Gregory Szorc <gregory.szorc@gmail.com> [Tue, 24 Apr 2018 13:55:25 -0700] rev 37818
hgweb: reuse body file object when hgwebdir calls hgweb (issue5851)
An unintended side-effect of f0a851542a05 was that the request body
file object (which uses a util.cappedreader) was constructed twice
when hgwebdir called into hgweb. Since we attempt to read all remaining
data from this file object when Content-Length is defined and since there
were two instances of this object and the client supplied no additional
data to read, this resulted in deadlock.
The fix implemented in this commit is to reuse the request body file
object when it is passed from hgwebdir to hgweb.
A test demonstrating `hg clone` and `hg push` via hgwebdir has been
added. Without this patch, the test hangs when doing `hg clone`.
Surprisingly, this must mean that we have effectively no test coverage
of the wire protocol when run via hgwebdir.
Differential Revision: https://phab.mercurial-scm.org/D3427
Pulkit Goyal <7895pulkit@gmail.com> [Wed, 25 Apr 2018 00:26:49 +0530] rev 37817
remotenames: mark the extension as EXPERIMENTAL
I still don't feel confident about locking the behavior of all the things in
the remotenames extension. Moreover the extension was introduced in this cycle
only. Let's mark this extension EXPERIMENTAL for now so that we can change
things especially the storage layer if required in next cycle.
I will like to use cbor at storage layer too.
Differential Revision: https://phab.mercurial-scm.org/D3426
Augie Fackler <augie@google.com> [Tue, 24 Apr 2018 22:47:14 -0400] rev 37816
tests: fix test-check-commit.t when all commits are public
I'm 99% sure this is a portable use of /bin/[, and it seems to fix the
issue I noticed on the buildbot on my machine.
Yuya Nishihara <yuya@tcha.org> [Tue, 24 Apr 2018 21:29:00 +0900] rev 37815
import: fix crash on --exact check of empty commit (issue5702)
Martin von Zweigbergk <martinvonz@google.com> [Tue, 24 Apr 2018 08:20:15 -0700] rev 37814
tests: mark test-check-interfaces.py as requiring a repo
This was failing our 4.6rc1 build like this:
mercurial.error.RepoError: repository /tmp/build-debs.zMTRhC/src-4.6rc1 not found
Differential Revision: https://phab.mercurial-scm.org/D3425
Boris Feld <boris.feld@octobus.net> [Mon, 23 Apr 2018 19:23:18 +0100] rev 37813
sshpeer: reflect actual command activity one handshake
The output from devel-peer-request is expected to give data about request and
roundtrip done to the server. Changeset a9cffd14aa04 changed some of that by
grouping hello and between commands call. However, the old sequence of command
was "emulated" in sshpeer.
Update the sshpeer to reflect this grouping of commands and update the tests
that use it.
Matt Harbison <matt_harbison@yahoo.com> [Mon, 23 Apr 2018 23:24:53 -0400] rev 37812
tests: drop a useless glob in test-infinite-bundlestore.t
With the previous breakage tamed, the lack of test output difference was causing
the test runner to report "no result code from test" because of this glob.
Matt Harbison <matt_harbison@yahoo.com> [Mon, 23 Apr 2018 23:22:52 -0400] rev 37811
infinitepush: ensure fileindex bookmarks use '/' separators (issue5840)
After loading up with status messages, I noticed that the subsequent matcher was
rejecting 'scratch\mybranch' on Windows. No bookmarks were reported back, and
the tests subsequently failed. I did a search for 'match', and nothing else
looks like it needs to be fixed up, but someone who understands this code should
also take a look.
I also tried setting `infinitepush.branchpattern=re:scratch\\.*` in
library-infinitepush.sh without this change, but that didn't work. Still,
should we ban '\' in these bookmarks to avoid confusion? I thought I saw code
that sandwiches a pattern between 're:^' and '.*', so perhaps regex characters
will need special care?
I also noticed comments in externalbundlestore.{read,write} that it won't work
on Windows because of opening an open file. But I don't see a test failure, so
this may lack test coverage.
Gregory Szorc <gregory.szorc@gmail.com> [Sun, 22 Apr 2018 11:54:10 -0700] rev 37810
interfaceutil: module to stub out zope.interface
The startup time of `hg` increased during the 4.6 development cycle. A
cause of that was importing more modules and doing more work at module
import time.
The import of zope.interface and the declaring of various interfaces
is partially responsible for the startup time regression.
Our current usage of zope.interface doesn't do much at run time: we are
merely declaring interfaces and stating that certain types implement
various interfaces. Core Mercurial is not (yet) using of any of
zope.interface features that actually require that interface plumbing be
defined. The only place we actually need the interface metadata is in
test-check-interfaces.py.
This commit establishes a new interfaceutil module. It exposes the subset
of the zope.interface API that we currently use. By default, the APIs
no-op. But if an environment variable is set, we export the real
zope.interface APIs.
Existing importers of zope.interface have been converted to use the new
module. test-check-interfaces.py has been updated to define the
environment variable so the real zope.interface is used.
The net effect of this change is we stop importing 9 zope.interface.*
modules and we no longer perform interface bookkeeping when registering
interfaces.
On my i7-6700K on Linux, a shell loop that runs `hg log -r .` 300 times
on a repo with 1 commit shows a significant CPU time improvement
(average of 4 runs):
4.5: 14.814s
before: 19.028s
after: 16.945s
And with `run-tests.py -j10` (single run):
4.5: ~3100s (~51.7m)
before: ~4450s (~74.2m)
after: ~3980s (~66.3m)
So this claws back about half of the regressions in 4.6.
Differential Revision: https://phab.mercurial-scm.org/D3419
Yuya Nishihara <yuya@tcha.org> [Mon, 23 Apr 2018 21:13:19 +0900] rev 37809
test-fix: normalize precision of mtime copied by 'cp -p'
Appears that MSYS cp only copies mtime in seconds.