You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by Eric Gillespie <ep...@pretzelnet.org> on 2007/12/19 23:55:49 UTC

dav_svn__replay_report repos_path

mod_dav_svn/replay.c:dav_svn__replay_report has the following:

  const char *base_dir = resource->info->repos_path;
  ...
  /* The request won't have a repos_path if it's for the root. */
  if (! base_dir)
    base_dir = "";

That was added in:

------------------------------------------------------------------------
r19314 | rooneg | 2006-04-11 15:36:28 -0700 (Tue, 11 Apr 2006) | 13 lines

Fix the base path portion of replay over ra_dav.  A while back we switched
to using the repos path for the base path, instead of passing it in as part
of the REPORT body.  This works fine on ra_svn, but ra_dav was using a vcc,
which has no repos path associated with it.

* subversion/libsvn_ra_dav/replay.c
  (svn_ra_dav__replay): Drop the vcc junk, just use the session url for
   the request, ensuring that the correct repos path will be passed on.

* subversion/mod_dav_svn/replay.c
  (dav_svn__replay_report): Use the repos_path for the value of base_dir,
   and handle the fact that it'll be NULL for requests to the root.
------------------------------------------------------------------------

However, it doesn't seem to be true, at least for trunk as of
today, for either neon or serf.  At least, judging from the
operational log after running svnsync_tests.py:

[19/Dec/2007:15:35:46 -0800] jrandom svnsync_tests-1 replay / r1
[19/Dec/2007:15:35:47 -0800] jrandom svnsync_tests-1 replay / r2

and so on.

-- 
Eric Gillespie <*> epg@pretzelnet.org

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org

Re: dav_svn__replay_report repos_path

Posted by Garrett Rooney <ro...@electricjellyfish.net>.
On Dec 19, 2007 6:55 PM, Eric Gillespie <ep...@pretzelnet.org> wrote:
> mod_dav_svn/replay.c:dav_svn__replay_report has the following:
>
>   const char *base_dir = resource->info->repos_path;
>   ...
>   /* The request won't have a repos_path if it's for the root. */
>   if (! base_dir)
>     base_dir = "";
>
> That was added in:
>
> ------------------------------------------------------------------------
> r19314 | rooneg | 2006-04-11 15:36:28 -0700 (Tue, 11 Apr 2006) | 13 lines
>
> Fix the base path portion of replay over ra_dav.  A while back we switched
> to using the repos path for the base path, instead of passing it in as part
> of the REPORT body.  This works fine on ra_svn, but ra_dav was using a vcc,
> which has no repos path associated with it.
>
> * subversion/libsvn_ra_dav/replay.c
>   (svn_ra_dav__replay): Drop the vcc junk, just use the session url for
>    the request, ensuring that the correct repos path will be passed on.
>
> * subversion/mod_dav_svn/replay.c
>   (dav_svn__replay_report): Use the repos_path for the value of base_dir,
>    and handle the fact that it'll be NULL for requests to the root.
> ------------------------------------------------------------------------
>
> However, it doesn't seem to be true, at least for trunk as of
> today, for either neon or serf.  At least, judging from the
> operational log after running svnsync_tests.py:
>
> [19/Dec/2007:15:35:46 -0800] jrandom svnsync_tests-1 replay / r1
> [19/Dec/2007:15:35:47 -0800] jrandom svnsync_tests-1 replay / r2
>
> and so on.

Well, at the time it seemed to be necessary, but it was a while back
and I don't remember the details.

In the end, I don't think it matters much.  If the base dir is /
instead of "" svn_repos_replay2 will simply turn it into "" anyway.
Heck, AFAICT a NULL base_dir would also work just as well, as it does
the check for that as well.

-garrett

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org