Augie Fackler <raf@durin42.com> [Thu, 03 Dec 2020 13:36:24 -0500] rev 45782
Added signature for changeset 1d5189a57405
Augie Fackler <raf@durin42.com> [Thu, 03 Dec 2020 13:36:23 -0500] rev 45781
Added tag 5.6.1 for changeset 1d5189a57405
Martin von Zweigbergk <martinvonz@google.com> [Wed, 02 Dec 2020 15:39:01 -0800] rev 45780
rebase: clear merge state when aborting in-memory merge on dirty working copy
Differential Revision: https://phab.mercurial-scm.org/D9509
Martin von Zweigbergk <martinvonz@google.com> [Wed, 02 Dec 2020 15:15:16 -0800] rev 45779
tests: show that in-memory rebase leaves state when working copy is dirty
When in-memory rebase falls back to on-disk rebase, it checks if the
working copy is dirty. If it is, it aborts the rebase. However, it
leaves the rebase state on disk. I broke it in feffeb18d412 (rebase:
teach in-memory rebase to not restart with on-disk rebase on conflict,
2020-09-18).
Differential Revision: https://phab.mercurial-scm.org/D9508
Matt Harbison <matt_harbison@yahoo.com> [Fri, 27 Nov 2020 14:54:37 -0500] rev 45778
extensions: avoid a crash when the version isn't properly byteified on py3
We already force bytestr on the `testedwith` and `buglink` attributes in
dispatch.py when generating a bug report with a similar comment about not
every extension being ported to py3. We should do the same here, so the
function can be properly typed.
Differential Revision: https://phab.mercurial-scm.org/D9433
Matt Harbison <matt_harbison@yahoo.com> [Mon, 23 Nov 2020 11:47:06 -0500] rev 45777
ui: ensure `getpass()` returns bytes
Previously, this could return either bytes or str. I'm not sure which direction
we should go in, but since the input is bytes, I guess bytes makes sense as
output. `debuguigetpass` crashed because it assumed bytes would be returned,
`sslcontext.load_cert_chain()` is happy with bytes or str if the type info in
PyCharm is correct, and `smtplib.SMTP.login()` wants str.
I couldn't figure out how to test this, because the test stalls for input with
`echo test | hg debuguigetpass --config ui.interactive=1`, likely because it
drains stdin before prompting. The custom input reading with `ui.nontty=1` does
not.
I'm also a bit concerned with all of this encoding/decoding. The existing code
in the mail module uses `encoding.strfromlocal()`, but the username and password
are ascii encoded/decoded in `mercurial.url.passwordmgr.find_user_password()`
with `pycompat.{str,bytes}url()`. I'm not sure if this inconsistency could
cause subtle compatability issues.
Differential Revision: https://phab.mercurial-scm.org/D9375
Matt Harbison <matt_harbison@yahoo.com> [Thu, 26 Nov 2020 02:28:42 -0500] rev 45776
packaging: regenerate the Windows requirements manifest on Windows
SecretStorage is a Linux package, and the other stuff removed is a dependency of
it. I assume this was last generated on Linux, and noticed this trying to add
another package and regenerating on Windows.
Differential Revision: https://phab.mercurial-scm.org/D9404
Matt Harbison <matt_harbison@yahoo.com> [Thu, 26 Nov 2020 03:09:56 -0500] rev 45775
pyoxidizer: point to the py3 requirements instead of py2 on Windows
Differential Revision: https://phab.mercurial-scm.org/D9406
Matt Harbison <matt_harbison@yahoo.com> [Sat, 21 Nov 2020 16:55:07 -0500] rev 45774
extensions: gracefully warn when doing min version check with no local version
After doing a `make clean`, I started getting cryptic failures to import
extensions with the `minimumhgversion` attribute on py3:
*** failed to import extension evolve: '>' not supported between instances of 'int' and 'NoneType'
*** failed to import extension topic: '>' not supported between instances of 'int' and 'NoneType'
This now handles the `(None, None)` tuple before comparing, and disables the
extension with the same friendly message as in py2.
Differential Revision: https://phab.mercurial-scm.org/D9363
Yuya Nishihara <yuya@tcha.org> [Sat, 28 Nov 2020 11:15:54 +0900] rev 45773
diff: do not concatenate immutable bytes while building a/b bodies (issue6445)
Use bytearray instead. I don't know what's changed since Python 2, but bytes
concatenation is 100x slow on Python 3.
% python2.7 -m timeit -s "s = b''" "for i in range(10000): s += b'line'"
1000 loops, best of 3: 321 usec per loop
% python3.9 -m timeit -s "s = b''" "for i in range(10000): s += b'line'"
5 loops, best of 5: 39.2 msec per loop
Benchmark using tailwind.css (measuring the fast path, a is empty):
% HGRCPATH=/dev/null python2.7 ./hg log -R /tmp/issue6445 -p --time \
--color=always --config diff.word-diff=true >/dev/null
(prev) time: real 1.580 secs (user 1.560+0.000 sys 0.020+0.000)
(this) time: real 1.610 secs (user 1.570+0.000 sys 0.030+0.000)
% HGRCPATH=/dev/null python3.9 ./hg log -R /tmp/issue6445 -p --time \
--color=always --config diff.word-diff=true >/dev/null
(prev) time: real 114.500 secs (user 114.460+0.000 sys 0.030+0.000)
(this) time: real 2.180 secs (user 2.140+0.000 sys 0.040+0.000)
Benchmark using random tabular text data (not the fast path):
% dd if=/dev/urandom bs=1k count=1000 | hexdump -v -e '16/1 "%3u," "\n"' > ttf
% hg ci -ma
% dd if=/dev/urandom bs=1k count=1000 | hexdump -v -e '16/1 "%3u," "\n"' > ttf
% hg ci -mb
% HGRCPATH=/dev/null python2.7 ./hg log -R /tmp/issue6445 -p --time \
--color=always --config diff.word-diff=true >/dev/null
(prev) time: real 3.240 secs (user 3.040+0.000 sys 0.200+0.000
(this) time: real 3.230 secs (user 3.070+0.000 sys 0.160+0.000)
% HGRCPATH=/dev/null python3.9 ./hg log -R /tmp/issue6445 -p --time \
--color=always --config diff.word-diff=true >/dev/null
(prev) time: real 44.130 secs (user 43.850+0.000 sys 0.270+0.000)
(this) time: real 4.170 secs (user 3.850+0.000 sys 0.310+0.000)
Augie Fackler <augie@google.com> [Tue, 01 Dec 2020 01:18:21 -0500] rev 45772
procutil: use rapply(tonativestr, ...) to preserve lists when they come in
This was broken when script was a list instead of a string. I caught this
with an internal extension at Google, and I'm not really sure why it wasn't
caught in any kind of CI.
Differential Revision: https://phab.mercurial-scm.org/D9471
Yuya Nishihara <yuya@tcha.org> [Tue, 03 Nov 2020 11:24:21 +0900] rev 45771
chg: reset errno prior to calling strtol()
Otherwise we can't figure out if the last strtol() invocation failed or
not.
Yuya Nishihara <yuya@tcha.org> [Tue, 03 Nov 2020 11:15:50 +0900] rev 45770
chg: do not close dir fd while iterating
It works so long as the dp is the last entry, but readdir(dp) would fail
with EBADF. Let's not do that and close the dir fd explicitly.
Yuya Nishihara <yuya@tcha.org> [Tue, 03 Nov 2020 11:12:25 +0900] rev 45769
chg: show debug message for each fd to be closed
It helps debugging. The number of file descriptors should be small in most
cases, so the console output wouldn't get bloated even with CHG_DEBUG=1.
Yuya Nishihara <yuya@tcha.org> [Tue, 03 Nov 2020 11:06:15 +0900] rev 45768
chg: apply clang-format
Mathias De Mare <mathias.de_mare@nokia.com> [Mon, 02 Nov 2020 11:58:34 +0100] rev 45767
chg: close file descriptors when starting the daemon
It's good practice to close file descriptors when forking to start a daemon.
This did not appear to happen yet, which results in flock hanging
in one location in our system (because the chg daemon keeps
the locked file open).
Differential Revision: https://phab.mercurial-scm.org/D9268
Barret Rennie <barret@brennie.ca> [Sat, 31 Oct 2020 17:42:31 -0400] rev 45766
crecord: render chunkpad on Windows (issue6427)
When using Windows wrappers of PDCurses (e.g., windows-curses), the chunkpad
does not render when executing `hg commit -i`. This is due to attempting to
refresh one too many columns of the pad.
Differential Revision: https://phab.mercurial-scm.org/D9267
Augie Fackler <raf@durin42.com> [Mon, 02 Nov 2020 14:26:19 -0500] rev 45765
Added signature for changeset 18c17d63fdab
Augie Fackler <raf@durin42.com> [Mon, 02 Nov 2020 14:26:18 -0500] rev 45764
Added tag 5.6 for changeset 18c17d63fdab
Dan Villiom Podlaski Christiansen <danchr@gmail.com> [Thu, 29 Oct 2020 11:12:25 +0100] rev 45763
i18n: fix coding tag unsupported by xgettext
Running `make update-pot` currently fails with the following error:
xgettext: mercurial/metadata.py:1: Unknown encoding "utf8". Proceeding with ASCII instead.
xgettext: Non-ASCII string at mercurial/metadata.py:311.
Please specify the source encoding through --from-code or through a comment
as specified in http://www.python.org/peps/pep-0263.html.
Differential Revision: https://phab.mercurial-scm.org/D9260
Dan Villiom Podlaski Christiansen <danchr@gmail.com> [Thu, 29 Oct 2020 11:11:54 +0100] rev 45762
py3: hggettext
`itervalues()` is only available on Python 2. Since this script
doesn't currently refer to the main Mercurial modules, and as a result
doesn't have easy access to the `pycompat` module, I simply changed it
to use `value()` instead. Although this allocates a list on Python 2,
I'd consider that acceptable for a utility script; Mercurial doesn't
have all _that_ many commands anyway.
Differential Revision: https://phab.mercurial-scm.org/D9259
Martin von Zweigbergk <martinvonz@google.com> [Thu, 29 Oct 2020 11:37:22 -0700] rev 45761
relnotes: copy "next" to "5.6" and clear "next"
The same procedure as every year^Wcycle.
Differential Revision: https://phab.mercurial-scm.org/D9263
Matt Harbison <matt_harbison@yahoo.com> [Fri, 23 Oct 2020 22:20:08 -0400] rev 45760
repoview: only pin obsolete wdir parents while there are unresolved conflicts
I noticed after doing an update from an obsolete revision with a dirty wdir that
the obsolete commit stayed visible for no obvious reason. It was decided in
85b03b1e4715 not to clear mergestate once all of the conflicts were resolved, in
order to allow re-resolving. Since the point of pinning the obsolete parents
was to allow resolving in the first place (aaeccdb6e654), it makes sense to also
gate it on whether or not there are any remaining files to resolve. This might
result in pinning again if files are marked unresolved again, but that seems
reasonable, given that it still solves the original issue.
Note that this isn't purely cosmetic- pushing with a pinned obsolete revision is
likely to cause complaints about pushing multiple heads or other unexpected
errors. So the faster it comes out of that state, the better.
Differential Revision: https://phab.mercurial-scm.org/D9248
Dan Villiom Podlaski Christiansen <danchr@gmail.com> [Wed, 28 Oct 2020 17:41:25 +0100] rev 45759
backout: don't assume that tip as what we just committed
Differential Revision: https://phab.mercurial-scm.org/D9256
Yuya Nishihara <yuya@tcha.org> [Fri, 23 Oct 2020 20:33:36 +0900] rev 45758
url: do not continue HTTP authentication with user=None (issue6425)
I initially thought this is a py3-compat bug of passwordmgr._writedebug(),
but actually returning (None, str) pair is wrong at all. HTTP authentication
would continue with user="None" in that case.
Since registering a password of user=None should also be wrong, this patch
simply adds early return.
Yuya Nishihara <yuya@tcha.org> [Fri, 23 Oct 2020 20:10:17 +0900] rev 45757
ui: fix echo back of ui.prompt() to not concatenate None as bytes
Spotted while writing tests for the issue6425. The default value may be
None.
Dan Villiom Podlaski Christiansen <danchr@gmail.com> [Thu, 29 Oct 2020 07:51:13 +0100] rev 45756
commit: don't change phases for preexisting commits
I noticed when pulling with hg-git in a repository that already had
the changes, but pulled from another Mercurial repository. This meant
that hg-git would re-create exact matches of the changesets, and
if they were public, they'd get reverted to drafts.
Differential Revision: https://phab.mercurial-scm.org/D9253
Matt Harbison <matt_harbison@yahoo.com> [Thu, 22 Oct 2020 18:38:41 -0400] rev 45755
demandimport: don't raise AttributeError if `exec_module` is missing
I assume this was meant to do the check gracefully. After shoveling a bunch of
modules into the ignore list in order to get keyring to work out of the box on
CentOS 8, I hit the following error accessing the password, which the change
fixes. Now the SecretStorage backend works out of the box, without any edits to
the ignore list.
** Unknown exception encountered with possibly-broken third-party extension mercurial_keyring
** which supports versions unknown of Mercurial.
** Please disable mercurial_keyring and try your action again.
** If that fixes the bug please report it to https://foss.heptapod.net/mercurial/mercurial_keyring/issues
** Python 3.6.8 (default, Apr 16 2020, 01:36:27) [GCC 8.3.1 20191121 (Red Hat 8.3.1-5)]
** Mercurial Distributed SCM (version 5.5.2)
** Extensions loaded: evolve, topic, rebase, absorb, mercurial_keyring
Traceback (most recent call last):
File "/home/mharbison/hg_py3.6.8_venv/lib64/python3.6/site-packages/mercurial_keyring.py", line 230, in _read_password_from_keyring
password = keyring.get_password(KEYRING_SERVICE, pwdkey)
File "/home/mharbison/hg_py3.6.8_venv/lib64/python3.6/site-packages/keyring/core.py", line 53, in get_password
return _keyring_backend.get_password(service_name, username)
File "/home/mharbison/hg_py3.6.8_venv/lib64/python3.6/site-packages/keyring/backends/chainer.py", line 51, in get_password
password = keyring.get_password(service, username)
File "/home/mharbison/hg_py3.6.8_venv/lib64/python3.6/site-packages/keyring/backends/SecretService.py", line 79, in get_password
return item.get_secret().decode('utf-8')
File "/home/mharbison/hg_py3.6.8_venv/lib64/python3.6/site-packages/secretstorage/item.py", line 105, in get_secret
decryptor = Cipher(aes, modes.CBC(aes_iv), default_backend()).decryptor()
File "/home/mharbison/hg_py3.6.8_venv/lib64/python3.6/site-packages/cryptography/hazmat/backends/__init__.py", line 15, in default_backend
from cryptography.hazmat.backends.openssl.backend import backend
File "<frozen importlib._bootstrap>", line 971, in _find_and_load
File "<frozen importlib._bootstrap>", line 955, in _find_and_load_unlocked
File "<frozen importlib._bootstrap>", line 665, in _load_unlocked
File "/home/mharbison/hg_py3.6.8_venv/lib64/python3.6/site-packages/hgdemandimport/demandimportpy3.py", line 53, in exec_module
self.loader.exec_module(module)
File "/home/mharbison/hg_py3.6.8_venv/lib64/python3.6/site-packages/cryptography/hazmat/backends/openssl/__init__.py", line 7, in <module>
from cryptography.hazmat.backends.openssl.backend import backend
File "<frozen importlib._bootstrap>", line 971, in _find_and_load
File "<frozen importlib._bootstrap>", line 955, in _find_and_load_unlocked
File "<frozen importlib._bootstrap>", line 665, in _load_unlocked
File "/home/mharbison/hg_py3.6.8_venv/lib64/python3.6/site-packages/hgdemandimport/demandimportpy3.py", line 53, in exec_module
self.loader.exec_module(module)
File "/home/mharbison/hg_py3.6.8_venv/lib64/python3.6/site-packages/cryptography/hazmat/backends/openssl/backend.py", line 14, in <module>
from six.moves import range
File "<frozen importlib._bootstrap>", line 971, in _find_and_load
File "<frozen importlib._bootstrap>", line 951, in _find_and_load_unlocked
File "<frozen importlib._bootstrap>", line 894, in _find_spec
File "/home/mharbison/hg_py3.6.8_venv/lib64/python3.6/site-packages/hgdemandimport/demandimportpy3.py", line 117, in find_spec
and getattr(spec.loader, "exec_module")
AttributeError: '_SixMetaPathImporter' object has no attribute 'exec_module'
Differential Revision: https://phab.mercurial-scm.org/D9243
Mitchell Plamann <mplamann@janestreet.com> [Thu, 22 Oct 2020 12:31:26 -0400] rev 45754
test: avoid bashisms in test-transaction-rollback-on-sigpipe.t
Differential Revision: https://phab.mercurial-scm.org/D9239
Gregory Szorc <gregory.szorc@gmail.com> [Wed, 21 Oct 2020 22:53:15 -0700] rev 45753
automation: upload Python 3.9 Windows wheels
We are producing these. We should be publishing them.
Gregory Szorc <gregory.szorc@gmail.com> [Wed, 21 Oct 2020 21:53:19 -0700] rev 45752
contrib: split Windows requirements into multiple files
Package support for Python 2 has diverged significantly. It is no
longer trivial to maintain a single requirements file that supports
both Python 2 and 3 because the set of packages and versions varies
wildly.
This commit split up the Windows requirements files so we have
variants for Python 2 and 3. As part of this, I also renamed the
files to have what I believe to be more reasonable naming ("win32"
felt like a weird identifier to me).
We can see that some package versions decreated on 2.7. This is
because the old pinned versions weren't compatible with Python 2.
Yuya Nishihara <yuya@tcha.org> [Sat, 17 Oct 2020 13:06:18 +0900] rev 45751
relnotes: add diffcontains() to new features list