Gregory Szorc <gregory.szorc@gmail.com> [Wed, 29 Jun 2016 19:49:39 -0700] rev 29448
tests: test case where default ca certs not available
I'm not a fan of TLS tests not testing both branches of a possible
configuration. While we have test coverage of the inability to validate
a cert later in this file, I insist that we add this branch so
our testing of security code is extra comprehensive.
Gregory Szorc <gregory.szorc@gmail.com> [Wed, 29 Jun 2016 19:38:24 -0700] rev 29447
sslutil: don't load default certificates when they aren't relevant
Before, we would call SSLContext.load_default_certs() when
certificate verification wasn't being used. Since
SSLContext.verify_mode == ssl.CERT_NONE, this would ideally
no-op. However, there is a slim chance the loading of system
certs could cause a failure. Furthermore, this behavior
interfered with a future patch that aims to provide a more
helpful error message when we're unable to load CAs.
The lack of test fallout is hopefully a sign that our
security code and tests are in a relatively good state.
Gregory Szorc <gregory.szorc@gmail.com> [Wed, 29 Jun 2016 19:37:38 -0700] rev 29446
sslutil: display a better error message when CA file loading fails
Before, sslcontext.load_verify_locations() would raise a
ssl.SSLError which would be caught further up the stack and converted
to a urlerror. By that time, we lost track of what actually errored.
Trapping the error here gives users a slightly more actionable error
message.
The behavior between Python <2.7.9 and Python 2.7.9+ differs. This
is because our fake SSLContext class installed on <2.7.9 doesn't
actually do anything during load_verify_locations: it defers actions
until wrap_socket() time. Unfortunately, a number of errors can occur
at wrap_socket() time and we're unable to ascertain what the root
cause is. But that shouldn't stop us from providing better error
messages to people running a modern and secure Python version.
Gregory Szorc <gregory.szorc@gmail.com> [Wed, 29 Jun 2016 18:15:28 -0700] rev 29445
tests: add test for empty CA certs file
smf reported that an environment with no loaded CA certs resulted
in a weird error. I'd like to detect this a bit better so we can
display an actionable error message.
The actual error being globbed over in this patch is "unknown error"
with a ssl.c line number. That isn't useful at all.
Maciej Fijalkowski <fijall@gmail.com> [Mon, 06 Jun 2016 13:08:13 +0200] rev 29444
internals: move the bitmanipulation routines into its own file
This is to allow more flexibility with the C sources -- now the
bitmanipulation routines can be safely imported without importing Python.h
Martijn Pieters <mjpieters@fb.com> [Fri, 24 Jun 2016 16:12:05 +0100] rev 29443
journal: new experimental extension
Records bookmark locations and shows you where bookmarks were located in the
past.
This is the first in a planned series of locations to be recorded; a future
patch will add working copy (dirstate) tracking, and remote bookmarks will be
supported as well, so the journal storage format should be fairly generic to
support those use-cases.
Augie Fackler <raf@durin42.com> [Mon, 27 Jun 2016 11:53:50 -0400] rev 29442
httpclient: update to 54868ef054d2 of httpplus
As of that revision, httpplus fully supports Python 3, including
mimicing all the subtle behavior changes around headers in Python 3's
http.client.
Yuya Nishihara <yuya@tcha.org> [Mon, 27 Jun 2016 20:44:14 +0900] rev 29441
revset: check invalid function syntax "func-name"() explicitly
Before the error was caught at func() as an unknown identifier, and the
optimizer failed to detect the syntax error. This patch introduces getsymbol()
helper to ensure that a string is not allowed as a function name.
Yuya Nishihara <yuya@tcha.org> [Tue, 28 Jun 2016 22:39:06 +0900] rev 29440
chg: silence warning of unused parameter 'sig'
Anton Shestakov <av6@dwimlabs.net> [Wed, 29 Jun 2016 22:48:32 +0800] rev 29439
hgweb: reindent atom/changelogentry.tmpl
It was mixing tabs and spaces, and not in a good way.
Indent style of other atom entries seems to be 1 space per level, so let's
apply it here as well.
Anton Shestakov <av6@dwimlabs.net> [Wed, 29 Jun 2016 21:07:59 +0800] rev 29438
hgweb: remove unused/nonexistent filelogentry from atom/map
filelog in atom style uses changelogentry to show commits, and
atom/filelogentry.tmpl doesn't even exist.
Anton Shestakov <av6@dwimlabs.net> [Wed, 29 Jun 2016 20:58:09 +0800] rev 29437
hgweb: remove unnecessary CDATA markup in atom/branchentry.tmpl
Since content is of type "text" (and is already escaped), using a CDATA section
is not required.
Looks like this was just an artifact of copying things from rss style in
add2f9ddcfb5, because other entries in atom style don't use CDATA in such
places.
Anton Shestakov <av6@dwimlabs.net> [Wed, 29 Jun 2016 20:19:04 +0800] rev 29436
hgweb: reindent rss/changelogentry.tmpl
It was mixing tabs and spaces, and not in a good way.
Indent style of other rss entries seems to be 4 spaces per level, so let's
apply it here as well.
Anton Shestakov <av6@dwimlabs.net> [Wed, 29 Jun 2016 19:40:04 +0800] rev 29435
hgweb: remove a couple of superfluous spaces in rss style
Pulkit Goyal <7895pulkit@gmail.com> [Mon, 27 Jun 2016 19:10:30 +0530] rev 29434
py3: add tests in check-code to load modules from util.py
The conditionalize imports are added in util.py and now we import modules from
there. So adding tests so that someone in future can use that.
Pulkit Goyal <7895pulkit@gmail.com> [Mon, 27 Jun 2016 16:48:54 +0530] rev 29433
py3: conditionalize SocketServer import
The SocketServer is renamed to socketserver in python 3
Pulkit Goyal <7895pulkit@gmail.com> [Mon, 27 Jun 2016 16:37:37 +0530] rev 29432
py3: conditionalize xmlrpclib import
The xmlrpclib library is renamed to xmlrpc.client in python 3
Pulkit Goyal <7895pulkit@gmail.com> [Mon, 27 Jun 2016 16:16:10 +0530] rev 29431
py3: conditionalize the urlparse import
The urlparse library is renamed to urllib.parse in python 3
Pulkit Goyal <7895pulkit@gmail.com> [Mon, 27 Jun 2016 15:53:38 +0530] rev 29430
py3: update tests/test-check-py3-compat.t
The lower part of the test runs with python 3 and hence remain unchanged.
Jun Wu <quark@fb.com> [Fri, 24 Jun 2016 15:21:10 +0100] rev 29429
chg: send SIGPIPE to server immediately when pager exits (issue5278)
If the user press 'q' to leave the 'less' pager, it is expected to end the
hg process immediately. We currently rely on SIGPIPE for this behavior. But
SIGPIPE won't arrive if we don't write anything (like doing heavy
computation, reading from network etc). If that happens, the user will feel
that the hg process just hangs.
The patch address the issue by adding a SIGCHLD signal handler and sends
SIGPIPE to the server as soon as the pager exits.
This is also an issue with hg's pager implementation.
Jun Wu <quark@fb.com> [Fri, 24 Jun 2016 17:06:41 +0100] rev 29428
chgserver: do not ignore SIGPIPE if pager is used
We rely on SIGPIPE to exit when the pager exits. And Python ignores SIGPIPE
by default. Explicitly set SIGPIPE handler to SIG_DFL (terminate) just like
pager.py.
Martin von Zweigbergk <martinvonz@google.com> [Fri, 24 Jun 2016 11:12:41 -0700] rev 29427
debug: make debug{revlog,index,data} --dir not just a flag
The directory argument (for tree manifests) should belong to to the
--dir argument. I had mistakenly made --dir a flag. One effect of this
was that I had meant for "-m" to be optional, but instead it changed
the behavior of --dir, so with "hg debugdata -m --dir dir1 0", the -m
took over and the "dir1" got treated as a revision in the root
manifest log.
Martin von Zweigbergk <martinvonz@google.com> [Fri, 24 Jun 2016 11:25:55 -0700] rev 29426
debugdata: disallow trailing option with -c/-m
Before this change, "hg debugdata -c 0 foo" was allowed.
Yuya Nishihara <yuya@tcha.org> [Sun, 26 Jun 2016 17:16:57 +0900] rev 29425
revset: get rid of redundant error checking from match()
Actually there was no additional error checking. It should be caught by
"not all(specs)".
Martijn Pieters <mjpieters@fb.com> [Fri, 24 Jun 2016 10:32:38 +0100] rev 29424
share: move magic string to a constant
Gregory Szorc <gregory.szorc@gmail.com> [Sat, 12 Mar 2016 16:08:19 -0800] rev 29423
branchmap: remove unused exception variable
Henrik Stuart <henriks@unity3d.com> [Mon, 27 Jun 2016 12:11:18 +0200] rev 29422
patch: allow copy information to be passed in
When displaying patches from graphical tools where you can browse through
individual files, with diff being called separately on each, recomputing the
limits of file copy history can become rather expensive on large repositories.
Instead, we can compute it once and pass it in for subsequent calls.
liscju <piotr.listkiewicz@gmail.com> [Thu, 23 Jun 2016 22:37:17 +0200] rev 29421
largefiles: check file in the repo store before checking remotely (issue5257)
Problem was files to check were gathered in the repository where
the verify was launched but verification was done on the remote
store. It was observed when user committed in cloned repository
and ran verify before pushing - committed files were marked
as non existing.
This commit fixes this by checking in the remote store only files
that are not existing in the repository store where verify was launched.
Solution is similiar to fd288d118074
liscju <piotr.listkiewicz@gmail.com> [Mon, 27 Jun 2016 10:33:33 +0200] rev 29420
largefiles: remove additional blank lines
It does not conform to the coding style.
liscju <piotr.listkiewicz@gmail.com> [Fri, 24 Jun 2016 09:08:16 +0200] rev 29419
largefiles: fix misleading comments in lfutil instore and storepath
Problem in both cases is cache in largefiles has assigned
meaning - user cache which is additional place to get/put
files. Those two function works on store - the main place
to store largefiles in the repository - .hg/largefiles and
using "cache" to describe it is misleading.
Gregory Szorc <gregory.szorc@gmail.com> [Sat, 25 Jun 2016 19:10:46 -0700] rev 29418
revset: implement match() in terms of matchany()
match() is the special case of a single element list being passed
to matchany() with the additional error checking that the revset
spec is defined. Change the implementation to remove the redundant
code and have match() call matchany().
Gregory Szorc <gregory.szorc@gmail.com> [Sat, 25 Jun 2016 19:12:20 -0700] rev 29417
scmutil: improve documentation of revset APIs
I can never remember the differences between the various revset
APIs. I can never remember that scmutil.revrange() is the one I
want to use from user-facing commands.
Add some documentation to clarify this.
While we're here, the argument name for revrange() is changed to
"specs" because that's what it actually is.
Gregory Szorc <gregory.szorc@gmail.com> [Sat, 25 Jun 2016 13:52:46 -0700] rev 29416
mdiff: remove use of __slots__
The use of __slots__ was added way back in 2006 in 4ec58b157265.
__slots__ isn't necessary for this class.
Gregory Szorc <gregory.szorc@gmail.com> [Sat, 25 Jun 2016 17:40:53 -0700] rev 29415
i18n: use unicode literal
Other parts of this expression are already using unicode literals.
We need this to make Python 3 happy and to avoid an implicit
conversion in Python 2.
Gregory Szorc <gregory.szorc@gmail.com> [Sat, 25 Jun 2016 17:22:06 -0700] rev 29414
pycompat: add HTTPPasswordMgrWithDefaultRealm to Python 3 block
Looks like we missed this in 800ec7c048b0.
Gregory Szorc <gregory.szorc@gmail.com> [Sun, 26 Jun 2016 07:59:02 -0700] rev 29413
ui: path option to declare which revisions to push by default
Now that we have a mechanism for declaring path sub-options, we can
start to pile on features!
Many power users have expressed frustration that bare `hg push`
attempts to push all local revisions to the remote. This patch
introduces the "pushrev" path sub-option to control which revisions
are pushed when no "-r" argument is specified.
The value of this sub-option is a revset, naturally.
A future feature addition could potentially introduce a "pushnames"
sub-options that declares the list of names (branches, bookmarks,
topics, etc) to push by default. The entire "what to push by default"
feature should probably be considered before this patch lands.
Gregory Szorc <gregory.szorc@gmail.com> [Sat, 25 Jun 2016 18:35:14 -0700] rev 29412
ui: don't fixup [paths] sub-options
As part of developing a subsequent patch I discovered that sub-option
values like "." were getting converted to paths. This is because the
[paths] section is treated specially during config loading.
This patch prevents post-processing sub-options from the [paths]
section.
Gregory Szorc <gregory.szorc@gmail.com> [Sat, 25 Jun 2016 07:26:43 -0700] rev 29411
sslutil: abort when unable to verify peer connection (BC)
Previously, when we connected to a server and were unable to verify
its certificate against a trusted certificate authority we would
issue a warning and continue to connect. This is obviously not
great behavior because the x509 certificate model is based upon
trust of specific CAs. Failure to enforce that trust erodes security.
This behavior was defined several years ago when Python did not
support loading the system trusted CA store (Python 2.7.9's
backports of Python 3's improvements to the "ssl" module enabled
this).
This commit changes behavior when connecting to abort if the peer
certificate can't be validated. With an empty/default Mercurial
configuration, the peer certificate can be validated if Python is
able to load the system trusted CA store. Environments able to load
the system trusted CA store include:
* Python 2.7.9+ on most platforms and installations
* Python 2.7 distributions with a modern ssl module (e.g. RHEL7's
patched 2.7.5 package)
* Python shipped on OS X
Environments unable to load the system trusted CA store include:
* Python 2.6
* Python 2.7 on many existing Linux installs (because they don't
ship 2.7.9+ or haven't backported modern ssl module)
* Python 2.7.9+ on some installs where Python is unable to locate
the system CA store (this is hopefully rare)
Users of these Pythongs will need to configure Mercurial to load the
system CA store using web.cacerts. This should ideally be performed
by packagers (by setting web.cacerts in the global/system hgrc file).
Where Mercurial packagers aren't setting this, the linked URL in the
new abort message can contain instructions for users.
In the future, we may want to add more code for finding the system
CA store. For example, many Linux distributions have the CA store
at well-known locations (such as /etc/ssl/certs/ca-certificates.crt
in the case of Ubuntu). This will enable CA loading to "just work"
on more Python configurations and will be best for our users since
they won't have to change anything after upgrading to a Mercurial
with this patch.
We may also want to consider distributing a trusted CA store with
Mercurial. Although we should think long and hard about that because
most systems have a global CA store and Mercurial should almost
certainly use the same store used by everything else on the system.
Gregory Szorc <gregory.szorc@gmail.com> [Sat, 25 Jun 2016 07:32:02 -0700] rev 29410
sslutil: remove out of place comment
This comment likely got orphaned as a result of refactoring in this
file. It isn't providing any useful value. So delete it.
liscju <piotr.listkiewicz@gmail.com> [Fri, 24 Jun 2016 11:51:00 +0200] rev 29409
largefiles: remove additional blank line between methods in localstore
According to the coding style it should be a single blank line
between functions.
Martin von Zweigbergk <martinvonz@google.com> [Thu, 23 Jun 2016 12:37:09 -0700] rev 29408
revset: make head() honor order of subset
The ordering of 'x & head()' was broken in 6a1a4c212d50 (revset:
improve head revset performance, 2014-03-13). Presumably due to other
optimizations since then, undoing that change to fix the order does
not slow down the simple case of "hg log -r 'head()'" mentioned in
that commit. I see a small slowdown from ~0.16s to about ~0.19s with
'not 0 & head()', but I'd say it's worth it for the correct output.
Martin von Zweigbergk <martinvonz@google.com> [Thu, 23 Jun 2016 13:08:10 -0700] rev 29407
revsets: use itervalues() where only values are needed
I don't think there will be a noticeable speedup, but it removes an
unused variable.
Martin von Zweigbergk <martinvonz@google.com> [Thu, 23 Jun 2016 12:39:05 -0700] rev 29406
revsets: passing a set to baseset() is not wrong
Since 69c6e9623bdc (revset: force ascending order for baseset
initialized from a set, 2016-04-04), it is safe to pass a revset to a
baseset.
Pierre-Yves David <pierre-yves.david@ens-lyon.org> [Fri, 24 Jun 2016 02:04:43 +0200] rev 29405
pyflakes: use pycompat.pickles to prevent error
The pyflakes in my test box complain about pickle in pycompat.
mercurial/pycompat.py:17: 'pickle' imported but unused
Kostia Balytskyi <ikostia@fb.com> [Tue, 21 Jun 2016 06:29:23 -0700] rev 29404
rebase: move local variable 'obsoletenotrebased' to the RR class
Kostia Balytskyi <ikostia@fb.com> [Tue, 21 Jun 2016 06:29:23 -0700] rev 29403
rebase: move restorestestatus function to be a method of the RR class
Kostia Balytskyi <ikostia@fb.com> [Tue, 21 Jun 2016 06:29:23 -0700] rev 29402
rebase: move local variables related to keeping things unchanged to the RR
This commit moves the following variables, local to the rebase function to be
fields of the rebaseruntime:
-keepf
-keepbranchesf
-keepopen
Kostia Balytskyi <ikostia@fb.com> [Tue, 21 Jun 2016 06:29:23 -0700] rev 29401
rebase: move local variables 'date' and 'extrafns' to the RR class
This commit moves the following variables, local to the rebase function to be
fields of the rebaseruntime:
-date
-extrafns