Joe Schaefer <joe_schaefer@yahoo.com> (joes)


Found Patch Suggested
r1382864 r1429718, r1300581, r1300403, r1300384, r1300276, r1300258, r1300242, r1300237, r1299519, r1299319, r1299317, r1298258, r1298021, r1297853, r1297848, r1297847, r1297824, r1295718, r1294796, r863889, r836401 r859639

r836401 | joes | 2009-11-15 19:53:16 +0000 (Sun, 15 Nov 2009)

create subversion dir

r859639 | rooneg | 2006-05-08 21:00:12 +0000 (Mon, 08 May 2006)

Clean up our XML parser via a pool cleanup callback instead of doing it
manually.  This will be more resiliant in the face of errors.

Suggested by: Joe Schaefer <joe@sunstarsys.com>

* contrib/server-side/mod_dontdothat/mod_dontdothat.c
  (dontdothat_filter): Stop cleaning up the parser.
  (clean_up_parser): New helper function.
  (dontdothat_insert_filters): Register a cleanup function to free the
   parser.


r863889 | danderson | 2007-03-14 15:51:08 +0000 (Wed, 14 Mar 2007)

Fix a bug in the authz code, related to issue #2388: if the authz global access
check is inconclusive (ie. no decision could be made), access should be denied, 
not authorized.

Patch by: Joe Schaefer <joe@sunstarsys.com>

* subversion/libsvn_repos/authz.c
  (authz_global_parse_section): Continue walking the authz config as long as
    the access is not determined.
  (authz_get_global_access): Return FALSE if the access could not be
    conclusively determined.

* subversion/tests/libsvn_repos/repos-test.c
  (authz): Add a section to the test configuration that triggers the bug. Fix
    the test loop to actually run all unit tests.


r1294796 | joes | 2012-02-28 19:17:52 +0000 (Tue, 28 Feb 2012)

+= joes


r1295718 | joes | 2012-03-01 18:12:28 +0000 (Thu, 01 Mar 2012)

add this back for the time being so we can continue running trunk in production at apache

r1297824 | joes | 2012-03-07 01:26:25 +0000 (Wed, 07 Mar 2012)

specify conf file

r1297847 | joes | 2012-03-07 03:18:01 +0000 (Wed, 07 Mar 2012)

Set gid first before dropping uid to non-root user.

* tools/server-side/svnpubsub/svnwcsub.py:
  (handle_options): set gid first

[accidental change to #! line reverted in r1297848]


r1297848 | joes | 2012-03-07 03:18:57 +0000 (Wed, 07 Mar 2012)

* tools/server-side/svnpubsub/svnwsub.py:
  (...): revert accidential change to the #! line


r1297853 | joes | 2012-03-07 03:38:12 +0000 (Wed, 07 Mar 2012)

* svnwcsub.py - drop --config-dir options as we now properly pass env


r1298021 | joes | 2012-03-07 17:02:09 +0000 (Wed, 07 Mar 2012)

* trunk/tools/server-side/svnpubsub/rc.d/svnwcsub* - make logfile and config file options configurable

r1298258 | joes | 2012-03-08 02:06:56 +0000 (Thu, 08 Mar 2012)

* svnpubsub/rc.d/svnwcsub* - need independent log dir due to new logrotation feature

r1299317 | joes | 2012-03-11 00:51:08 +0000 (Sun, 11 Mar 2012)

* trunk/tools/server-side/svnpubsub/svnwcsub.py
 need to reinstall default SIGCHLD handler for python before launching subprocesses (changed by daemonize.py)

r1299319 | joes | 2012-03-11 01:22:34 +0000 (Sun, 11 Mar 2012)

* trunk/tools/server-side/svnpubsub svnwcsub.py daemonize.py
 need to reinstall default sighandlers in the second child (the actual daemon process)

r1299519 | joes | 2012-03-12 02:02:37 +0000 (Mon, 12 Mar 2012)

* trunk/tools/server-side/svnpubsub/svnpubsub/client.py
Force the reconnect to eventually happen.  There is probably a better way of doing this, but for now this code is both necessary and sufficient.

r1300237 | joes | 2012-03-13 17:01:23 +0000 (Tue, 13 Mar 2012)

* trunk/tools/server-side/svnpubsub/svnpubsub/client.py
handle connect() exceptions a bit cleaner


r1300242 | joes | 2012-03-13 17:17:08 +0000 (Tue, 13 Mar 2012)

* trunk/tools/server-side/svnpubsub/README
todos


r1300258 | joes | 2012-03-13 17:55:56 +0000 (Tue, 13 Mar 2012)

* trunk/tools/server-side/svnpubsub/README
configurable LOGLEVEL plz


r1300276 | joes | 2012-03-13 18:30:15 +0000 (Tue, 13 Mar 2012)

* trunk/tools/server-side/svnpubsub/README
can get away with excluding certain base paths, probably don't need regexps


r1300384 | joes | 2012-03-13 22:28:49 +0000 (Tue, 13 Mar 2012)

* svnpubsub/README

cause is starting to look switch-related instead of host-related


r1300403 | joes | 2012-03-13 23:17:44 +0000 (Tue, 13 Mar 2012)

* svnpubsub/client.py

not good enough- 'stale' kicks in because it thinks the socket is available.
lets try closing the socket even tho connect has failed


r1300581 | joes | 2012-03-14 15:05:04 +0000 (Wed, 14 Mar 2012)

* trunk/tools/server-side/README.txt

Reconnect logic is working well now- can cross this off the todo list.


r1382864 | danielsh | 2012-09-10 14:18:51 +0000 (Mon, 10 Sep 2012)

[in tools/server-side/svnpubsub]

Fix a bogus logging.critical().

Found by: joes

* svnwcsub.py
  (BackgroundWorker.run): Fix if/else chain.


r1429718 | joes | 2013-01-07 09:15:11 +0000 (Mon, 07 Jan 2013)

* svnwcsub.py
   Better handling of nested directory checkouts.