
> A C
> svn: E155035: Cannot copy '/tmp/1.7.x/cp_fail_wcs/wc_fail/A' excluded by
server
>
> $svn st
> A + C
> ! C/B
>
> $svn info C
> Path: C
> Working Copy Root Path: /tmp/1.7.x/cp_fail_wcs/wc_fail
> URL: http://r52/svn/cp_fail_repo/C
> Repository Root: http://r52/svn/cp_fail_repo
> Repository UUID: 43eb70fe-f340-49b2-bf98-988167f4544c
> Revision: 1
> Node Kind: directory
> Schedule: add
> Copied From URL: http://r52/svn/cp_fail_repo/A
> Copied From Rev: 1
> Last Changed Author: admin
> Last Changed Rev: 1
> Last Changed Date: 2011-09-27 11:32:41 +0530 (Tue, 27 Sep 2011)
>
> $svn info C/B
> Segmentation fault
> $
>
> There is more to this issue:
> svn cp ^/A C
> svn cp A ^/C
> svn cp ^/A ^C
>
> All have different and interesting errors.
>
> I've attached a shell script which will help you reproduce this easily.
Also, I'm in the process of writing a test-case for this, but I don't know
which set of tests it belongs to - authz_tests.py or copy_tests.py or
info_tests.py. Clearly, the authz subsystem is doing what it is meant to do,
but on the working copy side, 'copy' is doing something nasty, which I
believe causes 'info' to Segfault.
>
> Regards,
> Arwin Arni
Received on 2011-09-27 12:59:21 CEST