Gregory Szorc <gregory.szorc@gmail.com> [Fri, 10 Aug 2018 14:20:47 -0700] rev 39269
manifest: remove addgroup() from manifestlog and imanifestlog
addgroup() is part of the storage interface for manifests.
Unfortunately, we don't yet have a formal storage interface
for manifests. (One will be established in subsequent commits.)
One thing is for sure, addgroup() doesn't belong on imanifestlog -
at least not unless we extend that interface to encompass storage.
For now, let's access addgroup() on the _revlog attribute, just like
we do for tree manifests. Even though this violates visibility,
it is consistent.
Differential Revision: https://phab.mercurial-scm.org/D4274
Gregory Szorc <gregory.szorc@gmail.com> [Fri, 10 Aug 2018 13:59:27 -0700] rev 39268
repository: clarify role of imanifestlog
Looking at the implementation of manifest.manifestlog, it is highly
tailored towards servicing the root manifest. So clarify that in the
interface docstring.
Differential Revision: https://phab.mercurial-scm.org/D4273
Gregory Szorc <gregory.szorc@gmail.com> [Fri, 10 Aug 2018 13:43:26 -0700] rev 39267
changegroup: change topics during generation
Changegroup generation and consumption currently uses different sets of
topics. Generation uses "bundling" and consumption uses the name of the
thing being consumed.
This commit makes the topic and unit names consistent on both ends.
The reason I chose to change the generation side is because when
"bundling" is used for 3 different stages, it is unclear which stage
of changegroup generation we're in. By making the topics "changesets,"
"manifests," and "files," one can get a better feel for how far along
in the bundling process we are.
Differential Revision: https://phab.mercurial-scm.org/D4272
Gregory Szorc <gregory.szorc@gmail.com> [Fri, 10 Aug 2018 12:44:15 -0700] rev 39266
changegroup: rename mfs to manifests
In the name of readability.
Differential Revision: https://phab.mercurial-scm.org/D4271
Gregory Szorc <gregory.szorc@gmail.com> [Fri, 10 Aug 2018 12:42:00 -0700] rev 39265
changegroup: clean up changelog callback
changelog.read() calls changelog.changelogrevion() then turns it into
a tuple. Let's call changelogrevision() directly and used named
attributes to make code easier to read.
While we're here, also change some variable names to make code easier
to read.
Differential Revision: https://phab.mercurial-scm.org/D4270
Gregory Szorc <gregory.szorc@gmail.com> [Fri, 10 Aug 2018 12:08:45 -0700] rev 39264
changegroup: call rev() on manifestlog instance
rev() is part of the imanifestlog interface and should be used instead
of using the private revlog instance, which is an implementation
detail.
Differential Revision: https://phab.mercurial-scm.org/D4269
Gregory Szorc <gregory.szorc@gmail.com> [Fri, 10 Aug 2018 11:00:06 -0700] rev 39263
manifest: rename dir to tree to avoid shadowing built-in
And update the argument name in the imanifestlog interface.
Differential Revision: https://phab.mercurial-scm.org/D4268
Gregory Szorc <gregory.szorc@gmail.com> [Thu, 09 Aug 2018 19:27:54 -0700] rev 39262
repository: remove candelta() from ifileindex
candelta() was previously called by changegroup code. With delta
generation moved to the storage API, there are no more external
consumers of this method and it can be removed from the storage
interface.
Differential Revision: https://phab.mercurial-scm.org/D4236
Gregory Szorc <gregory.szorc@gmail.com> [Thu, 09 Aug 2018 19:30:36 -0700] rev 39261
changegroup: rename dir to tree to avoid shadowing a built-in
Differential Revision: https://phab.mercurial-scm.org/D4235
Gregory Szorc <gregory.szorc@gmail.com> [Thu, 09 Aug 2018 16:11:24 -0700] rev 39260
repository: remove storedeltachains from ifilestorage
The ifilestorage interface was bootstrapped from requirements of
callers outside the storage implementation (revlogs). I believe we
even made some members public so they could be part of the interface!
Historically, the changegroup code was a gross offender when it
came to accessing low-level storage primitives. There are a handful
of members on the ifilestorage interface that are/were used only
for changegroup code.
With the recent refactor of changegroup code and the establishment
of a formal API on the storage interface for producing revision
deltas, the changegroup code is no longer accessing these low-level
primitives related to delta generation directly. Instead, things
are abstracted away in the storage implementation.
This means we can remove elements from the storage interface that
are no longer needed.
We start with "storedeltachains."
We remove it from the interface. Then we make it a private
attribute and update all references.
.. api:: storedeltachains has been dropped from ifilestorage interface
.. api:: storedeltachains on revlog classes is now _storedeltachains
Differential Revision: https://phab.mercurial-scm.org/D4227
Gregory Szorc <gregory.szorc@gmail.com> [Thu, 09 Aug 2018 16:02:14 -0700] rev 39259
repository: establish API for emitting revision deltas
With our revision delta and revision delta request interfaces
defined, it is now time to define a method on storage interfaces
for using them.
So far, the only storage interface that is well-defined and used
is file storage. So that is the only interface we need to add a
method on.
We define an ``emitrevisiondeltas()`` method that takes an
iterable of ``irevisiondeltarequest``s and turns them into
``irevisiondelta`` instances.
changegroup._handlerevisiondeltarequest() and the looping logic
from changegroup.deltagroup() has effectively been moved to
revlog.emitrevisiondeltas().
Our filelog wrapper class proxies its emitrevisiondeltas() to
the internal revlog instance.
The simple store test extension used to verify sanity of storage
abstractions has also implemented emitrevisiondeltas() for
file storage and the test harness when run with this extension doesn't
seem to exhibit any regressions.
Rather than create a shared type to represent revision deltas,
each storage backend has its own type and the class name identifies
where the revision delta was derived from.
Differential Revision: https://phab.mercurial-scm.org/D4226
Gregory Szorc <gregory.szorc@gmail.com> [Thu, 09 Aug 2018 15:40:14 -0700] rev 39258
repository: formalize interfaces for revision deltas and requests
Now that we've sufficiently abstracted how revision deltas are
produced in changegroup code, we can now start the process of
formalizing that as part of the interfaces defined in the
repository module.
This commit essentially converts the revisiondelta and
revisiondeltarequest classes into well-defined interfaces. This
is not strictly necessary. But I want all types formalized by the
storage interface to have interfaces. This makes it much easier
to test for interface conformance and for implementing new
storage backends.
Because the interface is documented, comments and docstrings from
changegroup.py have been dropped.
Differential Revision: https://phab.mercurial-scm.org/D4225
Gregory Szorc <gregory.szorc@gmail.com> [Thu, 09 Aug 2018 14:31:25 -0700] rev 39257
changegroup: move node sorting into deltagroup()
The 3 callers of deltagroup() all sort nodes and convert to
integer revs immediately before calling deltagroup().
Stepping back a bit, aside from the constraints on node ordering
that always apply (namely that a node must follow its ancestor),
it makes sense for sorting to be an implementation detail of the
store. Moving the sorting into deltagroup() will open the door
for moving sorting into a storage API.
The test for the changelog in deltagroup() to impact the sorting
method is a bit hacky. I suspect this will be resolved once
we establish a formal interface for delta group generation
and port changelog to it.
Differential Revision: https://phab.mercurial-scm.org/D4224
Matt Harbison <matt_harbison@yahoo.com> [Tue, 21 Aug 2018 21:05:15 -0400] rev 39256
cext: fix truncation warnings in revlog on Windows
Matt Harbison <matt_harbison@yahoo.com> [Tue, 21 Aug 2018 20:57:58 -0400] rev 39255
cext: fix revlog compiler error on Windows
Martin von Zweigbergk <martinvonz@google.com> [Sun, 05 Aug 2018 00:42:07 -0700] rev 39254
shortest: use nodetree for finding shortest node within revset
This speeds up `hg log -T '{shortest(node,1)}\n'` in my repo from 12s
to 4.5s. That's very close to the 4.1s it takes without the
disambiguation revset configured. My repo has 69.5k revisions, of
which 550 were in the configured revset ("not public()").
Differential Revision: https://phab.mercurial-scm.org/D4120
Martin von Zweigbergk <martinvonz@google.com> [Mon, 20 Aug 2018 15:57:03 -0700] rev 39253
index: fix a comment about overflow-checking
There's no "argument-checking" done in this method. This was a bad
copy paste.
Differential Revision: https://phab.mercurial-scm.org/D4339
Matt Harbison <matt_harbison@yahoo.com> [Mon, 20 Aug 2018 16:19:36 -0400] rev 39252
tests: demonstrate a problem with renames on the p2 side of a conversion
I think this is related to the octopus merge being sloppy, and that's having a
cascading affect on the fixup merge. If this change is made on p1 (specifically
with the 'Added parent file' commit), the failure doesn't occur.
The file modification with the rename doesn't seem to be necessary, but it's
what's happening in a production repo where I first noticed, so I left it. This
is an example of the manifest divergence I'd been seeing, which wasn't fixed by
Yuya's recent changes. This is separate from the changelog divergence I was
also seeing[1]. Probably nobody cares about bzr anymore, but this will also
affect git, since the octopus fixup code is in the hg sink.
[1] https://www.mercurial-scm.org/pipermail/mercurial-devel/2018-August/120473.html
Matt Harbison <matt_harbison@yahoo.com> [Fri, 13 Jul 2018 15:03:15 -0400] rev 39251
tests: add a test to show the added/modified/removed files in a bzr convert
Matt Harbison <matt_harbison@yahoo.com> [Fri, 13 Jul 2018 12:33:06 -0400] rev 39250
tests: show added/modified/removed files when logging repos converted from bzr
There's a bug related to incorrect status values that can be triggered by an
octopus fixup merge. Seeing what the current status is (instead of just the
aggregate list) will hopefully help track this down.
Matt Harbison <matt_harbison@yahoo.com> [Mon, 20 Aug 2018 22:04:22 -0400] rev 39249
tests: drop an unnecessary redirect to /dev/null
This eliminates the need to conditionalize the color output, because NUL on
Windows is a tty. Credit to Yuya for recognizing this.
Augie Fackler <augie@google.com> [Mon, 20 Aug 2018 11:29:43 -0400] rev 39248
py3: whitelist another test caught by the ratchet
Differential Revision: https://phab.mercurial-scm.org/D4335
Martin von Zweigbergk <martinvonz@google.com> [Fri, 03 Aug 2018 23:03:13 -0700] rev 39247
index: move raise_revlog_error() further up
I will add another caller below it.
Differential Revision: https://phab.mercurial-scm.org/D4119
Martin von Zweigbergk <martinvonz@google.com> [Fri, 06 Jul 2018 07:53:23 -0700] rev 39246
index: make node tree a Python object
Differential Revision: https://phab.mercurial-scm.org/D4118
Martin von Zweigbergk <martinvonz@google.com> [Sun, 19 Aug 2018 22:30:32 -0700] rev 39245
index: rename _fix_index() since it no longer fixes the index
Since c0d411ea6639 (index: drop support for negative indexes into the
index, 2018-07-20), it always returns the input (unless it raises).
Differential Revision: https://phab.mercurial-scm.org/D4334
Martin von Zweigbergk <martinvonz@google.com> [Sat, 18 Aug 2018 00:01:31 -0700] rev 39244
index: don't include nullid in boundary check in pure code
Yet another change that should probably have been part of a3dacabd476b
(index: don't allow index[len(index)] to mean nullid, 2018-07-20).
Differential Revision: https://phab.mercurial-scm.org/D4333
Augie Fackler <augie@google.com> [Mon, 20 Aug 2018 16:13:17 -0400] rev 39243
fastannotate: address pyflakes concerns
Two unused imports, one unused local variable assignment.
Differential Revision: https://phab.mercurial-scm.org/D4337
Augie Fackler <augie@google.com> [Thu, 09 Aug 2018 15:08:32 -0400] rev 39242
fastannotate: move some global state mutation to extsetup()
Differential Revision: https://phab.mercurial-scm.org/D4202
Augie Fackler <augie@google.com> [Thu, 09 Aug 2018 15:05:43 -0400] rev 39241
fastannotate: rip out specialized support for remotefilelog
remotefilelog can choose to collaborate with fastannotate for now if
it needs to, and in the future when we make good on our longstanding
desire to move remotefilelog-like features in-house we'll make sure
things are well-supported via a reasonable interface.
Differential Revision: https://phab.mercurial-scm.org/D4201
Augie Fackler <augie@google.com> [Thu, 09 Aug 2018 14:56:29 -0400] rev 39240
fastannotate: record some TODO items from indygreg's review
Some of these I'll address more or less right away, but some will take
a while for us to resolve.
Differential Revision: https://phab.mercurial-scm.org/D4200
Augie Fackler <augie@google.com> [Wed, 01 Aug 2018 11:10:41 -0400] rev 39239
fastannotate: write out files from server in a predictable order
This is mostly for test output stabilization, but it *might* be
slightly better with a ton of annotate caches because we'll write to
one directory at a time rather than bouncing around randomly.
Differential Revision: https://phab.mercurial-scm.org/D3995
Augie Fackler <augie@google.com> [Mon, 30 Jul 2018 22:50:00 -0400] rev 39238
fastannotate: initial import from Facebook's hg-experimental
I made as few changes as I could to get the tests to pass, but this
was a bit involved due to some churn in the blame code since someone
last gave fastannotate any TLC.
There's still follow-up work here to rip out support for old versions
of hg and to integrate the protocol with modern standards.
Some performance numbers (all on my 2016 MacBook Pro with a 2.6Ghz i7):
Mercurial mercurial/manifest.py
traditional blame
time: real 1.050 secs (user 0.990+0.000 sys 0.060+0.000)
build cache
time: real 5.900 secs (user 5.720+0.000 sys 0.110+0.000)
fastannotate
time: real 0.120 secs (user 0.100+0.000 sys 0.020+0.000)
Mercurial mercurial/localrepo.py
traditional blame
time: real 3.330 secs (user 3.220+0.000 sys 0.070+0.000)
build cache
time: real 30.610 secs (user 30.190+0.000 sys 0.230+0.000)
fastannotate
time: real 0.180 secs (user 0.160+0.000 sys 0.020+0.000)
mozilla-central dom/ipc/ContentParent.cpp
traditional blame
time: real 7.640 secs (user 7.210+0.000 sys 0.380+0.000)
build cache
time: real 98.650 secs (user 97.000+0.000 sys 0.950+0.000)
fastannotate
time: real 1.580 secs (user 1.340+0.000 sys 0.240+0.000)
mozilla-central dom/base/nsDocument.cpp
traditional blame
time: real 17.110 secs (user 16.490+0.000 sys 0.500+0.000)
build cache
time: real 399.750 secs (user 394.520+0.000 sys 2.610+0.000)
fastannotate
time: real 1.780 secs (user 1.530+0.000 sys 0.240+0.000)
So building the cache is expensive (but might be faster with xdiff
enabled), but the blame results are *way* faster.
Differential Revision: https://phab.mercurial-scm.org/D3994
Joerg Sonnenberger <joerg@bec.de> [Thu, 16 Aug 2018 05:50:49 +0200] rev 39237
util: improve handling of truncated compressed streams
If the compressed stream is over as marked by the reader providing
nothing new and the compression engine is not providing data, bail out.
This can happen in practise when the server misbehaves and would result
in an infinite loop otherwise.
Differential Revision: https://phab.mercurial-scm.org/D4297
Anton Shestakov <av6@dwimlabs.net> [Fri, 10 Aug 2018 23:24:33 +0800] rev 39236
zsh_completion: update all options
It's just too hard to further split this patch. What it does:
- adds missing flags
- removes flags that are no longer there
- updates flag descriptions and argument names
- adds * where using the same option many times is okay
- groups with () mutually exclusive flags that can't be used together
- adds + and = to options that need arguments
- removes + and = from options that don't take any arguments
- fixes minor issues and a typo
Differential Revision: https://phab.mercurial-scm.org/D4267
Anton Shestakov <av6@dwimlabs.net> [Fri, 10 Aug 2018 14:46:56 +0800] rev 39235
zsh_completion: run hg with HGPLAINEXCEPT=alias to list commands
This is what contrib/bash_completion does, looks like a sensible thing to do.
Differential Revision: https://phab.mercurial-scm.org/D4266
Anton Shestakov <av6@dwimlabs.net> [Fri, 10 Aug 2018 14:41:41 +0800] rev 39234
zsh_completion: remove completion for obsolete graphlog extension
Differential Revision: https://phab.mercurial-scm.org/D4265
Anton Shestakov <av6@dwimlabs.net> [Fri, 10 Aug 2018 09:09:52 +0800] rev 39233
zsh_completion: declare appropriate local parameters for ->string form
When ->string form is used for _arguments, the function that calls it must
declare appropriate local parameters.
Managing local return value is needed to tell the completion system if our
function succeeded in suggesting something or not, plus without that `hg diff
-<TAB>` doesn't look right.
While at it, fix a copypaste error (s/diff_files/revert_files/).
Docs: http://zsh.sourceforge.net/Doc/Release/Completion-System.html
Differential Revision: https://phab.mercurial-scm.org/D4264
Anton Shestakov <av6@dwimlabs.net> [Fri, 10 Aug 2018 08:59:52 +0800] rev 39232
zsh_completion: complete files from `hg st -mar` for commit and diff
Useful to have and is consistent with our bash completion file.
This is also an improvement coming from upstream, downstream or sidestream
(depends on how you look at it) hg completion file -- the one that lives in
zsh's own repo.
Differential Revision: https://phab.mercurial-scm.org/D4263
Anton Shestakov <av6@dwimlabs.net> [Fri, 10 Aug 2018 08:33:27 +0800] rev 39231
zsh_completion: drop -w, add -S to _arguments
From the docs: http://zsh.sourceforge.net/Doc/Release/Completion-System.html
_arguments -w: "allow option stacking even if one or more of the options take
arguments", so something like "-xy xarg yarg". Mercurial's command line parser
doesn't support that, so the flag got removed in zsh 92584634d3d3 and
bf8b61182043.
_arguments -S: "do not complete options after a ‘--’ appearing on the line". By
options it means things like "-x" and "--foo", but it will still complete
arguments like files.
Differential Revision: https://phab.mercurial-scm.org/D4262
Anton Shestakov <av6@dwimlabs.net> [Fri, 10 Aug 2018 08:19:36 +0800] rev 39230
zsh_completion: rename _hg_style_opts to _hg_template_opts
--style is a deprecated alias for --template, and doesn't show up in help
without --verbose, so completing it doesn't make much sense.
Let's s/style/template/ this variable to be more consistent with cmdutil.py,
which is helpful to look at in parallel next time someone wants to update the
completion rules.
Differential Revision: https://phab.mercurial-scm.org/D4261
Joerg Sonnenberger <joerg@bec.de> [Fri, 17 Aug 2018 00:51:46 +0200] rev 39229
pycompat: wrap xrange for py2 to provide efficient __contains__
The C implementation of xrange in Python 2 provides a O(n) membership
test, which is noticable on pull-based clones of large repositories.
Avoid this by providing a wrapper class with O(1) membership test based
on the edges of the range.
Differential Revision: https://phab.mercurial-scm.org/D4313
Martin von Zweigbergk <martinvonz@google.com> [Mon, 20 Aug 2018 09:48:08 -0700] rev 39228
merge with stable
Yuya Nishihara <yuya@tcha.org> [Sat, 18 Aug 2018 10:06:56 +0900] rev 39227
node: correct hex representation of pseudo node ids
Augie Fackler <augie@google.com> [Sun, 19 Aug 2018 21:53:41 -0400] rev 39226
merge with stable
Boris Feld <boris.feld@octobus.net> [Thu, 16 Aug 2018 04:50:32 +0200] rev 39225
revlog: fix typo in 'buildtext' name
There are no such things as 'deltacomputer._buildtext'. I'm not sure why this
code never ran, but we better fix the typo.
Martin von Zweigbergk <martinvonz@google.com> [Sat, 18 Aug 2018 23:17:06 -0700] rev 39224
revlog: fix pure version of _partialmatch() to include nullid
Before this patch, test-issue842.t and a few more tests fail when they
try to refer to the null revision by using a "000.." prefix of it (or
because they use the "shortest" template function which internally
does that).
This should have been part of my a3dacabd476b (index: don't allow
index[len(index)] to mean nullid, 2018-07-20), but I had forgotten to
update another part of the pure code there, so it didn't fail until
a1f934573c0b (parsers: adjust pure-python version to mimic
a3dacabd476b, 2018-08-09) and 65d5de1169dd (revlog: fix pure nodemap
to not access missing index entry, 2018-08-17) fixed the other things
I had missed.
Differential Revision: https://phab.mercurial-scm.org/D4332
Matt Harbison <matt_harbison@yahoo.com> [Sat, 18 Aug 2018 15:15:04 -0400] rev 39223
tests: conditionalize color output on abort
This broke in afc4ad706f9c. I'm not sure why other aborts don't need to be
conditionalized, but `hg help foo` does abort in red for both msys and cmd.exe.
Matt Harbison <matt_harbison@yahoo.com> [Sat, 18 Aug 2018 14:03:11 -0400] rev 39222
tests: conditionalize out tests that msys is mangling
This was exposed by ac0a87160012. The paths for convert and ssh respectively
are being turned into:
$TESTTMP\bzr+ssh:\foobar@selenic.com\baz\.hg
'C:\\C:\\Users\\Matt\\AppData\\Local\\Temp\\hgtests.8awa9x\\child1\\test-ssh-bundle1.t-sshv2\\nonexistent\\.hg
The errno value raised is 123, and the message is
The filename, directory name, or volume label syntax is incorrect
Matt Harbison <matt_harbison@yahoo.com> [Sat, 18 Aug 2018 13:27:33 -0400] rev 39221
test-absorb: stabilize for no-execbit platforms
Matt Harbison <matt_harbison@yahoo.com> [Sat, 18 Aug 2018 12:35:47 -0400] rev 39220
test-graft: add a missing output line for Windows
Matt Harbison <matt_harbison@yahoo.com> [Sat, 18 Aug 2018 01:07:43 -0400] rev 39219
share: allow more than one level of directories to be created
Matt Harbison <matt_harbison@yahoo.com> [Sat, 18 Aug 2018 00:51:26 -0400] rev 39218
clone: allow local cloning to create more than one level of directories
I figure cloning a remote repository is more common, thus it's more likely that
some people might be relying on the less restrictive behavior. Additionally,
`hg init` will also create more than one level of missing directories.
I also have a use case for reading the subrepos from .hgsub, and sharing them
into the normal nested location on the server. SCM Manager doesn't host
subrepos in the normal nested location, which is nice for deduplicating the
repository data, but confuses `hg verify`. Some of the subrepos are in the root
of the repositories, while others are one or two directories deep. So not
having to build up the parent path first is desirable.
Matt Harbison <matt_harbison@yahoo.com> [Fri, 17 Aug 2018 23:28:37 -0400] rev 39217
tests: demonstrate an inconsistency when cloning to a missing directory tree
I noticed that `hg share` is unable to create more than one missing directory on
the path, and thought it was inconsistent with clone. It turns out that the
path for copying/linking the remote store has the same limitation, but cloning
from a remote repo doesn't.
Matt Harbison <matt_harbison@yahoo.com> [Sat, 18 Aug 2018 23:39:26 -0400] rev 39216
tests: stabilize test-merge-tools.t on Windows
More fun with `hg import --bypass` to work around Windows limitations. The
diffs were generated on Linux, and had a tab to terminate the `+++b/...` lines.
But check-code complained about trailing whitespace, and it seems to run without
them.
Gregory Szorc <gregory.szorc@gmail.com> [Fri, 17 Aug 2018 21:26:34 +0000] rev 39215
dagutil: remove module
The previous commit removed the last consumer of this module.
.. api:: dagutil module has been removed
Some functionality has been moved to the dagop module. Other
functionality can be accomplished via revsets.
Differential Revision: https://phab.mercurial-scm.org/D4330
Gregory Szorc <gregory.szorc@gmail.com> [Fri, 17 Aug 2018 21:21:50 +0000] rev 39214
dagop: port revlogdag.linearize() to standalone function
The code should functionally be identical.
We also port the one consumer in changegroup to use the new
standalone function.
After this commit, dagutil is no longer used!
Differential Revision: https://phab.mercurial-scm.org/D4329
Gregory Szorc <gregory.szorc@gmail.com> [Fri, 17 Aug 2018 19:48:52 +0000] rev 39213
dagutil: use revlog.parentrevs() for resolving parent revisions
And remove parents() since it is no longer used.
revlog.parentrevs() is almost the same as parents(). The main
difference is that parentrevs() can return nullrev. dagop.headrevs()
already handles nullrev. We add an inline check for nullrev in the
other call site to account for the difference.
.. api:: parents() removed from dagutil classes
Use parentrevs() on the storage object instead.
Differential Revision: https://phab.mercurial-scm.org/D4328
Gregory Szorc <gregory.szorc@gmail.com> [Fri, 17 Aug 2018 19:45:13 +0000] rev 39212
dagop: extract headsetofconnecteds() from dagutil
The functionality for resolving the set of DAG heads from a
subset simply requires a function to resolve parent revisions.
Let's establish a function in the dagop module to do this, which
seems to be where generic DAG functionality goes these days.
Differential Revision: https://phab.mercurial-scm.org/D4327
Gregory Szorc <gregory.szorc@gmail.com> [Fri, 17 Aug 2018 19:35:24 +0000] rev 39211
setdiscovery: precompute children revisions to avoid quadratic lookup
Moving away from dagutil a few commits ago introduced quadratic
behavior when resolving children revisions during discovery.
This commit introduces a precompute step of the children revisions
to avoid the bad behavior.
I believe the new code should have near identical performance to
what dagutil was doing before. Behavior is still slightly different
because we take into account filtered revisions. But this change was
made when we moved off dagutil.
I added a comment about multiple invocations of this function
redundantly calculating the children revisions. I believe this
potentially undesirable behavior was present when we used dagutil,
as the call to inverse() previously in this function created a new
object and required computing children on every invocation. I thought
we should document the potential for a performance issue rather than
let it go undocumented.
Differential Revision: https://phab.mercurial-scm.org/D4326
Gregory Szorc <gregory.szorc@gmail.com> [Fri, 17 Aug 2018 19:24:36 +0000] rev 39210
dagutil: remove unused classes
We only directly use revlogdag in changegroup code. We don't need
all this abstraction. So remove various classes and levels
of inheritance.
Differential Revision: https://phab.mercurial-scm.org/D4325