Mads Kiilerich <madski@unity3d.com> [Sat, 08 Oct 2016 00:59:41 +0200] rev 30142
largefiles: use context for file closing
Make the code slightly smaller and safer (and more deeply indented).
Mads Kiilerich <madski@unity3d.com> [Sat, 08 Oct 2016 00:59:40 +0200] rev 30141
largefiles: when setting/clearing x bit on largefiles, don't change other bits
It is only the X bit that it matters to copy from the standin to the largefile
in the working directory. While it generally doesn't do any harm to copy the
whole mode, it is also "wrong" to copy more than the X bit we care about. It
can make a difference if someone should try to handle largefiles differently,
such as marking them read-only.
Thus, do similar to what utils.setflags does and set the X bit where there are
R bits and obey umask.
Mads Kiilerich <madski@unity3d.com> [Sun, 09 Oct 2016 15:54:49 +0200] rev 30140
eol: on update, only re-check files if filtering changed
Before, update would mark all files as 'normallookup' in dirstate if .hgeol
changed so all files would get the new filtering applied. That takes some time
... and is pointless if the filtering for that file didn't change.
Instead, keep track of the old filtering and only check files where the
filtering is changed.
To keep the old filtering, change to write the applied .hgeol content to
.hg/eol.cache instead of just touching it. That change is backwards/forwards
compatible.
In a real world test, this takes an update that is changing .hgeol and 30000
files from 12s to 4s - where the remaining eol overhead is 1-2s.
Gregory Szorc <gregory.szorc@gmail.com> [Thu, 13 Oct 2016 10:59:29 +0200] rev 30139
dirs: add comment about _PyBytes_Resize
So readers have a canonical function to compare this code to.
Pierre-Yves David <pierre-yves.david@ens-lyon.org> [Tue, 11 Oct 2016 01:29:08 +0200] rev 30138
checkcopies: extract the '_related' closure
There is not need for it to be a closure.
Pierre-Yves David <pierre-yves.david@ens-lyon.org> [Sat, 08 Oct 2016 23:00:55 +0200] rev 30137
checkcopies: add an inline comment about the '_related' call
This helps understanding the flow of the function.
Pierre-Yves David <pierre-yves.david@ens-lyon.org> [Sat, 08 Oct 2016 19:03:16 +0200] rev 30136
checkcopies: minor change to comment
This helped me understand the refactoring so this must be helpful.
Pierre-Yves David <pierre-yves.david@ens-lyon.org> [Sat, 08 Oct 2016 18:38:42 +0200] rev 30135
checkcopies: rename 'ca' to 'base'
This variable was named after the common ancestor. It is actually the merge
base that might differ from the common ancestor in the graft case. We rename the
variable before a larger refactoring to clarify the situation.
Pierre-Yves David <pierre-yves.david@ens-lyon.org> [Wed, 24 Aug 2016 05:09:46 +0200] rev 30134
bisect: extra a small initialisation outside of a loop
Having initialisation done during the first iteration is cute, but can be
avoided.
Martijn Pieters <mjpieters@fb.com> [Mon, 10 Oct 2016 23:11:15 +0100] rev 30133
pycompat: only accept a bytestring filepath in Python 2
Mateusz Kwapich <mitrandir@fb.com> [Mon, 10 Oct 2016 05:31:31 -0700] rev 30132
py3: use raw strings in line continuation (convert ext)
Our py2 to py3 string translations marks those as bytestrings.
Mateusz Kwapich <mitrandir@fb.com> [Mon, 10 Oct 2016 05:30:14 -0700] rev 30131
py3: namedtuple takes unicode (journal ext)
namedtuple usage consistent with changelog.py:141
Augie Fackler <augie@google.com> [Sun, 09 Oct 2016 09:42:46 -0400] rev 30130
debuginstall: use %d instead of %s for formatting an int
% formatting on bytes on Python 3 is pickier about which % character
we specify.
Pulkit Goyal <7895pulkit@gmail.com> [Sun, 09 Oct 2016 13:59:20 +0200] rev 30129
py3: test to check which commands run
This test helps us to keep track on the commands which runs to Python 3.
The full traceback is hidden. Thanks to Augie and Martijn to wrap it up
in four lines.
Pierre-Yves David <pierre-yves.david@ens-lyon.org> [Wed, 24 Aug 2016 05:06:21 +0200] rev 30128
bisect: build a displayer only once
There is multiple spot using this, building it early will help to extract more
of the logic into the bisect module.
Pierre-Yves David <pierre-yves.david@ens-lyon.org> [Wed, 24 Aug 2016 05:04:46 +0200] rev 30127
bisect: factor commonly update sequence
For now, This remains a closure in the module to avoid circular import with used
module.
Pierre-Yves David <pierre-yves.david@ens-lyon.org> [Wed, 24 Aug 2016 04:25:20 +0200] rev 30126
bisect: move check_state into the bisect module
Now that the function is simpler, we resume our quest to move the logic into the
bisect module. In the process, we add basic documentation.
Pierre-Yves David <pierre-yves.david@ens-lyon.org> [Wed, 24 Aug 2016 04:23:13 +0200] rev 30125
bisect: simplify conditional in 'check_state'
Now that extra code about "updating" flag have been removed, we can simplify the
condition flow and remove a level.
Pierre-Yves David <pierre-yves.david@ens-lyon.org> [Wed, 24 Aug 2016 04:22:40 +0200] rev 30124
bisect: remove code about "update-flag" in check_state
Now that the flag dedicated to updating the flag are handled earlier, we do not
need to handle them in the 'check_state' function.
Pierre-Yves David <pierre-yves.david@ens-lyon.org> [Sun, 09 Oct 2016 03:50:55 +0200] rev 30123
bisect: rename 'check_code' to match our naming scheme
We need to do it early, otherwise 'check-commit' will complain every time we
touch it.
Pierre-Yves David <pierre-yves.david@ens-lyon.org> [Wed, 24 Aug 2016 04:48:17 +0200] rev 30122
bisect: minor movement of code handle flag updating state
The code flag handling is quite complicated, we are moving code around to
prepare further simplification.
Augie Fackler <augie@google.com> [Sun, 09 Oct 2016 04:37:02 -0400] rev 30121
test-clone: discard lock-related messages
We can't predict where those will show up and they're not
super-important for the contents of this particular test, so just drop
them. Further reduces the flakiness of the test to zero.
Augie Fackler <augie@google.com> [Sun, 09 Oct 2016 04:25:22 -0400] rev 30120
test-clone: fix some instability in pooled clone race condition test
Healthy output (one log file mentioning "existing pooled" and one
mentioning "new pooled") will now print in a stable order, but
unhealthy output will print some sort of error.
This reduces the flakiness of the test from 55% to 38%. My next patch
makes it completely stable.
Martijn Pieters <mjpieters@fb.com> [Sun, 09 Oct 2016 17:44:23 +0200] rev 30119
py3: add an os.fsencode backport to ease path handling