tests: only stat a test file if we don't already know its "cost"

tests: write recent run times to a file named tests/.testtimes This is an effortless way to track how long each test takes without having to pass any extra command line options.

wireproto: support disabling bundle1 only if repo is generaldelta I recently implemented the server.bundle1* options to control whether bundle1 exchange is allowed. After thinking about Mozilla's strategy for handling generaldelta rollout a bit more, I think server operators need an additional lever: disable bundle1 if and only if the repo is generaldelta. bundle1 exchange for non-generaldelta repos will not have the potential for CPU explosion that generaldelta repos do. Therefore, it makes sense for server operators to continue to allow bundle1 exchange for non-generaldelta repos without having to set a per-repo hgrc option to change the policy depending on whether the repo is generaldelta. This patch introduces a new set of options to control bundle1 behavior for generaldelta repos. These options enable server operators to limit bundle1 restrictions to the class of repos that can be performance issues. It also allows server operators to tie bundle1 access to store format. In many server environments (including Mozilla's), legacy repos will not be generaldelta and new repos will or might be. New repos often aren't bound by legacy access requirements, so setting a global policy that disallows access to new/generaldelta repos via bundle1 could be a reasonable policy in many server environments. This patch makes this policy very easy to implement (modify global hgrc, add options to existing generaldelta repos to grandfather them in).

streamclone: use read() We have a convenience API for reading the full contents of a file. Use it.

internals: document revlog format It seems like a good idea to document the revlog format. There is a lot more that could be added to this documentation. But you have to start somewhere.

histedit: hide --outgoing warnings

histedit: list action when intervention is required

error: make InterventionRequired take a hint

histedit: hook afterresolvedstates

rebase: hook afterresolvedstates

graft: hook afterresolvedstates

resolve: suggest the next action Expose afterresolvedstates to allow graft and similar to suggest a message when resolving results in no unresolved files. If there isn't a matching state in afterresolvedstates, then if verbose, suggest commiting.

dispatch: report similar names consistently

cmdutil: pass node instead of ctx to diffordiffstat 93bcc73df8d5 changed showpatch to use ctx's more, but it accidentally passed prev as a context and node as a binary string, when both should be passed as binary strings (since diffordiffstat tries to resolve them via repo[X]). This affected hggit since the existing ctx belongs to the git overlay, but the resolved context (from the repo[X] resolution) should belong to the main repo. This broke a test because it tried to look in the git repo for data that didn't exist. This feels like a deeper issue in hggit somewhere, but the fix is here trivial and obviously more correct

import-checker: list package directory as stdlib module Before this patch, a directory containing __init__.py wasn't counted as a module and __init__.pyc was listed as foo.bar.__init__ module.

import-checker: normalize directory separator to get module name on Windows It didn't work if a path contains "\\". Therefore, ctypes.util couldn't be found on Windows.

mail: use print function We no longer use the print statement in mercurial.* \o/

lsprofcalltree: use print function While I was here, some single element tuples have been removed in favor of the shorter syntax. Some commented lines of code containing print statements have also been removed because it was unclear what purpose they served.

lsprof: use print function

keepalive: use print function

dispatch: use print function Python 3 doesn't have a print statement.

help: include section heading if section depth changes This makes it easier to distinguish between: "format" -------- "usestore" ... Enabled by default. and "progress.format" Format of the progress bar.

import: refactor nocommit and importbranch handling

import: refactor date and user handling

import: limit scope of msg in tryimportone

merge with stable

lazymanifest: check error return in filter

lazymanifest: add missing closing parenthesis in comment

help: rename the Format section to Structure This reduces the confusion from `hg help config.format` which deals with repository format. (But also profiling.format, and progress.format)

help: refactor version/defaults from format section The old messages implied that disabling a single setting would ensure compatibility, but if you disabled one of the older flags, and left a newer flag, that would not actually do what the docs say.

help: sort format section from newest to oldest

histedit: fix comment in newnodestoabort

histedit: fix comment in applychanges

run-tests: fix get port to try differing ports The code was moving its offset each time through the loop, but because it failed to update port, the port was not going to be available...

httpclient: update to 938f2107d6e2 of httpplus This enhances proxy support in httpclient a little bit, though I don't know that we used that functionality at all. It also switches httpplus to using absolute_import.

histedit: remove unused variable state

filemerge: default change/delete conflicts to 'leave unresolved' (BC) It makes far more sense to leave these conflicts unresolved and kick back to the user than to just assume that the local version be chosen. There are almost certainly buggy scripts and applications using Mercurial in the wild that do merges or rebases non-interactively, and then assume that if the operation succeeded there's nothing the user needs to pay attention to. (This wasn't possible earlier because there was no way to re-resolve change/delete conflicts -- but now it is.)

tests: explicitly request changed version in test-rebase-newancestor.t We're going to change the default for this in an upcoming patch, but in this instance we do want to continue picking the changed version.

tests: explicitly request changed version in c/d conflict in test-commit-amend.t We're going to change the default for this in an upcoming patch, but in this instance we do want to continue picking the changed version.

test-copy-move-merge.t: explicitly request changed version We're going to change the default for this in an upcoming patch, but in this instance we do want to continue picking the changed version.

match: add option to return line and lineno from readpattern This will be used to display the line and linenumber of ignorefile that matched an ignored file (issue4856).

dirstate: extract logic to compute the list of ignorefiles We are going to reuse this logic to improve debugignore in the next patches of the series.

dirstate: call the C implementation of nonnonormalentries when available Before this patch, we were using python code for computing the nonnormal dirstate entries. This patch makes us use the C implementation of the function when it is available. Using the nonnormal set in hgwatchman improves hg status performance. Below are the numbers for mozilla-central. with the changes: $ hg perfstatus ! wall 0.010632 comb 0.000000 user 0.000000 sys 0.000000 (best of 246) without the changes: $ hg perfstatus ! wall 0.036442 comb 0.030000 user 0.030000 sys 0.000000 (best of 100) On mozilla-central the improvement to hg status is ~20% (0.25s to 0.2s), on our big repos at Facebook, the win is ~40% (1.2s to 0.72s).

dirstate: add a C implementation for nonnormalentries Before this patch, there was only a python version of nonnormalentries. On mozilla-central we have a 10x win by putting this function in C: % python -m timeit -s \ 'from mercurial import hg, ui, parsers; \ repo = hg.repository(ui.ui(), "mozilla-central"); \ m = repo.dirstate._map' \ 'parsers.nonnormalentries(m)' 100 loops, best of 3: 3.15 msec per loop The python implementation runs in 31ms, a similar test gives: 10 loops, best of 3: 31.7 msec per loop On our big repos, the win is still of 10x with the python implementation running in 350ms and the C implementation running in 30ms.

dirstate: add test for non-normal set consistency This adds a test extension to check that the non-normal set contains the expected entries. It wraps several methods of the dirstate to check that the non-normal set has the correct values before and after the call. The extension lives in contrib so that paranoid developers can easily enable it to make sure that the non-normal set is consistent across more complex operations than the included tests.

dirstate: add code to update the non-normal set Before this patch, we were only populating the non-normal set when parsing or packing the dirstate. This was not enough to keep the non-normal set up to date at all time as we don't write and read the dirstate whenever a change happens. This patch solves this issue by updating the non-normal set when it should be updated. note: pack_dirstate changes the dmap and we have it keep it unchanged for retrocompatibility so we are forced to recompute the non-normal set after calling it.

dirstate: attach the nonnormalset to a propertycache This patch attaches the nonnormalset to a property cache so that we build it only when needed.

dirstate: add a function to compute non-normal entries from the dmap This patch adds a new python function in the dirstate to compute the set of non-normal files from the dmap. These files are useful to compute the repository status.

revset: use decorator to mark a predicate as safe Using decorator can localize changes for adding (or removing) a "safe" revset predicate function in source code. To avoid accidentaly treating unsuitable predicates as safe, this patch uses False as default value of "safe" argument. This forces safe predicates to be decorated with explicit 'safe=True'.

revset: use delayregistrar to register predicate in extension easily Previous patch introduced 'revset.predicate' decorator to register revset predicate function easily. But it shouldn't be used in extension directly, because it registers specified function immediately. Registration itself can't be restored, even if extension loading fails after that. Therefore, registration should be delayed until 'uisetup()' or so. This patch uses 'extpredicate' decorator derived from 'delayregistrar' to register predicate in extension easily. This patch also tests whether 'registrar.delayregistrar' avoids function registration if 'setup()' isn't invoked on it, because 'extpredicate' is the first user of it.

registrar: add delayregistrar class to register function in extensions 'delayregistrar' delays actual registration of function until 'setup()' invocation on it.

revset: use decorator to register a function as revset predicate Using decorator can localize changes for adding (or removing) a revset predicate function in source code. It is also useful to pick predicates up for specific purpose. For example, subsequent patch marks predicates as "safe" by decorator. This patch defines 'parsefuncdecl()' in 'funcregistrar' class, because this implementation can be uesd by other decorator class for fileset predicate and template function.

registrar: add funcregistrar class to register function for specific purpose This class centralizes the common logic to register function for specific purpose like below: - template keyword, filter and function - revset predicate - fileset predicate - webcommand 'funcregistrar' also formats help document of the function with the 'decl'(aration) specified at the construction. This can avoid (1) redundancy between 'decl' and help document, and (2) accidental typo of help document. For example, 'foo' should appear twice like below, if without such formatting: @keyword('foo') def foo(....): """:foo: Explanation of keyword foo ...""" Almost all cases needs very simple document formatting like below: - "``DECL``\n EXPLANATION" - ":DECL: EXPLANATION" But webcommand needs a little complicated formatting like: /PATH/SPEC ---------- EXPLANATION .... To make minirst recognize the section header, hyphen line should be as long as "/PATH/SPEC". It should be arranged by program. Implementing 'formatdoc()' in derived class can support complicated formatting in the latter case. But it seems redundant for simple one in the former case. Therefore, 'funcregistrar' does: - invoke 'self.formatdoc', if it is callable (for the latter case) - use it as the format string, otherwise (for the former case)

hgweb: support rendering a sub-topic If the requested topic contains a "." we assume a sub-topic is requested and display it.

hgweb: support rendering sub-topic indexes If the requested topic name is the name of a sub-topic, we now render an index of topics within that sub-topic.

templates: support linking to main help page Currently, the "helptopics" template assumes it is only used as the main index and therefore doesn't hyperlink "help" in the navigation list. Sub-topics will introduce an additional consumer of this template. So teach the template to hyperlink the "help" navigation entry when necessary.

templates: differentiate between partial and full topic name In order to support sub-topics, we need to support linking to a full topic name while displaying the base topic name. Change the {helpentry} template to grab the display name from an optional seperate variable (which will be defined in a future patch).

templates: make earlycommands and othercommands optional We now have sub-topics in the help system. The "helptopics" template serves as a mechanism for displaying an index of help topics. Previously, it was only used to show the top-level list of help topics, which includes special groupings of topics. In the near future, we'll adapt "helptopics" for showing the index of sub-topics. In this patch, we optionally render {earlycommands} and {othercommands} since they aren't present on sub-topics.

rebase: better error message when rebased changes are all in destination Before this patch, when rebasing a set of obsolete revisions that were plain pruned or already present in the destination, we were displaying: abort: no matching revisions This was not very helpful to understand what was going on, instead we replace the error message by: abort: all requested changesets have equivalents or were marked as obsolete (to force the rebase, set the config experimental.rebaseskipobsolete to False)

lrucachedict: add copy method This diff implements the standard dict copy() method for lrucachedicts, which will be used in the pushrebase extension to make a copy of the manifestcache.