As part of changing manifest.diff to accept a matcher, a previous patch added
matcher calls to each location in merge.manifestmerge that tested if 'x in mf'
to maintain the same behavior as before. After analyzing it further, this
matcher call isn't needed, and in fact hurts future patches ability to use the
matcher here.
Basically, all these 'if x in mf' checks were checking if a matched file's copy
source was in the matcher as well. This meant if you passed a matcher for just
file foo, it would not return file bar even if foo was a copy of bar. Since
manifestmerge cares about copy information, let's allow all lookups of copy
sources.
We also update one spot with a 'is not None' check, since it wasn't obvious that
the value could sometimes be None before, which broke when we called
matcher(None).
A future patch adds matcher optimizations to manifestmerge which causes this
code path to get covered by existing tests.
Recently we switched rebases to run the entire rebase inside a single
transaction, which dramatically improved the speed of rebases in repos with
large working copies. Let's also move the dirstate into a single dirstateguard
to get the same benefits. This let's us avoid serializing the dirstate after
each commit.
In a large repo, rebasing 27 commits is sped up by about 20%.
I believe the test changes are because us touching the dirstate gave the
transaction something to actually rollback.
This adds an option (which defaults to False) to run entire histedits in a
single transaction. This results in 20-25% faster histedits in large repos where
transaction startup cost is expensive.
I didn't want to enable this by default because it has some unfortunate side
effects. For instance, if a pretxncommit hook throws midway through the
histedit, it will rollback the entire histedit and lose any progress the user
had made. Same if the user aborts editting a commit message. It's still worth
turning this on for large repos, but probably not for normal sized repos.
Long term, once we have inmemory merging, we could do the entire histedit in
memory, without a transaction, then we could selectively rollback just parts of
it in the event of an exception.
Tested it by running the tests with
`--extra-config-opt=histedit.singletransaction=True`. The only failure was
related to the hook rollback issue I mention above.
We only want to pop the action after the action is completed, since if the
action aborts part way through we want it to remain at the front of the list so
continue/abort will start with it.
Previously we relied on the fact that we only serialized the state file at the
beginning of the action, so the pop wasn't serialized until the next iteration
of the loop. In a future patch we will be adding a large transaction around this
area, which means if we pop the list early it might get serialized if the action
throws a user InterventionRequired error, at which point the action is not in
the list anymore. So let's only pop it once the action is really truly done.
On Python 3, keys() is more like iterkeys(), so we got in trouble for
mutating the dict while we're iterating here. Since the list of caches
should be relatively small, work around this difference by just
forcing a copy of the key list.
The quoting logic here was actually insufficient, and would have had
bogus b-prefixes on Python 3. shellquote seems more appropriate
anyway. Surprisingly, only two tests have output changes, and both of
them look reasonable to me (both are in blackbox logs).
Spotted by Yuya during review.
I originally noticed that log output wasn't being colored after 348863ccec7e,
but there were other complications too. With a bunch of untracked files, only
the first 1K of characters were colored pink, and the rest were normal white. A
single modified file at the top would also be colored pink.
Line buffering and full buffering are treated as the same thing in Windows [1],
meaning the stream is either buffered or not. I can't find any explicit
documentation to say stdout is unbuffered by default when attached to a console
(but some internet postings indicated that is the case[2]). Therefore, it seems
that explicit flushes are better than just not reopening stdout.
NB: pager is now on by default, and needs to be disabled to see any color on
Windows.
[1] https://msdn.microsoft.com/en-us/library/86cebhfs(v=vs.140).aspx
[2] https://sourceforge.net/p/mingw/mailman/message/27121137/
The initial problem was `hg files` prints paths with '\', which gets removed
when piped (scanhelptopics.py failed to open 'hgext__init__.py'). Then, xargs
was invoking `hg help` with 'backout\r (esc)', which setting binary mode
prevents.
I think this makes the code much clearer. I had to think for a bit to
unpack the old-school `condition and if-true or if-false` dance, and
formatting argument lists here shouldn't be performance critical.
read() should never raise EOFError. If it did, UnboundLocalError would occur
due to unassigned 'ret' variable.
This issue was originally reported to TortoiseHg:
https://bitbucket.org/tortoisehg/thg/issues/4707/
Previously, `hg bundle zstd` on a non-generaldelta repo would
attempt to use a v1 bundle. This would fail because zstd is not
supported on v1 bundles.
This patch changes the behavior to automatically use a v2 bundle
when the user explicitly requests a bundlespec that is a compression
engine not supported on v1. If the bundlespec is <engine>-v1, it is
still explicitly rejected because that request cannot be fulfilled.
Version 1 bundles only support a fixed set of compression engines.
Before this change, we would accept any compression engine for v1
bundles, even those that may not work on v1. This could lead to
an error.
We define a fixed set of compression engines known to work with v1
bundles and we add checking to ensure a newer engine (like zstd)
won't work with v1 bundles.
I also took the liberty of adding test coverage for unknown compression
names because I noticed we didn't have coverage of it before.
This is useful for when repositories are nested in --web-conf, and in the future
with hosted subrepositories. The previous behavior was only to render an index
at each virtual directory. There is now an explicit 'index' child for each
virtual directory. The name was suggested by Yuya, for consistency with the
other method names.
Additionally, there is now an explicit 'index' child for every repository
directory with a nested repository somewhere below it. This seems more
consistent with each virtual directory hosting an index, and more discoverable
than to only have an index for a directory that directly hosts a nested
repository. I couldn't figure out how to close the loop and provide one in each
directory without a deeper nested repository, without blocking a committed
'index' file. Keeping that seems better than rendering an empty index.
The list parser is complex and reusable without ui. Let's move it to
config.py.
This allows us to parse a list from a "pure" config object without going
through ui. Like, we can make "_trustusers" calculated from raw configs,
instead of making sure it's synchronized by calling "fixconfig"s.
My environment (Python version? PYTHONPATH? something else?) raises
ModuleNotFoundError in test-check-py3-compat.t. This patch allows any
"*Error". The error string contains "error importing", so it seems
specific enough even after.
man(1) behaves as poorly as Mercurial without this change. This cribs
from git's run-command[0], which has a list of characters that imply a
string that needs to be run using 'sh -c'. If none of those characters
are present in the command string, we can use shell=False mode on
subprocess and get significantly better error messages (see the test)
when the pager process is invalid. With a complicated pager command
(that contains one of the unsafe characters), we behave as we do today
(which is no worse than git manages.)
I briefly tried tapdancing in a thread to catch early pager exits, but
it's just too perilous: you get races between fd duping operations and
a bad pager exiting, and it's too hard to differentiate between a
slow-bad-pager result and a fast-human-quit-pager-early result.
I've observed some weird variation in exit code handling in the "bad
experience" case in test-pager.t: on my Mac hg predictably exits
nonzero, but on Linux hg always exits zero in that case. For now,
we'll work around it with || true. :(
0: cddbda4bc8/run-command.c (L201)
With experimental.updatecheck=noconflict set, if one checks out
e1870a31325e (tests: add execute bit and fix shbang line, 2015-12-22)
and then try to check out its parent, hg will complain about
conflicting changes, even though the working directory is clean. We
need to also allow the 'e' action in merge.py. The 'e' action is used
when moving to a commit where the only change to the file is to its
executable flag, so it's just an optimized 'g' action.
Doesn't seem to be worth writing a test for, since the existing setup
in test-update-branches.t does not set any flags.