revlog: enforce chunk slicing down to a certain size Limit maximum chunk size to 4x final size when reading a revision from a revlog. We only apply this logic when the target size is known from the revlog. Ideally, revlog's delta chain would be written in a way that does not trigger this extra slicing often. However, having this second guarantee that we won't read unexpectedly large amounts of memory in all cases is important for the future. Future delta chain building algorithms might have good reason to create delta chain with such characteristics. Including this code in core as soon as possible will make Mercurial 4.7 forward-compatible with such improvement.

revlog: postprocess chunk to slice them down to a certain size After the density slicing is done, we enforce a maximum chunk size to avoid memory consumption issue.

revlog: add function to slice chunk down to a given size It is possible to encounter situations where the slicing based on density did not achieve chunk smaller than the 4*textlength limit. To avoid extra memory consumption in those cases, we need to be able to break down chunk to a given size. Actual caller comes in the next changesets.

revlog: extract density based slicing into its own function We are going to introduce another slicing step. We start by extracting the existing one into its own function.

revlog: document and test _slicechunk

revlog: add a doctest for _segmentspan

revlog: add a doctest to _trimchunk

revlog: introduce a tiny mock of a revlog class This will be used in various function doctest added in the next changesets. The class seems small enough to live in the module directly.

revlog: _segmentspan computes the byte span of a segment Refactor out this code to be used somewhere else in a next changeset

revlog: early return in _slicechunk when span is already small enough If the full span is smaller than the minimum gap size we'll consider, we know we won't do any slicing and we can return earlier.

revlog: early return in _slicechunk when density is already good We don't need to do anything if we know we won't enter the second while loop. Save the overhead of the first loop by returning earlier.

sparse-read: discard gap below 65K only Testing on actual data shows that 65K is more efficient in both time and memory than 256K.

sparse-read: target density of 50% instead of 25% The target density value is wrong. The default target chain span is 4*text-length. However, the target max chain payload is 2*text-length. So default target density should be 50% (2/4) not 25% (1/4).

grep: change default behaviour to search working directory files (BC) With this patch, grep searches on the working directory by default and looks for all files tracked by the working directory and greps on them. ### OLD BEHAVIOUR $ hg init a $ cd a $ echo "some text">>file1 $ hg add file1 $ hg commit -m "adds file1" $ hg mv file1 file2 $ hg grep "some" `file2:1:some text` `file1:0:some text` This behaviour is undesirable since file1 is not in the current history and was renamed as file2, so the second result was redundant and confusing. ### NEW BEHAVIOUR $ hg init a $ cd a $ echo "some text">>file1 $ hg add file1 $ hg commit -m "adds file1" $ hg mv file1 file2 $ hg grep "some" `file2:2147483647:some text` Differential Revision: https://phab.mercurial-scm.org/D3826

patch: don't separate \r and \n when colorizing diff output When displaying diffs, \r at the end of a line is treated as trailing whitespace. This causes an ANSI escape code to be inserted between \r and \n. Some programs, such as less since version 530 (maybe earlier, but at least not version 487) displays ^M when it encounters a lone \r. This causes a lot of noise in diff output on Windows, where \r\n is used to terminate lines. We avoid that by treating both \n and \r\n as end of line when considering trailing whitespace.

hook: add support for disabling the shell to native command translation I think having it on by default is the right thing to do, but this is an escape hatch if someone has a command that shouldn't be mangled. The inspiration is the priority prefix. The translation does nothing on non Windows platforms, so the default value is selected to avoid printing a useless note by default.

hook: narrow the 'priority' prefix check to align with the documentation A prefix like 'priorityfoo' is meaningless, but `hg help config.hooks` calls out the dot.

windows: don't consider '$$' to be an escaped '$' when translating to cmd.exe This functionality was inherited from `os.path.expandvars()`. But the point of adding this translating code is to be able to write a portable hook, and bash wouldn't replace '$$' with '$'. Escaping with '\' works, and is portable.

contrib: add heads(commonancestors(_)) to all-revsets This is mainly to check that we don't regress our optimization path.

revset: add optimization for heads(commonancestors()) Previously, the only way to get these commits were (tested on mozilla-central): hg perfrevset 'heads(::a7cf55 and ::d8b15)' ! wall 4.988366 comb 4.960000 user 4.780000 sys 0.180000 (best of 3) After this patch: (python) hg perfrevset 'heads(commonancestors(a7cf55 + d8b15))' ! wall 0.002155 comb 0.000000 user 0.000000 sys 0.000000 (best of 1107) (C) hg perfrevset 'heads(commonancestors(a7cf55 + d8b15))' ! wall 0.000568 comb 0.000000 user 0.000000 sys 0.000000 (best of 4646)

revsets: add commonancestors revset This is a method to reproduce "::x and ::y" such that a set can be sent in. For instance, it'd be convenient to have "::heads()" work like this but that already means "::x + ::y + ..." for each element in the "heads()" set. Therefore, we add the "commonancestors" method to mean "::x and ::y ..." for each head in the given set.

tweakdefaults: enable word-diff by default It's another creature comfort in diff output, we may as well include it alongside git-diffs and showfunc.

run-tests: fix a too long line Differential Revision: https://phab.mercurial-scm.org/D3900

run-tests: add missing life-cycle methods on the example custom test result A previous commit introduced `onStart` and `onEnd` methods on test result but the one used in tests lacked those two methods. Fix it and add some output to be sure they are called. Differential Revision: https://phab.mercurial-scm.org/D3899

run-tests: fix test result verbosity A previous refactoring created the test result with a verbosity of 0 hiding some of the outputs in the normal case. Differential Revision: https://phab.mercurial-scm.org/D3898

shelve: use more accurate description in conflict marker We use "shelve" and "working-copy" instead of "source" and "dest". This is a net win. Differential Revision: https://phab.mercurial-scm.org/D3694

shelve: directly handle the initial parent alignment Shelve is currently sub-contracting some of its work to the rebase extension. In order to make shelve more independent and flexible we would like shelve to handle the parent alignment directly. After this change, we no longer need to use rebase in shelve. Differential Revision: https://phab.mercurial-scm.org/D3693

run-tests: extract onStart and onEnd into the test result It would allow custom test result to display custom messages. Differential Revision: https://phab.mercurial-scm.org/D3701

run-tests: add support for external test result The goal is to begin experiment with custom test result. I'm not sure we should offers any backward-compatibility guarantee on that plugin API as it doesn't change often and shouldn't have too much clients. Differential Revision: https://phab.mercurial-scm.org/D3700

remotenames: synchronise remotenames after push also Earlier we use to pull remotenames information from the server in case of pull and clone only. This patch adds logic to push also command to pull remotenames information. This will help us in keeping the remotenames more upto date where there are a lot people changing state of branches and bookmarks at the server. Differential Revision: https://phab.mercurial-scm.org/D2874

fileset: pass in badfn to inner matchers Just for sanity. No idea if this will make a difference, but it should propagate the badfn because the matcher created by mctx.matcher() will be returned by fileset.match() in future patches.

fileset: restrict getfileset() to not return a computed set (API) And rename the functions accordingly. fileset.match() will be changed to not compute the initial subset. test-glog*.t get back to the state before 9f9ffe5f687c "match: compose 'set:' pattern as matcher."

match: add prefixdirmatcher to adapt subrepo matcher back This serves as an inverse function to the subdirmatcher, and will be used to wrap a fileset matcher of subrepositories. One of the root/prefix paths could be deduced from the matcher attributes to be wrapped, but we don't since the callers of this class know the root/prefix paths and can simply pass them in.

fileset: make debugfileset filter repository files This prepares for the structural change of the fileset. A computed fileset will no longer be a set of files, but a boolean function (i.e. matcher) to test if an input file matches the given fileset expression. --all-files option is added because some examples in the test need to scan files across revisions.

bundle2: use ProgrammingError to report bad use of addparam() This allows us to embed error message in bytes.

py3: byte-stringify literals in extension in test-bundle2-format.t # skip-blame just some b''

py3: drop b'' while formatting BundleUnknownFeatureError message

py3: suppress write() result and close file in test-bookmarks-pushpull.t

py3: drop b'' while formatting ResponseError

py3: fix revnums in bookmark discovery to be consumable more than once

py3: byte-stringify literals in extension in test-bundle2-exchange.t # skip-blame just some b''

aggressivemergedeltas: enabled the option by default The option has been around for a while (August 2015) but was never turned on by default. In-depth testing shows large wins for having that on with no significant drawbacks. When enabled, revlog consider delta against both p1 and p2 at the same time when storing a revision. Selecting a delta against "p2" can produce better deltas and chain. This raise large benefit for all repositories, especially if they have a lot of merges. Comparison of `.hg/store/` size: mercurial (6.74% merges): before: 54,225,348 bytes after: 47,279,959 bytes -13% pypy (8.30% merges): before: 459,041,759 bytes after: 346,090,067 bytes -25% netbeans (34.21% merges): before: 2,468,041,333 bytes after: 1,364,077,645 bytes -45% mozilla-central (4.84% merges): before: 2,731,799,546 bytes after: 2,157,718,019 bytes -21% Comparison of `00manifest.d` size: mercurial (6.74% merges): before: 11,682,516 bytes after: 6,143,044 bytes -47% pypy (8.30% merges): before: 156,447,163 bytes after: 52,941,780 bytes -66% netbeans (34.21% merges): before: 1,250,363,851 bytes after: 130,088,982 bytes -90% mozilla-central (4.84% merges): before: 468,202,733 bytes after: 215,096,339 bytes -54% In addition, the better deltas help with the performance of multiple core operations. However, better chains mean longer chains, which can affect performance negatively (mostly manifest revision retrieval time). Chains length is a deeper problem that also affects linear repository too. Overall we think the benefits of using p2 as a diff target are bigger than the downsizes. In addition, we are also working on ways to improve the performance impact of chain length, so theses downsizes get fixed in the future. Below are interesting items from the full benchmark run: bundling 100 revisions from pypy: before: 670ms after: 480ms -28% bundle 10000 revisions from pypy: before: 1.38s after: 1.10s -54% bundle 10000 revisions from pypy: before: 16.1s after: 7.81s -52% bundle 10000 revisions from netbeans: before: 19.3s after: 15.5s -19% unbundle 1000 revisions to pypy: before: 641ms after: 315ms - 51% clone mercurial (http): before: 26.0s after: 22.6s -23% pulling 1000 revisions from pypy (shh): before: 2.07s after: 1.36s -44% pushing 1000 revision through http (pypy repository) before: 2.18s after: 1.35s -48% diff time in mozilla-central: before: 1.420s after: 0.983s -31% status time in mozilla-central: before: 1.260s after: 0.828s -34% Impact in other cases seems minimal (within a couple of percent in worse cases) and can be seen in both direction: Timing for a simple `hg commit`: mozilla-central: before: 3.37s after: 3.22s -4% pypy: before: 194ms after: 197ms +2% Timing for status (from tip to parent of tip): mercurial: before: 52.4ms after: 52.4ms (same) pypy: before: 55.2 after: 56.9 +3% Timing for `hg update` mozilla-central, across 10 revisions: before: 4.82s after: 4.59s -5% mozilla-central, across 10000 revisions: before: 49.1s after: 49.9s +2% pypy, across 10 revisions: before: 213ms after: 216ms +1% pypy, across 10000 revisions: before: 5.31ms after: 5.24ms -1% The negative consequences are related to manifest fetch time: (timing for the tip revision tested by the benchmark) pypy-2018: before: 2.60ms after: 3.88ms +50% mozilla-central-2018: before: 565ms after: 652ms +15% (~+100ms) netbeans-2018: before: 101ms after: 250ms +48% (~+150ms) This shows up as a fixed overhead on some command we benchmarked: no-op push of mozilla-central: before: 945ms after: 1040ms +10% (~+100ms) pushing 10 changeset in netbeabs over ssh: before: 557ms after: 712ms +28% (+155ms) pushing 100 changeset in netbeabs over ssh: before: 592ms after: 771ms +30% (+179ms)

merge with stable

ui: make the large file warning limit fully configurable While add --large can be used to override it selectively, often enough the user simply doesn't care about machines with less than 100MB RAM or so, so make it possible to just specify a larger limit in hgrc. Differential Revision: https://phab.mercurial-scm.org/D3893

fileset: sort debugfileset output Unlike revset, the order of fileset result doesn't matter since it's used as a matcher predicate. This stabilizes debugfileset output for upcoming changes.

fileset: move helper functions to top

py3: fix bundle heads to be consumable more than once

py3: byte-stringify literals in hook script in test-bundle.t # skip-blame just some b''

py3: make 'None in lazyancestors' not crash This looks somewhat weird, but we have callers like 'torev(n) in futurecommon' around where torev(n) is dictlike.get(n). I could fix callers, but that would be unnecessarily verbose.

py3: convert server-string to unicode to make http library happy

py3: fix dumbhttp.py to convert --daemon-postexec arguments back to bytes

py3: use bytes.endswith() instead of bytes[n]

diff: graduate word-diff option from experimental Per 4.6 Sprint notes. I've also made it gated by "formatchanging" since it could change the output if we had an option to highlight words without using colors.

fix: add test case that shows why --whole with --base is useful Differential Revision: https://phab.mercurial-scm.org/D3894

context: raise ProgrammingError on repo['my-tag'] We had an internal extension that I had failed to migrate off of the deprecated API and its "'my-tag' in repo" check just started returning False. It took a while to figure out that that was what was happening. This patch would have helped. Differential Revision: https://phab.mercurial-scm.org/D3895

diffutil: move the module out of utils package mercurial.utils modules inherit the property of the mercurial.util, which is no dependency on ui, repo, ctx, etc. As the diffutil module seems to reside in the scmutil layer, it's probably better to not put it under the utils package.

diffutil: remove diffopts() in favor of diffallopts() patch.diffopts() exists only for backward compatibility. We don't need it in new module.

obsutil: use public interface to access to repo.ui

hgweb: pass ui to diffstatgen() explicitly

synthrepo: simply use the ui passed as a function argument

templatekw: obtain ui directly from the template context

context: remove unneeded alias of diffopts

match: remove ctx argument from code path down to _buildmatch() 'ctx' was there only for filesets.

match: compose 'set:' pattern as matcher Baby step towards porting fileset to matcher composition. We can't use the exactmatcher since it would provide a computed set as exact paths. That's why we use the predicatematcher with fset.__contains__. This will be cleaned up later. The test change in test-glog.t means that the "set:copied()" pattern is no longer be processed as a slow path. That's because the fset is empty. This will also change in future patches.

match: resolve 'set:' patterns first in _buildmatch() This just makes the next patch less complicated. The order of 'set:' and 'subinclude:' expansion doesn't matter.

match: explode if unsupported pattern passed down to _regex() builder

match: add basic wrapper for boolean function This serves as a generic wrapper for fileset predicates. In future patches, a fileset expression will be mapped to a tree of matchers for a better support of match attributes such as visitdir(). For example, $ hg debugwalk -v 'set:contrib/** and binary()' * matcher: <intersectionmatcher m1=<patternmatcher patterns='(?:contrib/.*$)'>, m2=<predicatematcher pred=binary>> ...

stringutil: move _formatsetrepr() from smartset I'll add a matcher subclass wrapping a boolean function, which will use buildrepr() to provide debugging information in a similar way to smartset.filteredset.

pycompat: move rapply() from util I want to use rapply() in utils.* modules, but that would introduce a reference cycle util -> utils.* -> util. Moving rapply() to pycompat should be okay since it mostly serves as a compatibility helper.

strip: improve help text for --no-backup option Help text is modified to clearly define the meaning of --no-backup option. Differential Revision: https://phab.mercurial-scm.org/D3886

test-convert: demonstrate an unstable hash issue for bzr -> hg -> hg It looks like the manifest value changing is the only difference, but I'm not sure why it's happening. I've got a similar divergence in a production repo that was also converted from bzr and has an octopus merge[1]. Unlike here, the manifest values for the destination merge commits reflect the initial merge only, instead of all four merges agreeing like this test. $ hg -R src_repo manifest -r 310 --debug | grep file # octopus fixup merge 2d8775bc2481bd28ac87038ecdf33e1dbddc80e9 644 file1 6adb9353a55bb8be76e71382efc724ec3ccf7ed5 644 file2 $ hg -R src_repo manifest -r 309 --debug | grep file # first merge 362e7cb5163153c4989daad1a834871ae849f205 644 file1 2c65d947191938c3ea616b7ceb7648ff3843261f 644 file2 $ hg -R dst_repo manifest -r 273 --debug | grep file # octopus fixup merge 362e7cb5163153c4989daad1a834871ae849f205 644 file1 2c65d947191938c3ea616b7ceb7648ff3843261f 644 file2 $ hg -R dst_repo manifest -r 272 --debug | grep file # first merge 362e7cb5163153c4989daad1a834871ae849f205 644 file1 2c65d947191938c3ea616b7ceb7648ff3843261f 644 file2 This divergence is espcially annoying because unlike changelog differences, I haven't figured out a way to fix this in code. The only way I found to work around it is to convert up to the point of divergence, `hg bundle` the bad revision in the source, apply it to the destination, add a line to the shamap, and fire off the conversion again. But I suspect that there's more to it than just the octopus merge because I also have a commit in the same repo, done in Mercurial (well after the conversion) that is exhibiting a similar issue (and it's not a merge commit). I'm almost positive that it was created with 4.4 or later. Any ideas? [1] https://www.mercurial-scm.org/pipermail/mercurial/2018-June/050924.html

convert: add a config knob for not saving the bzr revision Now that the timestamp is fixed, the log of the Mercurial repo is still unstable with --debug, because the bzr commit ID keeps changing. Both hg and git implement `convert.*.saverev`, so do the same here.

test-bzr: use a fixed commit time This will help stabilize the hashes when converting to a Mercurial repo.

diff: use `context.diff` to produce diff We want to make sure `context.diff` if full featured. Using it in the main diff code path is a good way to do so. The end goal is to be able to easily compute diff between in-memory context.

context: no longer accept diff options as dictionnary Since we already broke the API earlier in this stack, there are no point to introducing a new deprecation warning.

obsutil: pass a diffopts object to context.diff

synthrepo: pass a diffopts object to context.diff

webutil: pass a diffopts object to context.diff

revset: pass an explicit `diffopts` objet to context.diff

template: directly instantiate diff options for diffstat

tests: update test-context.py to use diffopts as diff argument

diffutil: extract diff options code into a dedicated util-module We want to be able to create and use diffoptions in more places. Currently the official function to instantiate diffoptions live into `mercurial.patch`. A module too high level to be easily imported in some places. So we extract the diff options related function in their own utility module.

packaging: add fedora 28 Differential Revision: https://phab.mercurial-scm.org/D3892

packaging: add missing path segments in cp statements in fedora build Differential Revision: https://phab.mercurial-scm.org/D3891

packaging: fix misplaced ( in docker-fedora rule definition Differential Revision: https://phab.mercurial-scm.org/D3890

packaging: make help output correct for fedora and centos targets Differential Revision: https://phab.mercurial-scm.org/D3889

packaging: remove stray = that confuses make(1) on my Mac Differential Revision: https://phab.mercurial-scm.org/D3888

manifest: define and implement addgroup() on manifestlog Changegroup code was bypassing our manifest interface and calling a method on the private revlog attribute. This commit formalizes the interface for adding a group of revisions from deltas and changes the changegroup code to use it. This enables alternate manifest storage to work with changegroup application operations (like `hg unbundle` and `hg pull`). Differential Revision: https://phab.mercurial-scm.org/D3883

manifest: define and implement rev() on manifestlog Various code is accessing repo.manifestlog._revlog - a private attribute. This bypasses our interface and makes it difficult to implement non-revlog manifest storage. This commit adds a rev() method to the manifestlog interface and class and teaches callers to use it. Ideally this method wouldn't exist, as very few consumers should need to resolve the revision number of a manifest. Again, the primary goal of interface work is to establish and use interfaces first and to improve them later. Differential Revision: https://phab.mercurial-scm.org/D3882

tests: don't allow reodering of glob/re lines across non-glob/re lines As shown in the test case added in the previous patch, it can be really hard to interpret diffs from the test runner if there are multiple lines that would match a given glob or regular expression. It looks like this has been the case since 1ad0ddf8cccc (run-tests: teach _processoutput to handle multiple lines of churn, 2016-03-17). It seems like the point of that was to preserve the "(glob)" annotation on lines even if they got moved. This patch tries to preserve that but only allows the lines to be moved past other glob/re lines. Differential Revision: https://phab.mercurial-scm.org/D3881

tests: add test showing puzzling test output with (glob) lines When using multiple (glob) lines, the test runner often moves the lines around in the output on failure. I have run into this problem many times. The added test shows one example of this. Note that it doesn't show the mismatching lines and instead shows diff hunks that appear to replace a literal by a glob, and other hunks that should clearly match. Differential Revision: https://phab.mercurial-scm.org/D3880

tests: move handling of None "el" out of linematch() It just seems odd for linematch() to know what None means. Differential Revision: https://phab.mercurial-scm.org/D3879

tests: remove some redundant code in run-tests These two lines also happen below if "r" is truthy, which it is in this case since it's equal to "retry". Differential Revision: https://phab.mercurial-scm.org/D3878

tests: don't reimplement enumerate() in run-tests Differential Revision: https://phab.mercurial-scm.org/D3877

histedit: improve help text for --no-backup option Improved help text for --no-backup option to clearly define it's meaning. Differential Revision: https://phab.mercurial-scm.org/D3885

histedit: add --no-backup option (issue5825) This option provides a functionality to not store a backup while aborting histedit in between. Also added tests for the same. Differential Revision: https://phab.mercurial-scm.org/D3872