Pulkit Goyal <pulkit@yandex-team.ru> [Tue, 05 Mar 2019 04:32:23 +0530] rev 41849
tests: make test-status.t compatible with test-check-module-imports.t
Otherwise the later fails on Python 3.
Differential Revision: https://phab.mercurial-scm.org/D6072
Pulkit Goyal <pulkit@yandex-team.ru> [Tue, 05 Mar 2019 04:11:32 +0530] rev 41848
py3: convert KEY_PRESSED value to bytes in crecord.py
This was a str before.
Differential Revision: https://phab.mercurial-scm.org/D6071
Augie Fackler <augie@google.com> [Mon, 04 Mar 2019 18:11:50 -0500] rev 41847
hghave: skip emacs tests on 24.3 and earlier
Turns out with-eval-after-load is new in 24.4.
Differential Revision: https://phab.mercurial-scm.org/D6074
Pulkit Goyal <pulkit@yandex-team.ru> [Tue, 05 Mar 2019 03:33:41 +0530] rev 41846
py3: use bytes instead of str in isinstance()
We use bytes everywhere.
Differential Revision: https://phab.mercurial-scm.org/D6070
Joerg Sonnenberger <joerg@bec.de> [Tue, 12 Feb 2019 19:08:17 +0100] rev 41845
server: allow customizing the default repo filter
hgweb has the (undocument) configuration option web.view that allows
restricting visible revisions to immutable. This is useful for serving
the same storage as publishing and non-publishing repo. Add the new
server.view option to serve the same purpose by changing the default
behavior of `getdispatchrepo`. Drop the hard-coded 'served' filter in the
batch handler of v1 of the wire proto, this is a left-over from the days
before `getdispatchrepo` existed.
Differential Revision: https://phab.mercurial-scm.org/D5946
Pulkit Goyal <pulkit@yandex-team.ru> [Sat, 02 Mar 2019 05:24:35 +0530] rev 41844
py3: convert filtername to str if it's None
I have not called pycompat.bytestr() and rather converted the value there
because I am starting to get concerned about the function call overhead of all
this bytes to str or vice versa convert functions.
Differential Revision: https://phab.mercurial-scm.org/D6042
Martin von Zweigbergk <martinvonz@google.com> [Sat, 02 Mar 2019 13:28:17 -0800] rev 41843
walkfilerevs: rename filerevgen() to filerevs() since it's not a generator
Differential Revision: https://phab.mercurial-scm.org/D6053
Gregory Szorc <gregory.szorc@gmail.com> [Sat, 02 Mar 2019 13:15:53 -0800] rev 41842
global: use raw string for setlocale() argument
Otherwise Python 2 will coerce a unicode to str, which fails
on HGUNICODEPEDANTRY=1.
Differential Revision: https://phab.mercurial-scm.org/D6052
Gregory Szorc <gregory.szorc@gmail.com> [Sat, 02 Mar 2019 13:07:58 -0800] rev 41841
encoding: use raw strings for encoding arguments
This prevents the internals of Python from coercing a unicode to str
on Python 2 and makes tests run with HGUNICODEPEDANTRY=1 a lot
happier.
Differential Revision: https://phab.mercurial-scm.org/D6051
Gregory Szorc <gregory.szorc@gmail.com> [Sat, 02 Mar 2019 13:02:39 -0800] rev 41840
revsetlang: use sysbytes() instead of blind encode()
Otherwise we will call str.encode() on Python 2, which is wrong.
sysbytes() does encode('utf-8') on Python 3. But the source is
guaranteed ASCII, so it shouldn't matter.
With this change, `hg` now runs with `HGUNICODEPEDANTRY=1` set.
However, several tests are failing.
Differential Revision: https://phab.mercurial-scm.org/D6050
Gregory Szorc <gregory.szorc@gmail.com> [Sat, 02 Mar 2019 12:57:00 -0800] rev 41839
global: use raw strings for namedtuple()
Otherwise Python 2 will attempt to coerce unicode to str, which
we don't want.
Differential Revision: https://phab.mercurial-scm.org/D6049
Gregory Szorc <gregory.szorc@gmail.com> [Sat, 02 Mar 2019 12:55:29 -0800] rev 41838
attr: don't attempt to .encode() a str on Python 2
Otherwise it coerces automatically.
Differential Revision: https://phab.mercurial-scm.org/D6048
Gregory Szorc <gregory.szorc@gmail.com> [Sat, 02 Mar 2019 12:51:55 -0800] rev 41837
procutil: use a raw string for module name
Otherwise Python 2 will coerce unicode to str.
Differential Revision: https://phab.mercurial-scm.org/D6047
Gregory Szorc <gregory.szorc@gmail.com> [Sat, 02 Mar 2019 12:51:04 -0800] rev 41836
global: use raw strings for __slots__
Otherwise Python 2 will coerce unicode to str at module load time.
We don't like automatic string coercions.
Differential Revision: https://phab.mercurial-scm.org/D6046
Martin von Zweigbergk <martinvonz@google.com> [Thu, 28 Feb 2019 07:45:51 -0800] rev 41835
absorb: run cleanupnodes() within transaction also when not using obsmarkers
scmutil.cleanupnodes() schedules stripping to be done after the
current transaction, so we can safely run it within the
transaction. This also means that the phase will be updated within the
transaction, which I believe means that the initial (possibly
incorrect) phase will not visible.
Differential Revision: https://phab.mercurial-scm.org/D6037
Pulkit Goyal <pulkit@yandex-team.ru> [Sat, 02 Mar 2019 05:23:15 +0530] rev 41834
py3: pass a str to getpass.getpass()
Otherwise I see `password: b''` on the prompt.
# skip-blame because just prefix
Differential Revision: https://phab.mercurial-scm.org/D6041
Pulkit Goyal <pulkit@yandex-team.ru> [Sat, 02 Mar 2019 05:12:45 +0530] rev 41833
py3: port things from chgserver.py
I have installed mercurial on my personal laptop using Python 3.7. I also have
hg aliased to chg. Before this commit, `hg version` didn't work. After this
patch, things are better and now chg works.
Differential Revision: https://phab.mercurial-scm.org/D6040
Pulkit Goyal <pulkit@yandex-team.ru> [Sat, 02 Mar 2019 05:01:00 +0530] rev 41832
py3: make sure return value of posix.groupname() is bytes
Differential Revision: https://phab.mercurial-scm.org/D6039
FUJIWARA Katsunori <foozy@lares.dti.ne.jp> [Fri, 01 Mar 2019 02:53:09 +0900] rev 41831
contrib: make check-code.py check code fragments embedded in test scripts
FUJIWARA Katsunori <foozy@lares.dti.ne.jp> [Fri, 01 Mar 2019 02:53:09 +0900] rev 41830
contrib: add line offset information to file check function of check-code.py
This is a part of preparation to apply checking with check-code.py on
code fragments embedded in *.t test scripts.
This information will be useful to show correct line number in an
actual file for errors detected in code fragments embedded in *.t test
scripts.
FUJIWARA Katsunori <foozy@lares.dti.ne.jp> [Fri, 01 Mar 2019 02:53:09 +0900] rev 41829
contrib: change return value of file checking function of check-code.py
This is a part of preparation to apply checking with check-code.py on
code fragments embedded in *.t test scripts.
After this patch, caller of _checkfiledata() can count number of
errors in each code fragments of an actual file.
This will be useful to share --per-file limitation by all embedded
code fragments.
FUJIWARA Katsunori <foozy@lares.dti.ne.jp> [Fri, 01 Mar 2019 02:53:09 +0900] rev 41828
contrib: factor out actual error check for file data of check-code.py
This is a part of preparation to apply checking with check-code.py on
code fragments embedded in *.t test scripts.
Newly added _checkfiledata() will be useful to apply checks on code
fragments embedded in *.t test scripts.
FUJIWARA Katsunori <foozy@lares.dti.ne.jp> [Fri, 01 Mar 2019 02:53:09 +0900] rev 41827
contrib: refactor preparation logic for patterns of check-code.py
This is a part of preparation to apply checking with check-code.py on
code fragments embedded in *.t test scripts.
Before this patch, preparation logic in _preparepats() of
check-code.py is not reusable. It can handle only module global list
"checks".
This patch splits preparation logic into small internal functions, and
add the loop to invoke them, in order to increase reusability of the
logic.
"c[-2]" is equivalent to "c[3]" for "checks". This patch uses the
former, because it will be more reusable for subsequent patch than the
latter.
FUJIWARA Katsunori <foozy@lares.dti.ne.jp> [Fri, 01 Mar 2019 02:53:05 +0900] rev 41826
contrib: split pypats list in check-code.py
This is a part of preparation to apply checking with check-code.py on
code fragments embedded in *.t test scripts.
Ideally, all patterns in "pypats" before this patch should be applied
on not only normal *.py files but also code fragments embedded in *.t
test scripts. But fixing test scripts for some patterns requires
many changes, and has less profit than effort.
Therefore, this patch splits pypats list into two below:
- commonpypats, which are applied on all (= including code fragments
embedded in *.t test scripts) *.py files
- pypats, which are applied only on normal *.py
FUJIWARA Katsunori <foozy@lares.dti.ne.jp> [Fri, 01 Mar 2019 02:51:52 +0900] rev 41825
tests: remove useless line wrapping in a code fragment embedded in test script
This is a part of preparation to apply checking with check-code.py on
code fragments embedded in *.t test scripts.
This revision avoids error of check-code.py below.
Use () to wrap long lines in Python, not \
Pierre-Yves David <pierre-yves.david@octobus.net> [Wed, 27 Feb 2019 12:40:18 +0100] rev 41824
storage: introduce a `revlog.reuse-external-delta` config
This option goes a bit further and provides a way to get the same behavior as the
`re-delta-all` optimisation from `hg debugupgraderepo`.
The effect of the option is a bit hard to test as we do not have multiple diff
algorithm at hand. However, we at least make sure the code path run.
Pierre-Yves David <pierre-yves.david@octobus.net> [Wed, 27 Feb 2019 10:49:25 +0100] rev 41823
storage: introduce a `revlog.reuse-external-delta-parent` config
As pointed in c59987ab96b4, we had no simple way to get a client to not blindly
reuse the delta parent from a bundle. Instead one had to rely on a side effect
of the deprecated `format.generaldelta` configuration.
We introduce an explicit `revlog.reuse-external-delta-parent` configuration
option (default to True) to control this behavior. If the option is not set,
`format.generaldelta` still control this behavior.
To test the new option, we convert a couple of place where `generaldelta` have
been used for its side effects.
Yuya Nishihara <yuya@tcha.org> [Sat, 02 Mar 2019 09:41:17 +0900] rev 41822
merge with stable