configitems: get default values from the central registry when available We do not have any registered config yet, but we are now ready to use them. For now we ignore this feature for config access with "alternates". On the long run, we expect alternates to be handled as "aliases" by the config item themself.

configitems: introduce a central registry for config option We now have the appropriate infrastructure to register config items. Usage will added in the next changeset.

configitems: add a basic class to hold config item information The goal of this class is allow explicit declaration for the available config option. This class will hold the data for one specific config item. To keep it simple we start centralizing the handling of the default config value. In the future we can expect more data to be carried on this class. For example: - documentation, - status (experimental, advanced, normal, deprecated), - aliases, - expected type, - etc...

run-tests: fix -i when "#testcases" is used in .t test The "#testcases" feature introduced by 7340465bd788 has issues with "-i" because "-i" uses "test.name.endswith('.t')" to test if a test is .t or not. test.name could now be something like "test-foo.t (caseA)" so the above endswith test is no longer valid. This patch changes the test to use "self.path" which won't have the issue.

run-tests: update .t reference output after reading the test The .t file is both test input and reference output. They should always match. However we have different code paths to read reference output (Test.__init__ -> Test.readrefout) and test input (TTest._run) so they might be inconsistent if somethings change the file between those two functions. This patch assigns "lines" read by "_run" back to "_refout" if "_refout" is not None (with --debug, see Test.readrefout) so reference output and test input will always match.

run-tests: do not prompt changes (-i) if a race condition is detected The race condition is like: 1. run-tests.py reads test-a.t as reference output, content A 2. run-tests.py runs the test (which could be content B, another race condition fixed by the next patch, but assume it's content A here) 3. something changes test-a.t to content C 4. run-tests.py compares test output (content D) with content A 5. with "-i", run-tests.py prompts diff(A, D), while the file has content C instead of A at this time This patch detects the above case and tell the user to rerun the test if they want to apply test changes.

patch: rewrite reversehunks (issue5337) The old reversehunks code accesses "crecord.uihunk._hunk", which is the raw recordhunk without crecord selection information, therefore "revert -i" cannot revert individual lines, aka. issue5337. The patch rewrites related logic to return the right reverse hunk for revert. Namely, 1. "fromline" and "toline" are correctly swapped [1] 2. crecord.uihunk generates a correct reverse hunk [2] Besides, reversehunks(hunks) will no longer modify its input "hunks", which is more expected. [1]: To explain why "fromline" and "toline" need to be swapped, take the following example: $ cat > a <<EOF > 1 > 2 > 3 > 4 > EOF $ cat > b <<EOF > 2 > 3 > 5 > EOF $ diff a b 1d0 <---- "1" is "fromline" and "0" is "toline" < 1 and they are swapped if diff from the reversed direction 4c3 | < 4 | --- | > 5 | | $ diff b a | 0a1 <---------+ > 1 3c4 <---- also "4c3" gets swapped to "3c4" < 5 --- > 4 [2]: This is a bit tricky. For example, given a file which is empty in working parent but has 3 lines in working copy, and the user selection: select hunk to discard [x] +1 [ ] +2 [x] +3 The user intent is to drop "1" and "3" in working copy but keep "2", so the reverse patch would be something like: -1 2 (2 is a "context line") -3 We cannot just take all selected lines and swap "-" and "+", which will be: -1 -3 That patch won't apply because of "2". So the correct way is to insert "2" as a "context line" by inserting it first then deleting it: -2 +2 Therefore, the correct revert patch is: -1 -2 +2 -3 It could be reordered to look more like a common diff hunk: -1 -2 -3 +2 Note: It's possible to return multiple hunks so there won't be lines like "-2", "+2". But the current implementation is much simpler. For deletions, like the working parent has "1\n2\n3\n" and it was changed to empty in working copy: select hunk to discard [x] -1 [ ] -2 [x] -3 The user intent is to drop the deletion of 1 and 3 (in other words, keep those lines), but still delete "2". The reverse patch is meant to be applied to working copy which is empty. So the patch would be: +1 +3 That is to say, there is no need to special handle the unselected "2" like the above insertion case.

profiling: cope with configwith default value handling changes Changeset 6ff6eb33f353 change 'configwith' behavior so that the default value is run through the conversion function. In parallel a new user of 'configwith' got introduced unaware of this coming behavior change. This broke profiling. We resolve the situation by having the new conversion function cope with a default value already using the right type.

py3: catch StopIteration from next() in generatorset IIUC, letting the StopIteration through would not cause any bugs, but not doing it makes the test-py3-commands.t pass. I have also diligently gone through all uses of next() in our code base. They either: * are not called from a generator * pass a default value to next() * catch StopException * work on infinite iterators * request a fixed number of items that matches the generated number * are about batching in wireproto which I didn't quite follow I'd appreciate if Augie or someone else could take a look at the wireproto batching and convince themselves that the next(batchable) calls there will not raise a StopIteration.

tests: adjust quoting to keep Windows happy with recent $PYTHON change I tried adding quotes to the $PYTHON variable, and also tried converting the path from the current 'c:/Python/python.exe' form to '/c/python/python.exe', but neither worked. I'm not sure why one of these needs '\"' around the variable and the other doesn't.

bundle2: don't use debug message "no-transaction" with transaction

py3: use pycompat.bytestr() in place of str()

py3: use r'' to access values from kwargs where keys are str These are the cases where either args is again passed as keyword argument or 1 or 2 elements are accessed. So it's better to add an r'' to prevent it converting to bytes rather than doing the conversion of args.

py3: convert keys of kwargs in template keywords functions to bytes This patch converts the args argument keys' to bytes wherever necessary as there are some places where either args is not used or using r'' is better or args is again passed as keyword arguments.

py3: make sure the commands name are bytes in test-devel-warnings.t

py3: replace str with bytes in isinstance() We were using str because on Python 2, str were bytes but now we have to use bytes. Otherwise the if conditions fails and we have weird results from commands on Python 3.

py3: catch binascii.Error raised from binascii.unhexlify Before Python 3, binsacii.unhexlify used to raise TypeError, now it raises binascii.Error.

shelve: allow unlimited shelved changes per name Previously, there is a 100 changes limit per name (bookmark or named branch). And the user will get "too many shelved changes named %s" when they are trying to shelve the 101th change. I hit that error message today. This limit was introduced by the shelve extension since the beginning. The function generating the names was called "gennames", under "getshelvename". There is another "gennames" under "backupfilename": def backupfilename(self): def gennames(base): yield base base, ext = base.rsplit('.', 1) for i in itertools.count(1): yield '%s-%d.%s' % (base, i, ext) "itertools.count" is an endless counter. Since the other "gennames" generates unlimited number of names, and the changeset introducing the limit (49d4919d21) does not say why the limit is useful. It seems safe to just remove the limit. The format "%02d" was kept intentionally so existing shelved changes won't break.

config: use the new '_unset' value for 'configsuboptions' This should let configsuboptions delegate all special processing of the default config value to the main 'config' method.

config: use the 'config' method in 'configsuboptions' There was unnecessary code duplication. It was getting in the way of the unification of the default value logic.

config: use the new '_unset' value for 'configpath' This should let 'configpath' delegate all special processing of the default config value to the main 'config' method.

config: use the new '_unset' value for 'configdate' This should let 'configdate' delegate all special processing of the default config value to the main 'config' method. The default value for date (None) is still enforced in this method if no other default were passed.

config: use the new '_unset' value for 'configlist' This should let 'configlist' delegate all special processing of the default config value to the main 'config' method. The default config value ([]) is still handled in this method.

config: use the new '_unset' value for 'configbytes' This should let 'configbytes' delegates all special processing of the default config value to the main 'config' method.

config: use the new '_unset' value for 'configint' This should let 'configint' delegates all special processing of the default config value to the main 'config' method.

config: use the new '_unset' value for 'configwith' This should let 'configwith' delegate all special processing of the default config value to the main 'config' method. This changeset introduce a small change in behavior since the default value is run through the 'convert' function. This does not seems harmful and no actual test break. This small change make the code simpler so I'm keeping it.

config: use the new '_unset' value for 'configbool' This should let 'configbool' delegate all special processing of the default config value to the main 'config' method. The default value for bool (False) is still enforced in this method if no other default were passed.

config: explicitly track the use of the standard default value We introduce a small object used to detect that no specific default value has been passed to 'ui.config'. We need this explicit special value since "None" is a valid and common default value. The end goal here is to make progress on a centralised and explicit declaration of the available config option. A first good usecase for this are "default" value. Before starting looking further down this alley we needs to rework the handling of default value in the 'ui' object to have all configxyz methods going through the same logic. This is the first changeset on this trek.

clonebundle: update hook arguments (BC) By calling applybundle() with 'clonebundles' and the url instead of calling processbundle(), the hooks will get different arguments: HG_SOURCE will be 'clonebundles' instead of 'bundle2' and HG_URL will be the url instead of 'bundle2'. This is consistent with the bundle1 behavior and seems like a bug fix, but I'm marking it BC anyway.

commands: move checkconflict to bookmarks module Again, commands.bookmark is getting too large. checkconflict already has a lot of state and putting it in the bmstore makes more sense than having it as a closure. This also allows extensions a place to override this behavior. While we're here, add a documentation string because, well, we should be documenting more of our methods.

commands: move checkformat to bookmarks module commands.bookmark has grown quite large with two closures already. Let's split this up (and in the process allow extensions to override the default behavior).

tests: tell pip not to check for a newer version pip will check to see if it's the latest version, and complain if it isn't. The --no-index flag implies the --disable-pip-version-check flag, and makes the warning (and any associated network activity) go away.

changeset_templater: backport parts map of [templates] section from formatter

formatter: add support for parts map of [templates] section Unlike a mapfile whose template is looked up by spec -> mapfile -> topic, [templates] section is global. We use :sub-section syntax to define parts per template.

changeset_templater: backport separator template from formatter As commented, this should be used with docheader and docfooter, not with header nor footer. That's one reason why no props are passed to templater when rendering a separator. (See map-cmdline.changelog to understand what the "header" is.)

formatter: add support for separator template This seems useful for writing JSON template.

formatter: add support for docheader and docfooter templates templatepartsmap() is a minimal copy of changeset_templater.__init__(). I tried to factor out a common function, but it was unnecessarily complicated.

formatter: extract helper function to render template

changeset_templater: do not enable verbosity postfix for [templates] section Since this postfix hack exists only for backward compatibility, we don't need it for new [templates] section. This isn't a BC as templates defined in [templates] section weren't loaded until recently.

changeset_templater: simplify handling of verbosity postfix

merge with stable

tests: don't touch the network when using virtualenv

tests: removed ReportedTest exception The only call site called addFailure before raising, which is exactly what the failure exception handler does. So this complexity is not needed. We have test coverage of this "server failed to start" scenario and nothing appeared to change.

tests: remove support for warned tests The previous changeset removed the last caller of addWarn(). So, we rip out that method and all the code related to tracking warned tests in the results system. There was even a comment saying we may want to fold warned tests into the "failed" state, which is what the previous changeset did.

contrib: add check-code rule to enforce non-use of `python` in tests

cleanup: use $PYTHON to run python in many more tests Spotted one of these, then wrote a check-code rule that caught them all. It will be the next change.

tests: enforce use of $PYTHON in tests

tests: use $PYTHON in #! so we always use the right Python

tests: sed away python #! in test-run-tests.t to avoid some upcoming insanity

tests: remove #! from primes.py in test-highlight.t It's about to be a source of trouble, but removing it changes a ton of test lines, so doing this change as a standalone commit.

keyword: use context manager for rollback locking

tests: remove WarnTest We would raise this if a test didn't return a result code. AFAICT this can only occur if there is a logic error in the test harness itself. I don't think it is worth the code complexity to distinguish this failure scenario from a regular test failure.

tests: remove unused IgnoreTest exception AFAICT its last use was removed in d839e4820da7.

tests: use unittest.SkipTest unittest.SkipTest was introduced in Python 2.7. We previously defined it with our own class so we could run on Python 2.6.

changegroup: delete "if True" and reflow

changegroup: let callers pass in transaction to apply() (API) I think passing in the transaction makes it a little clearer and more consistent with bundle2.

repair: create transaction for bundle1 unbundling earlier See earlier patch for motivation.

unbundle: create transaction for bundle1 unbundling earlier See earlier patch for motivation.

exchange: create transaction for bundle1 unbundling earlier changegroup.apply() currently creates a transation if there isn't already one. Having the callers of that method pass in an existing transaction seems a little cleaner. To do that, we need to make sure all callers have a transaction. Since the transaction name is used as a hook argument (HG_TXNNAME), we need to match the name from changegroup.apply().

changegroup: inline 'publishing' variable in apply()

repair: remove unnecessary locking for bookmarks The caller has already locked the repo.

repair: move check for existing transaction earlier Several benefits: * Gets close the comment describing it * Splits off unrelated comment about "backup" argument * Error checking is customarily done early * If we added an early return to the method, it would still consistently fail if there was an existing transaction (so we would find and fix that case quickly) One test needs updating with for this change, because we no longer create the backup bundle before we fail. I don't see much reason to create that backup bundle. If some command was adding content and then trying to strip it as well within the transaction, we would have a backup for the user, but the risk of that not being discovered in development seems very small.

strip: remove unncessary "del" and inline variable

repair: clarify in comment that caller must take lock, but not transaction I have checked that all callers have already taken the lock (and if they hadn't, we should have seen tests fail thanks to the 'transaction requires locking' devel warning in localrepo.transaction()).

amend: use context manager for locking

strip: use context manager for locking and transaction in stripcmd()

strip: use context manager for locking in strip()

rebase: use context manager for locking in pullrebase()

rebase: use context manager for locking in rebase()

test-http-proxy: redirect proxy stdout to /dev/null This output hasn't been getting flushed, but would alter the log if it ever grew large enough. See 23b07333a8b2.

test-logtoprocess: don't run on Windows Several bits of output were missing[1], unless the DETACHED_PROCESS flag is _not_ passed to subprocess.Popen(). The problem with that is it briefly opens and closes several cmd.exe windows on screen. Foozy also mentioned some other issues in that thread. With this, the last of the long standing Windows failures fixed, the test suite now runs cleanly (536 ran, 67 skipped) on Windows 7 x64, with python 2.7.13. \o/ [1] https://www.mercurial-scm.org/pipermail/mercurial-devel/2017-April/096987.html

revset: add support of keyword arguments to ancestors() and descendants() Prepares for adding depth parameter.

revsetlang: check arguments passed to ancestors() before optimizing to only() Future patches will add depth parameter to ancestors(), which isn't compatible with only().

revsetlang: factor out helper to match ancestors() in parsed tree More checks will be added.

templatekw: use common "rev:node" format as the default of predecessors I'm not sure if this is better. If we're planning to add a template keyword that returns obsoleted nodes unavailable in the repo (i.e. they have no valid revision numbers), we might want to use the current "node"-only format everywhere.

templatekw: populate all keywords depending on predecessor in map operation This is what showparents() does. repo[precnode] should never fail since its validity is tested by closestpredecessors().

templatekw: reference predecessor node id as {node} in map operation More predecessor-depending values will be populated by the next patch.

highlight: put pygments import inside demandimport.deactivated I tripped on some weirdness relating to _thread vs threading way down in a dep of highlight recently. I'm not really sure why I'm only just seeing this defect now, but experimentally this fixes the problem, and shouldn't cause any load-time slowness for people until pygments is actually about to be used since highlight.highlight is still lazily loaded in the highlight/__init__.py file.

run-tests: explicitly flush test runner output for Windows stability When hghave testing goes awry, the output order was changing on Windows. diff --git a/tests/test-run-tests.t b/tests/test-run-tests.t --- a/tests/test-run-tests.t +++ b/tests/test-run-tests.t @@ -920,10 +920,10 @@ > EOF > done $ rt -j 2 - .... + ....skipped: unknown feature: notarealhghavefeature\r (esc) + + # Ran 5 tests, 0 skipped, 0 warned, 0 failed. - skipped: unknown feature: notarealhghavefeature - $ cd .. $ rm -rf broken Since 'skipped: unknown feature: notarealhghavefeature\n\n' is printed to stdout and the rest to stderr, it seems like maybe stdio isn't line buffered on Windows. When a program exits, stdout is flushed before stderr[1]. [1] https://blogs.msdn.microsoft.com/oldnewthing/20060519-09/?p=31133

tinyproxy: explicitly flush logged messages On Windows, output streams are buffered when redirected to a file, and TerminateProcess() apparently doesn't trigger a flush. This left test-http-proxy.t missing part of the last line when it cat'd proxy.log[1]. Flushing stderr is all that is needed (on py27 anyway). I originally flushed stdout too, but that added additional output to the log: $ cat proxy.log + Accept: $LOCALIP (localhost)\r (esc) + Serving HTTP on 0.0.0.0 port 20810 ...\r (esc) + connect to localhost:$HGPORT\r (esc) * - - [*] "GET http://localhost:$HGPORT/?cmd=capabilities HTTP/1.1" - - (glob) + bye\r (esc) + connect to localhost:$HGPORT\r (esc) * - - [*] "GET http://localhost:$HGPORT/?cmd=branchmap HTTP/1.1" - - x-hgproto-1:0.1 0.2 comp=*zlib,none,bzip2 (glob) + bye\r (esc) + connect to localhost:$HGPORT\r (esc) * - - [*] "GET http://localhost:$HGPORT/?cmd=stream_out HTTP/1.1" - - x-hgproto-1:0.1 0.2 comp=*zlib,none,bzip2 (glob) + bye\r (esc) + connect to localhost:$HGPORT\r (esc) ... [1] https://www.mercurial-scm.org/pipermail/mercurial-devel/2017-April/096987.html

help: clarify ancestors() and descendants() include given set (issue5594) Also unified "a changeset" to "changesets".

dagop: move blockancestors() and blockdescendants() from context context.py seems not a good place to host these functions. % wc -l mercurial/context.py mercurial/dagop.py 2306 mercurial/context.py 424 mercurial/dagop.py 2730 total

dagop: split module hosting DAG-related algorithms from revset This module hosts the following functions. They are somewhat similar (e.g. scanning revisions using heap queue or stack) and seem non-trivial in algorithmic point of view. - _revancestors() - _revdescendants() - reachableroots() - _toposort() I was thinking of adding revset._fileancestors() generator for better follow() implementation, but it would be called from context.py as well. So I decided to create new module. Naming is hard. I couldn't come up with any better module name, so it's called "dag operation" now. I rejected the following candidates: - ancestor.py - existing, revlog-level DAG algorithm - ancestorset.py - doesn't always return a set - dagalgorithm.py - hard to type - dagutil.py - existing - revancestor.py - I want to add fileancestors() % wc -l mercurial/dagop.py mercurial/revset.py 339 mercurial/dagop.py 2020 mercurial/revset.py 2359 total

tests: protect tests involving git ext::sh with git-ext-sh

hghave: add has_git_range for testing if git understands ext::sh Even on CentOS 7, git is at version 1.8. It seems git 1.9 is when ext::sh was introduced so we a check for that. The way these functions are written follows the same style and format for the way we check svn and bzr versions.

rebase: rewrite "x in y.children()" as "y in x.parents()" children() is slow

shelve: rewrite "x in y.children()" as "y in x.parents()" children() is slow

py3: use open() instead of file() constructor The file() constructor has been removed in python 3.

py3: convert keys of kwargs back to bytes using pycompat.byteskwargs()

py3: use pycompat.strkwargs() to convert kwargs keys to str before passing

py3: use '%d' instead of '%s' for integers Python 3 does not let you use '%s' for integers.

py3: pass the path in hg.repository() as bytes This make test-ancestor.py pass on Python 3.

py3: pass range() into list() to get one explicitly range() on python 3 returns a generator whereas on python 2 returns a list. So to get a list on python 3, we passed it into list()

check-concurrency: expose the feature as 'concurrent-push-mode' We move the feature to a proper configuration and document it. The config goes in the 'server' section because it feels like something the server owner would want to decide. We pick and open field because it seems likely that other checking levels will emerge in the future. (eg: server like the mozilla-try server will likely wants a "none" value) The option name contains 'push' since this affects 'push' only. The option value 'check-related' is preferred over one explicitly containing 'allow' or 'deny' because the client still have a strong decision power here. Here, the server is just advising the client on the check mode to use.

exchange: switch to usual way of testing for bundle2-ness We used safehasattr() in one place, but we use isinstance() for this everywhere else, so switch to the latter.

setup: update a comment that blamed py26 for a Windows workaround

i18n: drop a py25 conditional I'm not sure how to test this one. `make update-pot` spews all kinds of warnings, though it did before this change too.

perf: ensure HGRCPATH is exported on Windows After dropping the garbage collector hack, `hg perfstartup` started yelling about not being able to import the evolve extension, which I have in my user config. Launching `env` shows that an empty HGRCPATH isn't exported to the environment. Since `env` doesn't quote, I have no idea if the variable is trimmed, but Mercurial doesn't complain when processing it.

test-status: glob fixes for Windows

plan9: drop py26 hacks

revset: fix negative ancestor spec to not return changectx objects The order was unstable because of this.

obsolete: use ProgrammingError over assert for volatile set registration We have ProgrammingError now.

strip: remove a redundant setting of hookargs bundle2.applybundle() will set both 'source' and 'url'.

unbundle: move streamclone error case away from bundle[12] cases The bundle1 and bundle2 cases are more similar, and having the streamclone case in between is distracting.

bundle2: use "else" instead of checking condition again

wireproto: update reference to deleted addchangegroup() Thanks to Yuya for catching this.

template: add predecessors template Add a 'predecessors' template that returns the list of all closest known predecessors for a changectx. The elements of the list are row changectx node id formatted by default as short nodes. The "closest predecessors" are the first locally known revisions encountered while, walking predecessors markers. For example: 1) If a (A, (B)) markers exists and both A and B are locally known A is a closest predecessors of B. 2) If a (A, (B)) and (B, (C)) markers exists and only A and C are known locally, A will be the closest precursors of C. This logic respect repository filtering. So hidden revision will be skipped by this logic unless --hidden is specified. Since we only display the visible predecessors, this template will not display anything in most case. It makes a good candidate for inclusion in the default log output. I added a new test-file for testing the precursors in various scenarios. This test file will also be used for the successors template. A new "obsutil" module has been added to start gathering utility function outside of the large obsolete.py module.

commit: select template by spec.ref name And load all templates defined in [committemplate] since the selected template is no longer be named as 'changeset'.

commit: do not look up committemplate in template paths (BC) From 5375ba75df40 and de5cee8ba088, I don't think the current behavior is intended. Commit templates should be processed as literal templates.

formatter: always store a literal template unnamed Now spec.ref should be '' if spec.tmpl is specified. Since spec.ref is the option to select the initial template to be rendered, it doesn't make sense to store the given literal template as spec.ref.

formatter: load templates section like a map file Since a map file has another level to select a template (spec -> mapfile -> topic), this isn't exactly the same as how a map file works. But I believe most users would expect the new behavior. A literal template is stored as an unnamed template so that it will never conflict with the templates defined in [templates] section.

test-command-template: rewrite recursion tests by using a map file A literal template will be unnamed soon, so no recursion will be practically made by using -T option.

templater: add simple interface for unnamed template (API) This provides a simpler API for callers which don't need full templating stack. Instead of storing the given template as the name specified by topic, use '' as the default template to be rendered.

ui: add support for a tweakdefaults knob We've been talking for years about a one-stop config knob to opt in to better behavior. There have been a lot of ideas thrown around, but they all seem to be too complicated to get anyone to actually do the work.. As such, this patch is the stupidest thing that can possibly work in the name of getting a good feature to users. Right now it's just three config settings that I think are generally uncontroversial, but I expect to add more soon. That will likely include adding new config knobs for the express purpose of adding them to tweakdefaults.

localrepo: remove unused addchangegroup() (API) This completes the cleanup started in d3775db748a0 (localrepo: move the addchangegroup method in changegroup module, 2014-04-01).

changegroup: rename "dh" to the clearer "deltaheads" We have a lot of frequently used abbreviations, but this is not one of them.

changegroup: rename "srccontent" to "cgnodes" It's the list of nodes in the incoming changegroup, so "cgnodes" made more sense to me.

revlog: rename list of nodes from "content" to "nodes" It seems like the reason for "content" is that the variable contains the nodes that the changegroup "contains", see e234eda20984 (revlog: make addgroup returns a list of node contained in the added source, 2012-01-13), but "nodes" seems much clearer.

revlog: delete obsolete comment The comment seems to refer to code that was deleted in 557988c691d1 (revlog.addgroup(): always use _addrevision() to add new revlog entries, 2010-10-08).

revlog: delete dead assignment in addgroup()