hgweb: strip trailing '/' in apppath before appending '/static/' (issue5943) Differential Revision: https://phab.mercurial-scm.org/D3978

config: rename `revlog` section into `storage` The idea was suggested by Gregory Szorc on IRC. It is more generic and seems better. It is probably best to rename the section before it ever makes into an official (non-rc) release. The only config option currently in this section have been prefixed with `revlog` to clarify it applies to `revlog` related storage.

Added signature for changeset e90130af47ce

Added tag 4.7rc0 for changeset e90130af47ce

configitems: restore alias for format.aggressivemergedeltas 913ca175c4ae broke BC by dropping support for reading format.aggressivemergedeltas. Let's restore it. Differential Revision: https://phab.mercurial-scm.org/D3966

worker: call selector.close() to release polling resources

release: merge default into stable for 4.7 release freeze

histedit: drop --no-backup option Dropping this option because now we have a better option than passing --no-backup flag every time, now user can set a config in hgrc: [ui] history-editing-backup = False This config aims to operate on every history editing command and it is still work in progress. As yuya suggessted it probably to late to add full support this config, so making this as an experimental config. Differential Revision: https://phab.mercurial-scm.org/D3965

aggressivemergedelta: document rename and move to `revlog` section The config does not follow our naming guideline and "Aggressive" is probably a word to keep away from users. The option does not truly fit in the `format` section. It can be turned on and off for existing repository without much consequence regarding compatibility. A new `revlog` option is created to control behavior related to revlog writing and reading. We can see multiple other config options that could be migrated there. * format.maxchainlen * experimental.mmapindexthreshold * experimental.sparse-read.density-threshold (in an updated form) * experimental.sparse-read.min-gap-size (in an updated form) In addition, we can foresee at least a couple of sparse-revlog related option coming too (to reduce delta chain length and increase snapshot reuse) These two extra options might fit there too. Unless we want to create a section dedicated to caches and performance. * format.chunkcachesize * format.manifestcachesize For now, we only migrate `optimize-delta-parent-choice` since it is getting out of experimental. It is too close to the release to move the other one. In addition, we still lack proper the prioritization of alias that would help renaming them without bad consequence for users. (Not fully happy about the `revlog` name but could not find better).

aggressivemergedeltas: rename variable internally The "aggressivemergedeltas" name is not great. First, it is quite long, second, we would rather have less "Aggressive" names within the project. We are about to rename the config option, so it seems the appropriate time to rename the internal variable.

config: document the purpose of the `format` config section The config section now have a top level documentation to clarify its intend and usage. In particular, user are now explain when the option are taken in account and how to convert repository. There are an handful of experimental options in this section that does not match its definition. They should be relocated to other section before getting out of experimental. (see next changeset for one example).

macosx: fixing macOS version generation after db9d1dd01bf0 With the Python3 change, the string is now something like version = b'4.6.2+848-88be288e8ac1' where it was previously just: version = '4.6.2+848-88be288e8ac1' Differential Revision: https://phab.mercurial-scm.org/D3964

histedit: add history-editing-backup config option Instead of passing --no-backup option every time you don't want to store backup, now you can set config option: [ui] history-editing-backup = False This option aims to operate on every history editing command. Differential Revision: https://phab.mercurial-scm.org/D3901

merge: mark file gets as not thread safe (issue5933) In default installs, this has the effect of disabling the thread-based worker on Windows when manifesting files in the working directory. My measurements have shown that with revlog-based repositories, Mercurial spends a lot of CPU time in revlog code resolving file data. This ends up incurring a lot of context switching across threads and slows down `hg update` operations when going from an empty working directory to the tip of the repo. On mozilla-unified (246,351 files) on an i7-6700K (4+4 CPUs): before: 487s wall after: 360s wall (equivalent to worker.enabled=false) cpus=2: 379s wall Even with only 2 threads, the thread pool is still slower. The introduction of the thread-based worker (02b36e860e0b) states that it resulted in a "~50%" speedup for `hg sparse --enable-profile` and `hg sparse --disable-profile`. This disagrees with my measurement above. I theorize a few reasons for this: 1) Removal of files from the working directory is I/O - not CPU - bound and should benefit from a thread pool (unless I/O is insanely fast and the GIL release is near instantaneous). So tests like `hg sparse --enable-profile` may exercise deletion throughput and aren't good benchmarks for worker tasks that are CPU heavy. 2) The patch was authored by someone at Facebook. The results were likely measured against a repository using remotefilelog. And I believe that revision retrieval during working directory updates with remotefilelog will often use a remote store, thus being I/O and not CPU bound. This probably resulted in an overstated performance gain. Since there appears to be a need to enable the thread-based worker with some stores, I've made the flagging of file gets as thread safe configurable. I've made it experimental because I don't want to formalize a boolean flag for this option and because this attribute is best captured against the store implementation. But we don't have a proper store API for this yet. I'd rather cross this bridge later. It is possible there are revlog-based repositories that do benefit from a thread-based worker. I didn't do very comprehensive testing. If there are, we may want to devise a more proper algorithm for whether to use the thread-based worker, including possibly config options to limit the number of threads to use. But until I see evidence that justifies complexity, simplicity wins. Differential Revision: https://phab.mercurial-scm.org/D3963

worker: ability to disable thread unsafe tasks The worker on Windows is implemented using a thread pool. If worker tasks are not thread safe, badness can occur. In addition, if tasks are executing CPU bound code and holding onto the GIL, there will be non-substantial overhead in Python context switching between active threads. This can result in significant slowdowns of tasks. This commit teaches the code for determining whether to use a worker to take thread safety into account. Effectively, thread unsafe tasks don't use the thread-based worker on Windows. Differential Revision: https://phab.mercurial-scm.org/D3962

worker: rename variable to reflect constant Differential Revision: https://phab.mercurial-scm.org/D3961

worker: use one pipe per posix worker and select() in parent process This allows us to pass results larger than PIPE_BUF through the pipes without interleaving them. This is necessary now because "hg fix" sends file contents as the result from workers. Differential Revision: https://phab.mercurial-scm.org/D3960

tests: refactor common bundle2 capabilities Differential Revision: https://phab.mercurial-scm.org/D1945

debug: move extensions debug behind a dedicated flag Since b86664c81833, we process the `--debug` flag earlier. This is overall good and useful, but has at least one negative side effect. Previously the debug message we report when trying to import extensions were issued before we processed the `--debug` flag. Now they happen after. Before: $ ./hg id --debug 21f507b8de2f9c1606e9aeb5ec7d2a6dedb7a4a7 tip After: $ ./hg id --debug ☿ (revset-bench) could not import hgext.evolve (No module named evolve): trying hgext3rd.evolve could not import hgext.mercurial_keyring (No module named mercurial_keyring): trying hgext3rd.mercurial_keyring could not import hgext3rd.mercurial_keyring (No module named mercurial_keyring): trying mercurial_keyring could not import hgext.hggit (No module named hggit): trying hgext3rd.hggit could not import hgext3rd.hggit (No module named hggit): trying hggit 21f507b8de2f9c1606e9aeb5ec7d2a6dedb7a4a7 tip (This get worse if --traceback is used). To work around this, we move this extensions related debug message behind a new flag 'devel.debug.extensions' and restore the previous output. I'm not fully happy about using the 'devel' section for a flag that can be used by legitimate users to debug extensions issues. However, it fits well next to other `devel.devel.*` options and is mostly used by extensions author anyway. We might move it to another, more appropriate section in the future (using alias).

curses: use "text" interface if TERM=dumb Differential Revision: https://phab.mercurial-scm.org/D3948

windows: expand '~/' and '~\' to %USERPROFILE% when translating to cmd.exe It's convenient to be able to reference hooks in a portable location on any platform.

windows: replace single quote with double quote when translating to cmd.exe Since cmd.exe doesn't understand single quotes, single quotes to prevent $var expansion is basically unusable without this. Single quote isn't allowed in a path name, so it seems unlikely to come up otherwise.

hook: only print the note about native cmd translation if it actually changes This makes it so that it will never occur on a non Windows platform.

hook: disable the shell to native command translation by default There are other things I want to add like ~ expansion and translating single to double quotes for cmd.exe. So off by default is safer. I'm having second thoughts about the name, but I don't have any better ideas.

setup: allow to run setup.py with python 3 without a mercurial checkout Some people may want to test mercurial in a python 3 environment through e.g. pip, in which case setup.py doesn't run in a mercurial checkout, so the hack in setup.py to allow python 3 cannot be overcome. This change allows a manual override with the HGPYTHON3 environment variable. Additionally, when for some reason the version is unknown (for crazy people like me, who have a git checkout of the mercurial repo), the version variable ends up being an unicode string, which fails the `isinstance(version, bytes)` assertion. So fix that at the same time. Differential Revision: https://phab.mercurial-scm.org/D3958

upgrade: enable adding or removing sparse-revlog requirement

upgrade: add information about sparse-revlog Show information about sparse-revlog in debugformat, just like other requirements.

sparse-revlog: implement algorithm to write sparse delta chains (issue5480) The classic behavior of revlog._isgooddeltainfo is to consider the span size of the whole delta chain, and limit it to 4 * textlen. Once sparse-revlog writing is allowed (and enforced with a requirement), revlog._isgooddeltainfo considers the span of the largest chunk as the distance used in the verification, instead of using the span of the whole delta chain. In order to compute the span of the largest chunk, we need to slice into chunks a chain with the new revision at the top of the revlog, and take the maximal span of these chunks. The sparse read density is a parameter to the slicing, as it will stop when the global read density reaches this threshold. For instance, a density of 50% means that 2 of 4 read bytes are actually used for the reconstruction of the revision (the others are part of other chains). This allows a new revision to be potentially stored with a diff against another revision anywhere in the history, instead of forcing it in the last 4 * textlen. The result is a much better compression on repositories that have many concurrent branches. Here are a comparison between using deltas from current upstream (aggressive-merge-deltas on by default) and deltas from a sparse-revlog Comparison of `.hg/store/` size: mercurial (6.74% merges): before: 46,831,873 bytes after: 46,795,992 bytes (no relevant change) pypy (8.30% merges): before: 333,524,651 bytes after: 308,417,511 bytes -8% netbeans (34.21% merges): before: 1,141,847,554 bytes after: 1,131,093,161 bytes -1% mozilla-central (4.84% merges): before: 2,344,248,850 bytes after: 2,328,459,258 bytes -1% large-private-repo-A (merge 19.73%) before: 41,510,550,163 bytes after: 8,121,763,428 bytes -80% large-private-repo-B (23.77%) before: 58,702,221,709 bytes after: 8,351,588,828 bytes -76% Comparison of `00manifest.d` size: mercurial (6.74% merges): before: 6,143,044 bytes after: 6,107,163 bytes pypy (8.30% merges): before: 52,941,780 bytes after: 27,834,082 bytes -48% netbeans (34.21% merges): before: 130,088,982 bytes after: 119,337,636 bytes -10% mozilla-central (4.84% merges): before: 215,096,339 bytes after: 199,496,863 bytes -8% large-private-repo-A (merge 19.73%) before: 33,725,285,081 bytes after: 390,302,545 bytes -99% large-private-repo-B (23.77%) before: 49,457,701,645 bytes after: 1,366,752,187 bytes -97% The better delta chains provide a performance boost in relevant repositories: pypy, bundling 1000 revisions: before: 1.670s after: 1.149s -31% Unbundling got a bit slower. probably because the sparse algorithm is still pure python. pypy, unbundling 1000 revisions: before: 4.062s after: 4.507s +10% Performance of bundle/unbundle in repository with few concurrent branches (eg: mercurial) are unaffected. No significant differences have been noticed then timing `hg push` and `hg pull` locally. More state timings are being gathered. Same as for aggressive-merge-delta, better delta comes with longer delta chains. Longer chains have a performance impact. For example. The length of the chain needed to get the manifest of pypy's tip moves from 82 item to 1929 items. This moves the restore time from 3.88ms to 11.3ms. Delta chain length is an independent issue that affects repository without this changes. It will be dealt with independently. No significant differences have been observed on repositories where `sparse-revlog` have not much effect (mercurial, unity, netbeans). On pypy, small differences have been observed on some operation affected by delta chain building and retrieval. pypy, perfmanifest before: 0.006162s after: 0.017899s +190% pypy, commit: before: 0.382 after: 0.376 -1% pypy, status: before: 0.157 after: 0.168 +7% More comprehensive and stable timing comparisons are in progress.

sparse-revlog: new requirement enabled with format.sparse-revlog The meaning of the new 'sparse-revlog' requirement is that the revlogs are allowed to contain wider delta chains with larger holes between the interesting chunks. These sparse delta chains should be read in several chunks to avoid a potential explosion of memory usage. Former version won't know how to read a delta chain in several chunks. They would keep reading them in a single read, and therefore would be subject to the potential memory explosion. Hence this new requirement: only versions having support of sparse-revlog reading should be allowed to read such a revlog. Implementation of this new algorithm and tools to enable or disable the requirement will follow in the next changesets.

revlog: extract `deltainfo.distance` for future conditional redefinition This commit exist to make the next one clearer.

shelve: pick the most recent shelve if none specified for --patch/--stat Differential Revision: https://phab.mercurial-scm.org/D3950

shelve: improve help text for --patch and --stat It's not currently obvious why "hg shelve -p" fails, since -p doesn't take an argument. Differential Revision: https://phab.mercurial-scm.org/D3949

ssh: avoid reading beyond the end of stream when using compression Compressed streams can be used as part of getbundle. The normal read() operation of bufferedinputpipe will try to fulfill the request exactly and can deadlock if the server sends less as it is done. At the same time, the bundle2 logic will stop reading when it believes it has gotten all parts of the bundle, which can leave behind end of stream markers as used by bzip2 and zstd. To solve this, introduce a new optional unbufferedread interface and provided it in bufferedinputpipe and doublepipe. If there is buffered data left, it will be returned, otherwise it will issue a single read request and return whatever it obtains. Reorganize the decompression handlers to try harder to read until the end of stream, especially if the requested read can already be fulfilled. Check for end of stream is messy with Python 2, none of the standard compression modules properly exposes it. At least with zstd and bzip2, decompressing will remember EOS and fail for empty input after the EOS has been seen. For zlib, the only way to detect it with Python 2 is to duplicate the decompressobj and force some additional data into it. The common handler can be further optimized, but works as PoC. Differential Revision: https://phab.mercurial-scm.org/D3937

revset: add larger test for heads(ancestors(…)) It is important to not regress on this benchmark so we move it into the "base" file. And we add another benchmark with more than two revisions.

revset-benchmark: use a generic revset to test `heads(commonancestors())` This allow to benchmark revset performance in other repositories than just the mercurial one.

revlog: reintroduce `revlog.descendant` as deprecated Reintroduce `revlog.descendant` to help extensions authors update their extensions in order to use the new API.

context: reintroduce `ctx.descendant` as deprecated Reintroduce `ctx.descendant` to help extensions authors update their extensions in order to use the new API.

obsolete: explode if metadata contains invalid UTF-8 sequence (API) The current metadata API can be a source of bugs since it forces callers to process encoding conversion by themselves. So let's make it reject bad data as a last ditch. I assume there's no metadata field which is supposed to store arbitrary BLOB like transplant_source.

obsolete: store user name and note in UTF-8 (issue5754) (BC) Before, user names were stored in local encoding and transferred across repositories, which made it impossible to restore non-ASCII user names on different platforms. This patch fixes new markers to be encoded in UTF-8 and decoded back to local encoding when displaying. Existing markers are unfixable so they may result in mojibake. I don't like the API that requires metadata dict to be UTF-8 encoded, which is a source of bugs, but there's no abstraction layer to process the encoding thingy efficiently. So we apply the same rule as extras dict to obsstore metadata.

obsolete: clarify users in markerusers() never contain None

revset: special case commonancestors(none()) to be empty set This matches the behavior of ancestor(none()). From an implementation perspective, ancestor() and commonancestors() are intersection, and ancestors() is union, so it would make some sense that commonancestors(none()) returned all revisions. However, ancestor(none()) isn't implemented as such, which breaks ancestor(x) == max(commonancestors(x)). From a user perspective, ancestors of nothing is nothing whichever type of operation the ancestor predicate does.

revset: clarify heads() order doesn't matter while computing common ancestors Follows up 5460926352ee and 52f19a840543.

hghave: require clang-format >= 6 due to output change

cext: reformat with clang-format 6.0 It appears some changes in clang-format affect our code. I didn't dig into that deeper since the new output looks better.

py3: use bytes() to byte-stringify Abort message in handleremotechangegroup()

py3: don't str() to byte-stringify object in test-bundle2-remote-changegroup.t

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

py3: open file in binary mode in test-bundle2-remote-changegroup.t

statprof: small if cleanup Explicitly testing for None to avoid comparison bugs.

store: assert the fncache have been loaded if dirty This should catch fncache corruption as the one that existed in `perffncachewrite`.

perffncachewrite: load fncache after lock is acquired Without this patch, running perffncachewrite on a repository destroy its fncache. Lock Acquisition drops various caches, including the fncache one. Then writing of an non-loaded fncache result into an empty one.

perf: add a 'perf.all-timing' option to display more than best time Minimal time is a useful information, but it is useful to have a wider view on the performance picture.

perf: document the perfmanifest command

fileset: remove fullmatchctx class It's exactly the same as matchctx.

fileset: remove subset and unused filtering functions from matchctx

fileset: remove callexisting flag and mctx.existing() (API) They are no longer needed since any files are included as long as they are passed in to the matcher.

fileset: rewrite predicates to return matcher not closed to subset (API) (BC) This makes fileset expression open to any input, so that we can just say "hg status 'set: not binary()'" to select text files including unknowns. With this and removal of subset computation, 'set:**' becomes as fast as 'glob:**'. Further optimization will probably be possible by narrowing the file tree to compute status for example. This also fixes 'subrepo()' to not ignore the current mctx.subset. .. bc:: The fileset expression may include untracked files by default. Use ``tracked()`` to explicitly filter out files not existing at the context revision.

highlight: use matcher API to test if file should be pygmentized Prepares for the upcoming API change. We'll no longer have to call the fileset function directly since the cost of the matcher-based fileset will become O(number of tests) from O(number of files in subset).

fileset: parse argument of size() by predicate function This change is necessary to pass in a size expression to predicatematcher. See the next patch.

fileset: add "tracked()" to explicitly select files in the revision I'm going to rewrite filesets to be match predicates, which means basic patterns such as '*' will no longer be "closed" to the subset constructed from the ctx. Good thing is that 'hg status "set:not binary()"' can include unknown files out of the box, and fileset computation will likely to be faster as we won't have to walk dirstate twice, for example. Bad thing is that we can't select files at a certain revision by 'set:revs(REV, **)' since '**' is "open" to any paths. So, this patch introduces "tracked()" as a replacement for the '**' in the example above.

fileset: rewrite andset() to not use mctx.narrow() New code is less efficient than the original, but it helps porting andset() to matcher composition. This will be cleaned up later. This effectively disables the fullmatchctx magic since mctx will never be demoted to the matchctx. The fullmatchctx class will be removed later.

fileset: add helpers to make predicatematcher and nevermatcher These functions will be used to compose a tree of matchers from a fileset expression.

phases: micro-optimize newheads() to not create context objects

phases: remove excessive optimization from newheads() (issue5939) This function is intended to compute 'heads(::heads - roots::)', but it failed because 'heads + parents(roots)' missed sibling branches of the roots. That's why the public heads slipped down from D to B in the example added by 2a227782e754 "tests: add test demonstrating phase loss when cloning": > E draft > |\Z draft > | Y draft > D | public > | X draft > C/ public > B public > A public where heads = {E, Z}, roots = {X}

rebase: remove unused variable "release" and an extra blank line Differential Revision: https://phab.mercurial-scm.org/D3947

rebase: remove unnecessary confirm block Removed unnecessary 'if confirm:' block as that thing is now handled at 'finally'. Differential Revision: https://phab.mercurial-scm.org/D3945

rebase: in --confirm option just abort if hit a conflict Before this patch, it was prompting the user in both cases 1) when there is no conflict 2) when there is at least one conflict. But for simplicity we can just abort if we hit a conflict and no need to prompt in that case. Differential Revision: https://phab.mercurial-scm.org/D3944

pullbundle: fix handling of gzip bundlespecs Differential Revision: https://phab.mercurial-scm.org/D3933

tests: add test demonstrating phase loss when cloning (issue5939) The added tests demonstrate that phases exchange when using the listkeys based phases exchange fails to preserve public phase in a certain scenario when a merge is a phase root. Both non-bundle2 and bundle2 prior to the binary phase data part are buggy. Differential Revision: https://phab.mercurial-scm.org/D3932

scmutil: rewrite docstring for filecache The old docstring was incorrect in that it said that subsequent calls perform a stat() and refresh the object if things change. This is not how things work: __get__ populates obj.__dict__[self.sname] with the result of the decorated function and returns this value without validation on subsequent calls, if available. The correct usage of this type is kinda wonky. It would probably benefit from a refactor. But I don't have time to do that right now. But we can change the docstring so others aren't entrapped by its lies (like I was when using repofilecache in a Mozilla extension). Differential Revision: https://phab.mercurial-scm.org/D3943

rebase: correct misleading message in --confirm option Differential Revision: https://phab.mercurial-scm.org/D3939

rebase: make sure we don't loose the return code in --confirm option return _dorebase() to make sure it returns 'return code'. Differential Revision: https://phab.mercurial-scm.org/D3938

rebase: use revnums (not nodes) for set of extinct revisions There is no need to convert extinct revisions to nodeids. Differential Revision: https://phab.mercurial-scm.org/D3942

rebase: avoid converting from nodes to revnums twice In the case where the node has successors, but none of them is an ancestor of the destination, we would iterate over the successor nodes twice, check if they're in the repo and convert them to revnums. I doubt it's a measureable cost, but it gets simpler this way too. Differential Revision: https://phab.mercurial-scm.org/D3941

rebase: reduce scope of a variable Differential Revision: https://phab.mercurial-scm.org/D3940

context: rename descendant() to isancestorof() This makes the direction much clearer, IMO. Differential Revision: https://phab.mercurial-scm.org/D3936

copies: delete now-unnecessary check for "a == b" before "a.descendant(b)" Unnecessary since 879cbdde63df (revlog: do inclusive descendant testing (API), 2018-06-21). Differential Revision: https://phab.mercurial-scm.org/D3935

revlog: delete isdescendantrev() in favor of isancestorrev() As agreed on by Boris, Yuya, and me on D3929. Differential Revision: https://phab.mercurial-scm.org/D3934

rebase: add --confirm option This feature adds a functionality in rebase to confirm before applying changes. When there is no conflict and user confirm to apply actions, we just finish the unfinished rebase. But when there is a conflict and user confirm to apply actions then we can't just finish rebasing using rbsrt._finishrebase() because in-memory merge doesn't support conflicts, so we have to abort and run on-disk merge in this case. And if user doesn't confirm to apply actions then simply abort the rebase. Differential Revision: https://phab.mercurial-scm.org/D3870

revlog: introduce a isancestorrev() and use it in rebase Differential Revision: https://phab.mercurial-scm.org/D3931

revlog: make isdescendantrev(a, b) check if a < b This check was taken from rebase.py. It seems to make sense to move it here, but I haven't done any measurements. Differential Revision: https://phab.mercurial-scm.org/D3930

revlog: replace descendant(b, a) by isdescendantrev(a, b) (API) The "is" is to match "isancestor" and to make it clear that it doesn't return a descendant. The "rev" is to make it clear that it's not about nodeids (unlike e.g. isancestor()). The argument order change is just seems more natural (and makes isancestor() less confusing). Differential Revision: https://phab.mercurial-scm.org/D3929

revlog: move comment about commonancestorsheads where it's used Differential Revision: https://phab.mercurial-scm.org/D3928

removeemptydirs: add test for `hg split` inside a disappearing directory Differential Revision: https://phab.mercurial-scm.org/D3927

fncache: avoid loading the filename cache when not actually modifying it With time, fncache can become very large. The mozilla-central repo for example, has a 31M and growing fncache file. Loading this file takes time (280ms for the mozilla-central repository). In many scenarios, we don't need to load fncache at all. For example, when committing changes to existing files, or pushing such commits to another clone. This patch detects when a name is added via store.vfs(), and only loads the cache if a) the data metadata file doesn't already exist, or b) when opening for appending, the data or metadata file exists but has size (a transaction rollback leaves behind such files). Benchmarks (run on Macos 10.13 on a 2017-model Macbook Pro with Core i7 2.9GHz and flash drive), each test without and with patch run 5 times: * committing to an existing file, against the mozilla-central repository. Baseline real time average 2.3736, with patch 1.9884. * unbundling a large changeset consisting *only* of existing-file modifications (159 revisions, 1050 modifications, mozilla-central 4a250a0e4f29:beea9ac7d823), into a clone limited to the ancestor revision of that revset). Baseline real time average 1.5048, with patch 1.3108.

tests: add diff color trailing whitespace test

scmutil: fix shortesthexnodeidprefix on Python 3 for 0-prefixed nodes This fixes test-bookmarks.t on Python 3 (which had regressed.) Differential Revision: https://phab.mercurial-scm.org/D3926

tests: add missing b prefix in test-context.py # skip-blame just a b prefix Differential Revision: https://phab.mercurial-scm.org/D3925

manifest: just duplicate the definition of items as iteritems The forwarding trick was failing test-check-interfaces on Python 3. Duplicating a line of code is easy enough I'm doing that rather than try and figure out what's going on in any kind of detail. Differential Revision: https://phab.mercurial-scm.org/D3924

tests: properly sort imports in test-nointerrupt.t Somehow this was only detected in Python 3. Differential Revision: https://phab.mercurial-scm.org/D3923

revlog: adjust doctest examples to be portable to Python 3 The range() builtin returns a special object rather than a list, but we really wanted a list in these examples. Differential Revision: https://phab.mercurial-scm.org/D3922

py3: whitelist 3 more passing tests caught with the ratchet Differential Revision: https://phab.mercurial-scm.org/D3921

grep: hide cryptic revision number by default I believe nobody would want to see the "2147483647:" prefix while grepping working directory files. The wdir revision is still visible if "-r wdir()" is specified since a revset may point to more than one revisions, and it seems confusing to change the visibility dynamically by the number of the matched revisions. Differential Revision: https://phab.mercurial-scm.org/D3920

grep: restore pre-9ef10437bb88 behavior, enable wdir search by tweakdefaults Unfortunately, python-hglib relies on the original grep behavior and is documented as such. Even though we agreed to introduce the BC, we shouldn't break existing libraries. So this patch flips the default again and move the new default to ui.tweakdefaults. We could instead use HGPLAIN to turn this flag off, but that would be rather confusing as the old/new behaviors are quite different. Differential Revision: https://phab.mercurial-scm.org/D3919

grep: add config knob to enable/disable the default wdir search This allows us to conditionally enable the new behavior to unbreak python-hglib. See the next patch. The config option is undocumented since the --all-files is still experimental and isn't fully implemented. It can be moved to [experimental] if that's preferred. Differential Revision: https://phab.mercurial-scm.org/D3918

grep: search working directory files by default if --all-files is specified The default was -rtip:0 before, but "--all-files -rtip:0" would be the most useless combination, and it isn't supported yet. Let's change the default to something useful. This isn't flagged as BC since --all-files isn't released. Differential Revision: https://phab.mercurial-scm.org/D3917

grep: reject --diff --all-files This combination doesn't make any sense since --diff is the flag to search change history, whereas --all-files is to include unmodified contents. Differential Revision: https://phab.mercurial-scm.org/D3916

grep: rename --allfiles to --all-files This matches the config name to be added. Note that we don't have to keep the old flag since it isn't released yet. Differential Revision: https://phab.mercurial-scm.org/D3915

debugdeltachain: protect against 0 readsize If you don't read any data, read density should be 1.

debugdeltachain: avoid division by zero when a chain is empty The two ratios chainratio and extraratio are computed using dividers that may be zero when the file is empty. As the denominators are integers, the limit of the ratio "just before zero" is the numerator value itself. If the numerator itself is zero, the ratio value is still meaningful: in both cases, a "good" value is a low ratio, and a size of zero is the optimal case.

tests: prevent conflict markers in test case from triggering warnings Broadly, having lines that begin with several '<' tends to trigger tools that try to warn you about unresolved conflicts. It will save some headaches to tweak this string literal to avoid that. Differential Revision: https://phab.mercurial-scm.org/D3897

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.