Sun, 13 Mar 2016 23:16:44 +0100 largefiles: replace invocation of os.path module by vfs in lfcommands.py
liscju <piotr.listkiewicz@gmail.com> [Sun, 13 Mar 2016 23:16:44 +0100] rev 28559
largefiles: replace invocation of os.path module by vfs in lfcommands.py
Sun, 13 Mar 2016 02:06:23 +0100 cache: rebuild branch cache from scratch when inconsistencies are detected
Mads Kiilerich <madski@unity3d.com> [Sun, 13 Mar 2016 02:06:23 +0100] rev 28558
cache: rebuild branch cache from scratch when inconsistencies are detected This should recover automatically from some corruptions that for unknown reasons are seen in the wild.
Sun, 13 Mar 2016 02:06:22 +0100 cache: safer handling of failing seek when writing revision branch cache
Mads Kiilerich <madski@unity3d.com> [Sun, 13 Mar 2016 02:06:22 +0100] rev 28557
cache: safer handling of failing seek when writing revision branch cache If the seek for some reason fails (perhaps because the file is too short to search to the requested position), make sure we seek to the start and rewrite everything. It is unknown if this fixes a real problem that ever happened.
Sun, 13 Mar 2016 02:06:21 +0100 cache: remove branch revision file before rewriting the branch name file
Mads Kiilerich <madski@unity3d.com> [Sun, 13 Mar 2016 02:06:21 +0100] rev 28556
cache: remove branch revision file before rewriting the branch name file New branch names are usually appended to the branch name file. If that fails or the file has been modified by another process, it is rewritten. That left a small opportunity that there could be references to non-existent entries in the file while it was rewritten. To avoid that, remove the revision branch cache file with the references to the branch name file before rewriting the branch name file. Worst case, when interrupted at the wrong time, the cache will be lost and rebuilt next time. It is unknown if this fixes a real problem that ever happened.
Mon, 14 Mar 2016 15:05:25 +0000 chg: do not redirect stdout to /dev/null
Jun Wu <quark@fb.com> [Mon, 14 Mar 2016 15:05:25 +0000] rev 28555
chg: do not redirect stdout to /dev/null Redirecting stdout to /dev/null has unwanted side effects, namely ui.write will stop working. This patch removes the redirection code and helps chg to pass test-bad-extension.t.
Mon, 14 Mar 2016 15:03:19 +0000 pager: skip uisetup if chg is detected
Jun Wu <quark@fb.com> [Mon, 14 Mar 2016 15:03:19 +0000] rev 28554
pager: skip uisetup if chg is detected chg has its own pager implementation that it wants to skip pager's uisetup. It is currently done by redirecting stdout to /dev/null, which has unintended side effects. This patch makes pager aware of chg and skip uisetup directly from pager. We may want to merge chg and pager's pager implementation to make this unnecessary in the future.
Mon, 14 Mar 2016 23:08:03 +0000 chgserver: add an explicit gc to trigger __del__
Jun Wu <quark@fb.com> [Mon, 14 Mar 2016 23:08:03 +0000] rev 28553
chgserver: add an explicit gc to trigger __del__ SocketServer.ForkingMixIn uses os._exit which will skip all cleanup handlers. We want to run __del__ to make things like transactions, {ssh,http}peer, atomictempfile, dirstateguard, etc. work. This patch adds a "gc.collect()" to trigger __del__. It is helpful for chg to pass some test cases in test-devel-warnings.t.
Tue, 15 Mar 2016 10:36:02 +0000 blackbox: do not assume self._bb{vfs,repo,fp} are set in blackboxui.__init__
Jun Wu <quark@fb.com> [Tue, 15 Mar 2016 10:36:02 +0000] rev 28552
blackbox: do not assume self._bb{vfs,repo,fp} are set in blackboxui.__init__ It's possible for the blackboxui code to do a "del self._bbvfs", then ui.copy() or similar attempt will fail. It will also fail when constructing a blackboxui from a non-blackbox ui. This patch fixes the issue by not assuming any _bb* attr is set.
Mon, 14 Mar 2016 23:28:14 +0000 chg: downgrade "failed to read channel" from abortmsg to debugmsg
Jun Wu <quark@fb.com> [Mon, 14 Mar 2016 23:28:14 +0000] rev 28551
chg: downgrade "failed to read channel" from abortmsg to debugmsg If the server has an uncaught exception, it will exit without being able to write the channel information. In this case, the client is likely to complain about "failed to read channel", which looks inconsistent with original hg. This patch silences the error message and makes uncaught exception behavior more like original hg. It will help chg to pass test-fileset.t.
Tue, 15 Mar 2016 00:42:33 +0000 histedit: do not close stdin
Jun Wu <quark@fb.com> [Tue, 15 Mar 2016 00:42:33 +0000] rev 28550
histedit: do not close stdin Closing stdin is unexpected by chgserver and is not a good idea generally. This patch refactors related code a bit and make sure stdin is not closed. It will make chg much happier on test-histedit*.t.
(0) -10000 -3000 -1000 -300 -100 -10 +10 +100 +300 +1000 +3000 +10000 tip