manifest: use itertools.chain() instead of + for Python 3 compat This is all pure-Python code, so I'm not too worried about perf here, but we can come back and fix it should it be a problem. With this change, the manifest code passes most unit tests on Python 3 (once the tests are corrected with many b prefixes. I've got a little more to sort out there and then I'll mail that change too.

manifest: fix some pure-Python parser bits to work on Python 3

tests: make test-manifest finish importing in Python 3 The test is still broken, but now it executes.

tests: drop assertIn polyfill now that we're 2.7-only

perf: benchmark command for revlog indexes We didn't have explicit microbenchmark coverage for loading revlog indexes. That seems like a useful thing to have, so let's add it. We currently measure the low-level nodemap APIs. There is room to hook in at the actual revlog layer. This could be done as a follow-up. The hackiest thing about this patch is specifying revlog paths. Other commands have arguments that allow resolution of changelog, manifest, and filelog. I needed to hook in at a lower level of the revlog API than what the existing helper functions to resolve revlogs allowed. I was too lazy to write some new APIs. This could be done as a follow-up easily enough. Example output for `hg perfrevlogindex 00changelog.i` on my Firefox repo (404418 revisions): ! revlog constructor ! wall 0.003106 comb 0.000000 user 0.000000 sys 0.000000 (best of 912) ! read ! wall 0.003077 comb 0.000000 user 0.000000 sys 0.000000 (best of 924) ! create index object ! wall 0.000000 comb 0.000000 user 0.000000 sys 0.000000 (best of 1803994) ! retrieve index entry for rev 0 ! wall 0.000193 comb 0.000000 user 0.000000 sys 0.000000 (best of 14037) ! look up missing node ! wall 0.003313 comb 0.000000 user 0.000000 sys 0.000000 (best of 865) ! look up node at rev 0 ! wall 0.003295 comb 0.010000 user 0.010000 sys 0.000000 (best of 858) ! look up node at 1/4 len ! wall 0.002598 comb 0.010000 user 0.010000 sys 0.000000 (best of 1103) ! look up node at 1/2 len ! wall 0.001909 comb 0.000000 user 0.000000 sys 0.000000 (best of 1507) ! look up node at 3/4 len ! wall 0.001213 comb 0.000000 user 0.000000 sys 0.000000 (best of 2275) ! look up node at tip ! wall 0.000453 comb 0.000000 user 0.000000 sys 0.000000 (best of 5697) ! look up all nodes (forward) ! wall 0.094615 comb 0.100000 user 0.100000 sys 0.000000 (best of 100) ! look up all nodes (reverse) ! wall 0.045889 comb 0.050000 user 0.050000 sys 0.000000 (best of 100) ! retrieve all index entries (forward) ! wall 0.078398 comb 0.080000 user 0.060000 sys 0.020000 (best of 100) ! retrieve all index entries (reverse) ! wall 0.079376 comb 0.080000 user 0.070000 sys 0.010000 (best of 100)

perf: rename perfrevlog to perfrevlogrevisions We have a couple of commands beginning with "perfrevlog." The actual "perfrevlog" command actually measures resolving the fulltext of multiple revisions. So let's rename it to reflect what it actually does.

server: use pycompat to get argv

encoding: make wide character class list a sysstr That's what east_asian_width returns, so just match it.

cmdutil: use sorted(dict) instead of x = dict.keys(); x.sort() The former both does less work and has the virtue of working on Python 3.

util: look for empty-sysstr instead of empty-bytesstr in textwrap code Fixes behavior on Python 3.

minirst: look for column delimiters using slices instead of indicies This works on both Python 2 and 3.

minirst: grab a byte, not an int, for the underline style

minirst: use bytes.strip instead of str.strip

py3: use pycompat.bytestr so that we don't get ascii values This fixes `hg files 'set:(**.py)'` which makes test-check-py3-compat.t able to run on Python 3. So if you now do `python3 ./run-tests.py test-check-py3-compat`, the test will actually run.

py3: update test-check-py3-compat.t

loader: pywatchman appears to already be py3 compatible Our loader was doing some confusing things in pywatchman, but it looks like we shouldn't be using it there anyway.

py3: update test expectations for py3-commands test

context: do not cache manifestctx This will make sure when ctx.repo.manifestlog changes, a correct new manifestctx is returned. repo.manifestlog takes care of caching so the manifestctx won't be reconstructed every time.

test-context: add a case demonstrating manifest caching problem This issue was discovered when testing absorb on Windows. What happens are: 1. ctx.p1().manifestctx gets cached. let's call ctx.p1().manifestctx._revlog() "mrevlog1" 2. repo.manifestlog gets invalidated. let's call repo.manifestlog._revlog "mrevlog2" 3. repo.commitctx(ctx) commitctx uses ctx.p1().manifestctx and writes to "mrevlog1" 4. repo[n].manifest() cannot find the manifest node in "mrevlog2" This patch adds a test case to reproduce the issue.

debugbundle: display the content of obsmarkers parts We parse and display the markers in the part when possible.

bundle: add an experimental knob to include obsmarkers in bundle The "hg bundle" command is a good place to test if the inclusion of obsmarkers within a bundle is working well (part exists, content is correct etc). So we add a way to have them included. Ideally, this would be controlled by a change around bundlespec (bundlespec "v3" + arguments). However, my main goal is to have obsmarkers included in bundle created by the 'hg strip' command, not the 'hg bundle' so for now I'm avoiding the detour through bundlespec rework territory. Better debug output for obsmarkers in 'debugbundle' will be added in later changesets. The 'test-obsolete-bundle-strip.t' test will also get updated in a later changeset to keep the current changeset smaller.

bundle2: move function building obsmarker-part in the bundle2 module We move it next to similar part building functions. We will need it for the "writenewbundle" logic. This will allow us to easily include obsmarkers in on-disk bundle, a necessary step before having `hg strip` also operate on markers. (Yes, the bundle2 module was already too large, but there any many interdependencies between its components so it is non-trivial to split, this is a quest for another adventure.)

policy: remove unused policynoc and policynocffi constants

cffi: remove superfluous "if True" blocks

cffi: split modules from pure The copyright lines are updated per change history. cffi/osutil.py isn't tested since I have no access to OS X machine right now, sorry.

policy: extend API version checks for cffi This is just a stub for future extension. I could add a version constant to CFFI modules by putting it to both ffi.set_source() and ffi.cdef(), but that doesn't seem right. So for now, cffi modules will be explicitly unversioned (i.e. version constant must be undefined or set to None.) We can revisit it later when we need to consider CFFI support more seriously.

filterpyflakes: allow reexporting pure symbols from cffi modules cffi modules will do 'from ..pure.<module> import *'.

import-checker: guess names of C extension modules Since extension modules aren't included in the list of source files, they need to be populated somehow. Otherwise the import from cext/cffi would be treated as a global one.

import-checker: convert localmods to a set of module names This makes it easy to add a source-less module name to the set.

import-checker: allow importing symbols from pure modules This allows us to re-export pure functions from cffi modules: # mercurial/cffi/base85.py from ..pure.base85 import *

cffi: put compiled modules into mercurial.cffi package Don't pollute the top-level namespace.

cffi: rename build scripts This frees up cffi package for modules to be split from pure.

match: remove support for includes from patternmatcher Includes (and excludes) are now delegated to the includematcher.

match: simplify includematcher a bit The "include" we have in symbols is redundant and the double negative in visitdir() can be removed.

match: remove support for non-include patterns from includematcher The includematcher will always get at least one include pattern and will never get any non-include patterns, so we can remove most of the code in it. This patch does mostly straight-forward deletions of code. We will clean up further later.

match: split up main matcher into patternmatcher and includematcher At this point the includematcher is an exact copy of the main matcher class. We will specialize and simplify both classes in the following patches. This initial unmodified copy is just to make the differences clearer. We also rename the main matcher to "patternmatcher" for consistency. I may eventually merge this new includematcher back into the main matcher, but I think doing it this way makes the intermediate steps clearer regardless.

match: remove support for exact matching from main matcher class Exact matching is now handled by the exactmatcher class. We can safely remove _files from the __repr__() implementation, because even though the field is set, the patternspat field is enough for the representation to be unambiguous (which was not the case when the matcher could handle exact matches).

match: handle exact matching using new exactmatcher

merge: use intersectmatchers() in "m2-vs-ma optimization" It doesn't seem like this can actually happen, but seems like cleaner anyway.

match: handle includes using new intersectionmatcher

match: move entire uipath() implementation to basematcher Even though most matchers will always want to use the relative path in uipath(), when we add support for intersecting matcher, we will want to control which form to use for any kind of matcher without knowing the type (see next patch), so we need the implementation on the base class. Also rename the attribute from "pathrestricted" to "relativeuipath" since there actually are cases where we match everything but still use relative paths (like when the user runs "hg files .." from inside mercurial/).

local-clone: also copy tags related caches This caches provide a large speedup for some repositories. Keeping it around is valuable.

local-clone: also copy revs-branch-cache files This cache provides a large speedup for some repositories. Keeping it around is valuable.

local-clone: extract the listing of caches to copy Right now, the clone only copies the branchmap caches. There are multiple other valuable caches that we should copy and extensions might add their own. So we add a function to list the cache files to copy from the repository. The repository argument is unused but extensions will want it.

local-clone: extract the closure copying caches Closures often get on the way. They are not much value in having that as a closure so I'm extracting it at the module level.

test: add isolated prune case (to test-obsolete-bundle-strip.t) This adds a test where the prune marker is not related to any other obsmarkers.

test-obsolete-bundle-strip: add a complex split and fold case This is a more complex case that checks the logic used when split and fold gets into play.

test-obsolete-bundle-strip: add cases with prune on missing revs Same as the previously added case, but the prune is no longer known locally. This will mostly matter for the strip testing. Introducing the test early will help clarify patches related to strip.

obsolete: fix relevant-obsmarkers computation on pruned changeset The markers pruning a node was not directly considered relevant for the pruned node, only to its parents. This went unnoticed during obsmarkers exchange because all ancestors of the pruned node would be included in the computation. This still affects obsmarkers exchange a bit since "inline" prune markers would be ignored (see second test case). This went unnoticed, because in such case, we always push another obsolescence markers for that node. We add explicit tests covering this case. (The set of relevant changeset is use in the obsmarkers discovery protocol used in the evolve experimental extension, the impact will be handled on the extension side).

test: add a test file for relevant obsmarkers and its usage The logic around obsmarkers "relevant" to a set of revs have a couple of test around in other places but no systematic testing. In addition, all the current testing focus on the exchange case (we looks at relevant markers for '::heads'). For bundles, we'll need something a bit different. We'll no longer have set of revision going down to the repository roots. So we'll have to test these cases too. In addition, stripping obsmarkers will introduce new logic around obsmarkers that will need testing too. So a new test file make sense here. We start with a simple tests, more advanced cases are coming in the next changesets. The extra testing catch a minor bug (later in the series).

annotate: add a new experimental --skip option to skip revs This option is most useful for mechanical code modifications, especially ones that retain the same number of lines.

annotate: add core algorithm to skip a rev The core algorithm is inspired by git hyper-blame, implemented at https://chromium.googlesource.com/chromium/tools/depot_tools.git/+/master/git_hyper_blame.py. The heuristic is as documented in the comments.

annotate: make pair take all parents to pair against In upcoming patches we'll need to be aware of all parents at the same time. This also exposes a potential bug: if a line can be annotated with both parents of a merge commit, it'll always be annotated with p2, not p1. I'm not sure if that's what we want, but at least the code makes it clear now.

annotate: move pair function to top level We'll want to make this more complicated and have unit tests for it in upcoming patches.

bookmarks: fix check of hash-like name to not abort by ambiguous identifier 'mark in repo' may raise LookupError. I set it to not be warned since bookmark names shorter than 4 chars aren't checked and short names are likely to be ambiguous.

localrepo: document that __contains__() may raise LookupError

hidden: drop outdated comment about "dynamic" performance This comment is now irrelevant since we have a faster algorithm and no cache.

hidden: unify the static and dynamic blocker logic We no longer have cache and they both work the same way. Unifying the logic simplify the code and reduce the amount of set copies.

hidden: drop the hidden cache logic The improvement in time complexitty and the speed-up in computation is large enough that the has little use now. Its update time can even gets in the way. So we drop it. This will allow us to unify the static/dynamic blockers logic in the next changeset.

hidden: simplify the computation of consistency blocker For a couple of years, we now have precomputed set for all mutable phases. We can use this set restrict our search and quickly detect non-hideable children of hideable changesets. This speeds up the hidden computation. See docstring of the new function for details. This new version reuses the '_domainancestors' function to keep the computation of revealed changeset in O(len(visible)) Below are perfvolatilesets timing from two Mozilla repositories with different contents. hidden cache is disabled while obtaining them. 1) Mozilla repository with: * 400667 changesets * 35 hidden changesets (first rev-268334) * 288 visible drafts * 1 unstable changeset Before: ! visible ! wall 0.001744 comb 0.000000 user 0.000000 sys 0.000000 (best of 1563) After: ! visible ! wall 0.000742 comb 0.000000 user 0.000000 sys 0.000000 (best of 3755) The timing above include the computation of obsolete changeset: ! obsolete ! wall 0.000396 comb 0.000000 user 0.000000 sys 0.000000 (best of 6816) So adjusted time give 1.3ms before versus 0.3ms after. A 4x speedup. 2) Mozilla repository with: * 405645 changesets * 4312 hidden changesets (first rev-326004) * 264 visible drafts * 1 unstable changeset Before: ! visible ! wall 0.025476 comb 0.030000 user 0.030000 sys 0.000000 (best of 111) After ! visible ! wall 0.007703 comb 0.010000 user 0.010000 sys 0.000000 (best of 358) The timing above include the computation of obsolete changeset: ! obsolete ! wall 0.006408 comb 0.010000 user 0.010000 sys 0.000000 (best of 404) So adjusted time give 19ms before versus 1.3ms after. A 17x speedup.