Robert Pluim <rpluim@bigfoot.com>


Patch
r846350, r845596, r845500, r845110

r845110 | philip | 2003-02-23 14:39:10 +0000 (Sun, 23 Feb 2003)

Reduce the memory used by 'svn status <list>'.

Patch by: Robert Pluim <rpluim@bigfoot.com>

* subversion/clients/cmdline/status-cmd.c (svn_cl__status): Use
  and clear a subpool when iterating.


r845500 | kfogel | 2003-03-21 20:56:10 +0000 (Fri, 21 Mar 2003)

Make mod_dav_svn leak a little less.

Patch by: Robert Pluim <rpluim@bigfoot.com>
(From issue #1179.)
(With feedback from Greg Stein <gstein@collab.net>, and a few filigrees by me.)

* svn/subversion/mod_dav_svn/dav_svn.h: 
    (dav_svn_getetag): Add pool argument.

* svn/subversion/mod_dav_svn/repos.c 
  (dav_svn_getetag): Take new pool argument, use it to allocate the
    return value and for temporary work.
  (getetag_pathetic): New wrapper for dav_svn_getetag.
  (dav_svn_set_headers): Pass resource->pool to dav_svn_getetag.
  (dav_svn_hooks_repos): Use getetag_pathetic here.

* svn/subversion/mod_dav_svn/liveprops.c
  (dav_svn_insert_prop): Pass the local subpool to dav_svn_getetag.
    This makes about 100k per prop get allocated in a subpool that gets
    cleared, rather than in resource->pool, thus reducing the peak
    size of the pool (and it's safe, since the result is always copied
    into the response pool).


r845596 | kraai | 2003-04-02 16:25:56 +0000 (Wed, 02 Apr 2003)

* subversion/libsvn_client/diff.c
  (do_diff): Check that the filesystem actually has the item specified
  as URL1@N (URL2 was already being checked).

* subversion/tests/clients/cmdline/diff_tests.py
  (diff_nonextant_urls): New.

Patch by: Robert Pluim <rpluim@bigfoot.com>


r846350 | mbk | 2003-06-18 13:40:50 +0000 (Wed, 18 Jun 2003)

Apply a patch generated by Robert Pluim <rpluim@bigfoot.com> that fixes
a scalability/pool-usasge issue in svn_client_delete noticed by Mats Nilsson
<mats.nilsson@xware.se>.   Thanks, Mats and Robert!

Patch by: Robert Pluim <rpluim@bigfoot.com>

* subversion/libsvn_client/delete.c
  (svn_client_delete): Actually *use* the iteration subpool.