run-tests: print more information on unnecessary glob matching Extend the message with the test name and the approximate line number. (The line number is the one of the command producing the output.) Finding the line to fix is easier now. old message: ...... Info, unnecessary glob: at a/b/c (glob) .. new message: ...... Info, unnecessary glob in test-example.t (after line 9): at a/b/c (glob) .. The test result is still pass as before.

run-tests: suggest to append glob when only path sep does not match When the line does not match because of \ instead of / (on windows), append (glob) in the expected output. This allows to rename test-bla.t.err to test-bla.t for getting a correct output. This worked for other failures like missing (esc), but not here. Output example (only +- lines of diff): Before: - path/with/local/sep + path\\with\\local/sep Now: - path/with/local/sep + path/with/local/sep (glob)

run-tests: test each line matching function on its own This has several advantages. * Each match function can return some information to the caller runone (used in the next patch). * It is not checked that the line ends in " (glob)" when rematch() returns false. * And it looks more readable.

tests: new test for line matching functions in run-tests Test for failing matches and warnings. (The existing test-run-tests.t can not do both by design.) And simulate matching on other os.

test-backout: add multiple summary calls to monitor result wc The main goal is to monitor that working directory parent are correct after backout. This will be useful the next changeset introducting magic merge usage.

transplant: use "ui.extractchoices()" to show the list of available responses Before this patch, transplant extension shows the list of available responses by specific string, even though the prompt string passed to "ui.promptchoice()" has enough (maybe i18n-ed) information. This patch uses "ui.extractchoices()" to show the list of available responses.

transplant: use "ui.promptchoice()" for interactive transplant Before this patch, transplant extension uses "ui.prompt()" for interactive transplant, and has to check whether user response returned by "ui.prompt()" is valid or not in own code. In addition to it, transplant extension uses response characters (e.g. "y", "n", and so on) directly in own code, and this disallows to use another response characters by translation, even though the help shown by '?' typing is translatable. This patch uses "ui.promptchoice()" instead of "ui.prompt()" to resolve problems above.

transplant: add test for interactive transplant

record: use "ui.extractchoices()" to get the list of available responses Before this patch, record extension gets the list of available responses from online help document of "hg record" in the tricky way, even though the value passed to "ui.promptchoice()" has enough (maybe i18n-ed) information. This patch uses "ui.extractchoices()" to get the list of available responses.

ui: add "extractchoices()" to share the logic to extract choices from prompt

branchmap: use set for update code We are doing membership test and substraction. new code is marginally faster.

branchmap: simplify update code We drop iterrevs which are not needed anymore. The know head are never a descendant of the updated set. It was possible with the old strip code. This simplification make the code easier to read an update.

branchmap: stop useless rev -> node -> rev round trip We never use the node of new revisions unless in the very specific case of closed heads. So we can just use the revision number. So give another handfull of percent speedup.

branchmap: stop membership test in update logic Now that no user try to update the cache on a truncated repo we can drop the extra lookup. Give an handfull percent speedup on big branchmap update.

branchmap: remove silly line break The line fit in 80 character limit without it. It is even shorter without it.

run-tests: set the thread name to the test name for info on error This does not happen when running normal. But when fiddling around with the test infrastructure, this helps a lot. Old traceback messge Exception in thread Thread-7: Traceback (most recent call last): ... new traceback message Exception in thread test-something.t: Traceback (most recent call last): ...

run-tests: report tests as failed when run-test raises an error Before no message was returned to the main thread. No result was registered and no new thread was started. This does not happen when running normal. But when fiddling around with the test infrastructure, this helps a lot.

hgweb: infinite scroll support for coal style

test-hgweb-*: output change fixes from b1d65cb8

hgweb: infinite scroll support for monoblue style

hgweb: infinite scroll support for gitweb style

hgweb: avoid invalid infinity scroll request when overwritten web.style Infinity scroll is broken when you override the web.style in the following ways: $ hg --config='web.style=gitweb' serve $ open http://localhost:8080/shortlog?style=paper ajaxScrollInit should use http://localhost:8080/shortlog/%next%?style=paper. however, http://localhost:8080/shortlog/%next% is used actually. It is missing style parameter. This patch add style parameter to request url.

hgweb: fix regexp for other styles like monoblue Some styles have indentation.

tests: introduce test for rebasing on named branches with closed heads

rebase: remove old code for handling empty rebaseset

rebase: improve error message for --base being empty or causing emptiness Before it just said 'nothing to rebase'. Now 'if "base" is an empty set: abort: empty "base" revision set - can't compute rebase set If the set of changesets to rebase can't be found from "base", it will fail as before but with more explanation of what the problem was. The name of the "base" option is not obvious - it is more like "samples identifying the branch to rebase". The error messages for problems with the specified "base" value will use that term and might thus also not be obvious, but at least they are consistent with the option name. The name "base" will not be used if the base only was specified implicitly as the working directory parent.

rebase: improve error message for empty --source set Before, it just said 'nothing to rebase' in this case. Now, it aborts mentioning the reason: 'empty "source" revision set'. Specifying revisions that cannot be rebased is a 'soft' error, but specifying an empty set deserves an abort that explains exactly what the problem is.

rebase: improve error message for empty --rev set Before, it just said 'nothing to rebase' in this case. Now, it aborts mentioning the reason: 'empty "rev" revision set'. Specifying revisions that cannot be rebased is a 'soft' error, but specifying an empty set deserves an abort that explains exactly what the problem is.

rebase: test for empty dest revision

help: branch names primarily denote the tipmost unclosed branch head Was the behavior correct and the description wrong so it should be updated as in this patch? Or should the code work as the documentation says? Both ways could make some sense ... but none of them are obvious in all cases. One place where it currently cause problems is when the current revision has another branch head that is closer to tip but closed. 'hg rebase' refuses to rebase to that as it only see the tip-most unclosed branch head which is the current revision. /me kind of likes named branches, but no so much how branch closing works ...

util: introduce util.debugstacktrace for showing a stack trace without crashing This is often very handy when hacking/debugging. Calling util.debugstacktrace('hey') from a place in hg will give something like: hey at: ./hg:38 in <module> /home/user/hgsrc/mercurial/dispatch.py:28 in run /home/user/hgsrc/mercurial/dispatch.py:65 in dispatch /home/user/hgsrc/mercurial/dispatch.py:88 in _runcatch /home/user/hgsrc/mercurial/dispatch.py:740 in _dispatch /home/user/hgsrc/mercurial/dispatch.py:514 in runcommand /home/user/hgsrc/mercurial/dispatch.py:830 in _runcommand /home/user/hgsrc/mercurial/dispatch.py:801 in checkargs /home/user/hgsrc/mercurial/dispatch.py:737 in <lambda> /home/user/hgsrc/mercurial/util.py:472 in check ...

check-code: print debug output when an ignore pattern matches

check-code: drop now unused check-code-ignore Using check-code-ignore to skip the failures on a line has several disadvantages: * It skips all check-code failures on a line, not only the one it was created for. * It does not give any hint for which rule it was added, making it difficult to see when it is not needed anymore. So drop this pragma in favor of better alternatives promoted before.

check-code: explain what to do when a check-code rule mismatches In the past several approaches were used when a check-code rule triggered without a good reason. Not all of them looked nice, some were even wrong. Suggest some good practices which should be used instead.

cleanup: Remove the only ever used skip-check-code pragma Use the work-around suggested by the rule instead

check-code: always report when a file is skipped by "no-check-code" Skipping an entire file generally from checking is an important event, so report it always. Do not tell the check name because skipping does not depend on it. Directly skip the entire file instead of checking more patterns and skip again. The pragma no-check-code was introduced by accident in the past. (Fixed in e033a7d444ac and ee07f9d142c9.) This now is prevented because the files to skip have to be listed in the test output of test-check-code-hg.t.

check-code: do not skip entire file, skip only one match instead Skipping of the entire file has been introduced in bc3b48b0f5c8.

bisect: --command without --noupdate should flag the parent rev it tested 14913fcb30c6 not only introduced the 'bisect(current)' revset predicate, it also changed how the 'current' revision is used in combination with --command. The new behaviour might be ok for --noupdate where the working directory and its revision shouldn't be used, but it also did that when --command is used to run a command on the currently checked out revision then it could register the test result on the wrong revision. An example: Before, bisect with --command could use the wrong revision when recording the test result: $ hg up -qr 0 $ hg bisect --command "python \"$TESTTMP/script.py\" and some parameters" changeset 31:58c80a7c8a40: bad abort: inconsistent state, 31:58c80a7c8a40 is good and bad Now it works as before and as expected and uses the working directory revision for the --command result: $ hg up -qr 0 $ hg bisect --command "python \"$TESTTMP/script.py\" and some parameters" changeset 0:b99c7b9c8e11: bad ...

context: drop caching 'copies' method The 'copies' method has no test coverage and calls copies.pathcopies with an incorrect number of parameters and is thus (fortunately) not used. Kill it.

commands.bookmarks: move hexfn to inside list block This isn't used outside this block, nor is it expected to be.

commands.bookmarks: move cur initialization to inside wlock This is more correct because we now fetch '.' while nothing else can interfere with it.

commands.bookmarks: pass cur in explicitly to checkconflict cur will be moved inside the wlock in a future patch, so we need to pass it into checkconflict explicitly.

commands.bookmarks: hold wlock for write operations Any invocations of bookmarks other than a plain 'hg bookmarks' will likely cause a write to the bookmark store. These should be guarded by the wlock. The repo._bookmarks read should be similarly guarded by the wlock if we're going to be subsequently writing to it.

commands.bookmarks: separate out 'no bookmarks set' status messages Upcoming patches will acquire the wlock for write operations, such as make inactive, but not read-only ones, such as list bookmarks. Separate out the status messages so that the code paths can be separated.

tests: test-convert-cvs-synthetic.t requires cvs 1.12 9589227657bc introduced hghave cvs112 10 months ago. Let's assume it has stabilized so much that we can start using it now.

merge with stable

i18n-de: rename noun "entfernt" to "Gegenseite" The German translation for "remote" as "entfernt" can be misleading in situations where remote is used as a noun. "entfernt" is not a noun and can also mean "removed". To clarify this we rename "remote" to "Gegenseite" when used as a noun.

doc: bump copyright year

localrepo: drop unused variable

discovery: stop using nodemap for membership testing Nodemap is not aware of filtering so we need to ask the changelog itself if a node is known. This is probably a bit slower but such check does not dominated discovery time. This is necessary if we want to run discovery on filtered repo.

discovery: enforce filtering into revlogbaseddag._internalizeall One more step toward discovery running on filtered repo.

discovery: make revlogdag work on filtered repo The revlogdag class is a core part of discovery. We need its initialisation to exclude revision filtered out.

pull: run findcommon incoming on unfiltered repo The discovery is not yet ready for filtered repo. Pull was using filtered for its discovery which is wrong. It worked by dumb luck because discovery mainly use funtion that does not respect the filtering. Trying to makes discovery work on filtered repo revealed this bug.

push: more robust check for bundle fast path When all changesets in the local repo are either being pushed or remotly known, we can take a fast path when bundling changeset because we are certain all local deltas are computed againts base known remotely. So we have a check to detect this situation, when we did a bare push and nothing was excluded. In a coming refactoring, the discovery will run on filtered view and the content of `outgoing.excluded` will just include unserved (secret) changeset not filtered by the repoview used to call push (usually "visible"). So we need to check if there is both no excluded changeset and nothing filtered by the current repoview.

pull: fix post-pull common computation Before that changes, pulled revision that happend to be already known locally (so, not actually added) was not taken into account when computing the new common set between local and remote. It appears that we already know the heads of the pulled set. It is in the `rheads` variable, so we are just using it and everything is works fine. We are dropping the, now useless, computation of `added` set in the process.

run-tests: better check for python version Compare version by using pythons tuple comparison. So we do not match on python 3.0 or newer.

templatekw: allow tagtypes other than global in getlatesttags hg-git uses tagtype 'git', for example, so it's better to check for tagtype != 'local', not strictly for 'global'

revlog: move file writing to a separate function Moves the code that actually writes to a file to a separate function in revlog.py. This allows extensions to intercept and use the data being written to disk. For example, an extension might want to replicate these writes elsewhere. When cloning the Mercurial repo on /dev/shm with --pull, I see about a 0.3% perf change. It goes from 28.2 to 28.3 seconds.

discovery: prefer loop to double-for list comprehension in changegroupsubset The double-for form of list comprehensions gets particularly unreadable when you throw in an 'if' condition. This expands the only remaining instance of the double-for syntax in our codebase into a loop.

merge with stable

merge with stable

Added signature for changeset ca387377df7a

Added tag 2.8.2 for changeset ca387377df7a

merge with i18n

i18n-ja: synchronized with d4be314b2071

merge with stable

fileset, revset: do not use global parser object for thread safety parse() cannot be called at the same time because a parser object keeps its states. This is no problem for command-line hg client, but it would cause strange errors in multi-threaded hgweb. Creating parser object is not too expensive. original: % python -m timeit -s 'from mercurial import revset' 'revset.parse("0::tip")' 100000 loops, best of 3: 11.3 usec per loop thread-safe: % python -m timeit -s 'from mercurial import revset' 'revset.parse("0::tip")' 100000 loops, best of 3: 13.1 usec per loop

obsolete: order of magnitude speedup in _computebumpedset Reminder: a changeset is said "bumped" if it tries to obsolete a immutable changeset. The previous algorithm for computing bumped changeset was: 1) Get all public changesets 2) Find all they successors 3) Search for stuff that are eligible for being "bumped" (mutable and non obsolete) The entry size of this algorithm is `O(len(public))` which is mostly the same as `O(len(repo))`. Even this this approach mean fewer obsolescence marker are traveled, this is not very scalable. The new algorithm is: 1) For each potential bumped changesets (non obsolete mutable) 2) iterate over precursors 3) if a precursors is public. changeset is bumped We travel more obsolescence marker, but the entry size is much smaller since the amount of potential bumped should remains mostly stable with time `O(1)`. On some confidential gigantic repo this move bumped computation from 15.19s to 0.46s (×33 speedup…). On "smaller" repo (mercurial, cubicweb's review) no significant gain were seen. The additional traversal of obsolescence marker is probably probably counter balance the advantage of it. Other optimisation could be done in the future (eg: sharing precursors cache for divergence detection)

obsolete: add an allprecursors method mirroring allsuccessors one. Detection of bumped changeset should use `allprecursors(<mutable>)` instead or `allsuccessors(<immutable>)` so we need the all precursors function to exists.

perf: fix perfvolatilesets The repoview's `filteredrevs` has been renamed to `filterrevs` at some point. perf was never informed.

obsolete: improve allsuccessors doc string The fact original nodes are also yield is not obvious. We update the docstring to highlight it.

obsolete: fix bad comment We cannot afford such extra "with" they are far too pricy.

util: remove unused realpath (issue4063) util.realpath was in use for only 5 days from dbdb777502dc until it was backed out in c519cd8f0169 because it caused issue3077 and issue3071.

import-checker: suppress check-code about any() ast is a new enough module that this script can't work on any version of Python without any(), so we'll just use it.

import-checker: use any() and a genexp to avoid awkward for/else construction

import-checker: backout 40f79b9a2cc8 (issue4129) This patch backs out 40f79b9a2cc8, which caused test-module-imports.t to be skipped when the test was run using virtualenv. Since the test now passes when using virtualenv, the skip is no longer necessary.

import-checker: make test-module-imports.t work using virtualenv (issue4129) This patch modifies contrib/import-checker.py so that test-module-imports.t will pass if run using virtualenv. The patch achieves this by adding two new prefixes to the list of allowable sys.path prefixes. The added prefixes are the directories of two modules in the stdlib. The modules selected are a minimal set that allowed the return value of list_stdlib_modules() to match the return value without virtualenv, when run on the patch author's machine: Mac OS X 10.8, Python 2.7.6.

import-checker: refactor sys.path prefix check (issue4129) This patch refactors the logic in contrib/import-checker.py responsible for checking the beginnings of the paths in sys.path. In particular, it adds a variable that defines the set of allowed prefixes. The primary purpose of this change is to make it easier to add more allowed prefixes. This will be useful in resolving issue4129, which involves making the function list_stdlib_modules() work when run from a virtualenv.

filter: add a comment so that people do not forget to update subsettable Changeset 175c6fd8cacc moved `subsettable` from `mercurial/repoview.py` to `mercurial/branchmap.py`. This mean that `filtertable` and `subsettable` are no longer next to each other. So we add a comment to remind people to update both.

localrepo: remove unused repo.branchtags()/_branchtip() methods

hgweb: simplify branches with repo.branchmap().iterbranches()

hgweb: simplify summary with repo.branchmap().iterbranches()

branches: simplify with repo.branchmap().iterbranches() Running hg branches on the PyPy repo (with 996) over a busy NFS server, before this change: $ time hg --profile branches > /dev/null CallCount Recursive Total(s) Inline(s) module:lineno(function) 1069 0 1.2955 1.2955 <open> 1063 0 0.5576 0.5576 <method 'close' of 'file' objects> 4122 0 0.1993 0.0449 mercurial.repoview:161(changelog) 8240 0 0.0771 0.0299 mercurial.changelog:133(tip) 4122 0 0.0422 0.0204 mercurial.localrepo:26(__get__) 8240 0 0.0252 0.0131 <len> 4122 0 0.0060 0.0037 mercurial.repoview:112(filterrevs) 8240 0 0.0028 0.0028 <hash> 3029 0 0.2139 0.0390 mercurial.context:202(__init__) 3029 0 0.1402 0.0339 mercurial.repoview:161(changelog) 3029 0 0.0240 0.0087 mercurial.changelog:183(rev) 9087 0 0.0067 0.0067 <isinstance> 1096 0 0.0025 0.0025 <binascii.unhexlify> 4125 0 0.0015 0.0015 <len> 4229 0 0.0344 0.0344 mercurial.revlog:296(rev) 1061 0 0.0343 0.0343 <method 'seek' of 'file' objects> 1063 0 0.0339 0.0339 <method 'read' of 'file' objects> 40476 16488 0.0479 0.0311 <len> 16488 0 0.0216 0.0168 mercurial.revlog:262(__len__) 8240 0 0.0771 0.0299 mercurial.changelog:133(tip) 8240 0 0.0281 0.0203 mercurial.changelog:190(node) 8240 0 0.0191 0.0095 <len> 1342 0 0.0278 0.0278 <zlib.decompress> 1074 0 2.2143 0.0266 mercurial.changelog:270(read) 1074 0 2.1328 0.0230 mercurial.revlog:907(revision) 1073 0 0.0208 0.0108 mercurial.changelog:28(decodeextra) 2148 0 0.0072 0.0072 <method 'split' of 'str' objects> 2148 0 0.0211 0.0038 mercurial.encoding:61(tolocal) 1074 0 0.0028 0.0028 <method 'index' of 'str' objects> 1061 0 1.9811 0.0237 mercurial.revlog:817(_loadchunk) real 0m2.742s user 0m0.811s sys 0m0.188s After this change: $ time hg --profile branches > /dev/null CallCount Recursive Total(s) Inline(s) module:lineno(function) 2092 0 0.1444 0.0292 mercurial.context:202(__init__) 2092 0 0.0908 0.0216 mercurial.repoview:161(changelog) 2092 0 0.0164 0.0057 mercurial.changelog:183(rev) 6276 0 0.0045 0.0045 <isinstance> 1096 0 0.0024 0.0024 <binascii.unhexlify> 3188 0 0.0013 0.0013 <len> 2218 0 0.0230 0.0230 mercurial.revlog:296(rev) 2111 0 0.1028 0.0218 mercurial.repoview:161(changelog) 4218 0 0.0387 0.0146 mercurial.changelog:133(tip) 2111 0 0.0238 0.0104 mercurial.localrepo:26(__get__) 4218 0 0.0122 0.0062 <len> 2111 0 0.0038 0.0021 mercurial.repoview:112(filterrevs) 4218 0 0.0014 0.0014 <hash> 20240 8444 0.0233 0.0149 <len> 8444 0 0.0110 0.0084 mercurial.revlog:262(__len__) 4218 0 0.0387 0.0146 mercurial.changelog:133(tip) 4218 0 0.0144 0.0103 mercurial.changelog:190(node) 4218 0 0.0097 0.0048 <len> 2398 1 0.0271 0.0115 mercurial.localrepo:26(__get__) 2398 1 0.0146 0.0046 mercurial.scmutil:939(__get__) 2124 0 0.0009 0.0009 mercurial.localrepo:330(unfiltered) 274 0 0.0002 0.0002 mercurial.repoview:192(unfiltered) 4 0 0.1409 0.0112 mercurial.branchmap:19(read) 1096 0 0.1113 0.0028 mercurial.localrepo:407(__contains__) 1098 0 0.0020 0.0020 <method 'split' of 'str' objects> 1097 0 0.0019 0.0019 <binascii.unhexlify> 1096 0 0.0093 0.0018 mercurial.encoding:61(tolocal) 1096 0 0.0010 0.0010 <method 'append' of 'list' objects> 4349 0 0.0150 0.0105 mercurial.changelog:190(node) 4349 0 0.0045 0.0045 mercurial.revlog:317(node) real 0m0.362s user 0m0.329s sys 0m0.024s

debuglabelcomplete: simplify with repo.branchmap().iterbranches()

branchmap: introduce iterbranches() method

localrepo: refactor repo.branchheads() to use repo.branchmap().branchheads()

branchmap: introduce branchheads() method

localrepo: refactor repo.branchtip() to use repo.branchmap().branchtip()

branchmap: introduce branchtip() method

branchmap: cache open/closed branch head information This lets us determine the open/closed state of a branch without reading from the changelog (which can be costly over NFS and/or with many branches).

bookmarks: allow push -B to create a new remote head (issue2372) Push is currently allowed to create a new head if there is a remote bookmark that will be updated to point to the new head. If the bookmark is not known remotely then push aborts, even if a -B argument is about to push the bookmark. This change allows push to continue in this case. This does not require a wireproto force.

template: modify showextras to return a hybrid This modifies slightly the behavior introduced in 519120a96c63 to allow showextras to return a hybrid, rather than showlist. The example in the template help file now executes and returns meaningful results.

branches: avoid unnecessary changectx.branch() calls This requires reading from the changelog, which can be costly over NFS. Note that this does not totally remove reading from the changelog; we still do that when calling changectx.closesbranch(). That call will be removed in a later patch. Running hg branches on the PyPy repo (with 996) over a busy NFS server, before this change: $ time hg --profile branches > /dev/null CallCount Recursive Total(s) Inline(s) module:lineno(function) 2042 0 2.2827 2.2827 <open> 2036 0 0.9840 0.9840 <method 'close' of 'file' objects> 2036 0 0.0464 0.0464 <method 'read' of 'file' objects> 5233 0 0.1985 0.0453 mercurial.repoview:161(changelog) 10462 0 0.0791 0.0314 mercurial.changelog:133(tip) 5233 0 0.0388 0.0176 mercurial.localrepo:26(__get__) 10462 0 0.0250 0.0126 <len> 5233 0 0.0059 0.0039 mercurial.repoview:112(filterrevs) 10462 0 0.0029 0.0029 <hash> 2034 0 0.0444 0.0444 <method 'seek' of 'file' objects> 5340 0 0.0390 0.0390 mercurial.revlog:296(rev) 2582 0 0.0371 0.0371 <zlib.decompress> 3155 0 0.1963 0.0366 mercurial.context:202(__init__) 3155 0 0.1238 0.0306 mercurial.repoview:161(changelog) 3155 0 0.0261 0.0080 mercurial.changelog:183(rev) 9465 0 0.0061 0.0061 <isinstance> 1096 0 0.0023 0.0023 <binascii.unhexlify> 4251 0 0.0014 0.0014 <len> 2059 0 3.7341 0.0332 mercurial.changelog:270(read) 2059 0 3.6304 0.0307 mercurial.revlog:907(revision) 2057 0 0.0262 0.0137 mercurial.changelog:28(decodeextra) 4118 0 0.0094 0.0094 <method 'split' of 'str' objects> 4118 0 0.0270 0.0048 mercurial.encoding:61(tolocal) 2059 0 0.0040 0.0040 <method 'index' of 'str' objects> 10462 0 0.0791 0.0314 mercurial.changelog:133(tip) 10462 0 0.0289 0.0207 mercurial.changelog:190(node) 10462 0 0.0188 0.0091 <len> 52433 20932 0.0478 0.0310 <len> 20932 0 0.0221 0.0168 mercurial.revlog:262(__len__) 2059 0 3.6304 0.0307 mercurial.revlog:907(revision) real 0m4.361s user 0m0.986s sys 0m0.237s After this change: $ time hg --profile branches > /dev/null CallCount Recursive Total(s) Inline(s) module:lineno(function) 1069 0 1.1098 1.1098 <open> 1063 0 0.4865 0.4865 <method 'close' of 'file' objects> 4122 0 0.1811 0.0404 mercurial.repoview:161(changelog) 8240 0 0.0712 0.0272 mercurial.changelog:133(tip) 4122 0 0.0378 0.0177 mercurial.localrepo:26(__get__) 8240 0 0.0221 0.0115 <len> 4122 0 0.0057 0.0033 mercurial.repoview:112(filterrevs) 8240 0 0.0025 0.0025 <hash> 3029 0 0.1979 0.0371 mercurial.context:202(__init__) 3029 0 0.1278 0.0310 mercurial.repoview:161(changelog) 3029 0 0.0230 0.0081 mercurial.changelog:183(rev) 9087 0 0.0061 0.0061 <isinstance> 1096 0 0.0026 0.0026 <binascii.unhexlify> 4125 0 0.0014 0.0014 <len> 4229 0 0.0337 0.0337 mercurial.revlog:296(rev) 1061 0 0.0296 0.0296 <method 'seek' of 'file' objects> 1063 0 0.0292 0.0292 <method 'read' of 'file' objects> 8240 0 0.0712 0.0272 mercurial.changelog:133(tip) 8240 0 0.0271 0.0196 mercurial.changelog:190(node) 8240 0 0.0169 0.0083 <len> 40476 16488 0.0422 0.0271 <len> 16488 0 0.0193 0.0152 mercurial.revlog:262(__len__) 1342 0 0.0241 0.0241 <zlib.decompress> 9445 0 0.0336 0.0224 mercurial.changelog:190(node) 9445 0 0.0112 0.0112 mercurial.revlog:317(node) 1074 0 1.9102 0.0224 mercurial.changelog:270(read) 1074 0 1.8397 0.0202 mercurial.revlog:907(revision) 1073 0 0.0187 0.0099 mercurial.changelog:28(decodeextra) 2148 0 0.0061 0.0061 <method 'split' of 'str' objects> 2148 0 0.0184 0.0034 mercurial.encoding:61(tolocal) real 0m2.402s user 0m0.735s sys 0m0.177s

branchmap: add documentation on the branchcache on-disk format

revlog: allow tuning of the chunk cache size (via format.chunkcachesize) Running perfmoonwalk on the Mercurial repo (with almost 20,000 changesets) on Mac OS X with an SSD, before this change: $ hg --config format.chunkcachesize=1024 perfmoonwalk ! wall 2.022021 comb 2.030000 user 1.970000 sys 0.060000 (best of 5) (16,154 cache hits, 3,840 misses.) $ hg --config format.chunkcachesize=4096 perfmoonwalk ! wall 1.901006 comb 1.900000 user 1.880000 sys 0.020000 (best of 6) (19,003 hits, 991 misses.) $ hg --config format.chunkcachesize=16384 perfmoonwalk ! wall 1.802775 comb 1.800000 user 1.800000 sys 0.000000 (best of 6) (19,746 hits, 248 misses.) $ hg --config format.chunkcachesize=32768 perfmoonwalk ! wall 1.818545 comb 1.810000 user 1.810000 sys 0.000000 (best of 6) (19,870 hits, 124 misses.) $ hg --config format.chunkcachesize=65536 perfmoonwalk ! wall 1.801350 comb 1.810000 user 1.800000 sys 0.010000 (best of 6) (19,932 hits, 62 misses.) $ hg --config format.chunkcachesize=131072 perfmoonwalk ! wall 1.805879 comb 1.820000 user 1.810000 sys 0.010000 (best of 6) (19,963 hits, 31 misses.) We may want to change the default size in the future based on testing and user feedback.

revlog: read/cache chunks in fixed windows of 64 KB When reading a revlog chunk, instead of reading up to 64 KB ahead of the request offset and caching that, this change caches a fixed window before and after the requested data that falls on 64 KB boundaries. This increases cache hits when reading revlogs backwards. Running perfmoonwalk on the Mercurial repo (with almost 20,000 changesets) on Mac OS X with an SSD, before this change: $ hg perfmoonwalk ! wall 2.307994 comb 2.310000 user 2.120000 sys 0.190000 (best of 5) (Each run has 10,668 cache hits and 9,304 misses.) After this change: $ hg perfmoonwalk ! wall 1.814117 comb 1.810000 user 1.810000 sys 0.000000 (best of 6) (19,931 cache hits, 62 misses.) On a busy NFS share, before this change: $ hg perfmoonwalk ! wall 17.000034 comb 4.100000 user 3.270000 sys 0.830000 (best of 3) After: $ hg perfmoonwalk ! wall 1.746115 comb 1.670000 user 1.660000 sys 0.010000 (best of 5)