graphmod: refactor state handling Move ASCII graph state to a dictionary, to clarify what is being tracked. Move the 'seen' state (tracking currently active edges) into this structure.

templater: move label() function from color extension ui.label() is no-op by default, so we can just call ui.label() by label() template function no matter if the color is enabled or not.

templater: make label() take unknown symbol as color literal Instead of the mapping hack introduced by b775a2029e8d, this patch changes the way how a label symbol is evaluated. This is still hackish, but should be more predictable in that it doesn't depend on the known color effects. This change is intended to eliminate the reference to color._effects so that color.templatelabel() can be merged with templater.label().

convert: monotone use absolute_import

convert: p4 use absolute_import

convert: hg use absolute_import

convert: cvsps use absolute_import

convert: darcs use absolute_import

convert: filemap use absolute_import

convert: gnuarch use absolute_import

convert: git use absolute_import

branchmap: check node against changelog instead of repo Testing 'node in repo' requires constructing a changectx, which is a little expensive. Testing 'repo.changelog.hasnode(node)' is notably faster. This saves 10-20ms off of every command, when testing a few thousand nodes from the branch cache. I considered changing the implementation of localrepository.__contains__ so every place would benefit from the change, but since localrepository.__contains__ uses changectx to check if the commit exists, it means it supports a wider range of possible inputs (like revs, hashes, '.', etc), so it seemed unnecessarily risky.

changegroup: clear progress callback after changelog processing The progress callback is replaced by one for manifests after changelog processing is done, but let's not depend on manifests replacing the value and instead explicitly clear it.

tests: adjust for code move in Hypothesis 2.0.0 It appears that the Settings class was renamed to settings, and because of this, the settings module was renamed to configuration. https://github.com/DRMacIver/hypothesis/commit/a0e663b

changegroup: progress for added files is not measured in "chunks" The "prog" class cg1unpacker.apply() has the unit set to "chunks". This is not correct for files, where the file itself is the unit. The unit is not usually printed, which is probably why this has not been fixed yet. It can be show with e.g. "--config progress.format='topic number unit'".

changegroup: exclude submanifests from manifest progress The progress callback for manifests is cleared outside of _unpackmanifests(), which means it will remain in effect while pulling subdirectory manifests when using treemanifests. Since the total number of revisions used for the progress is the number of changesets, the total number of treemanifest revisions is usually larger than that. One effect of this is that the ETA is negative. It's hard to estimate the number of subdirectory revisions, so let's just exclude them from progress for now.

commit: block amend while histedit is in progress (issue4800)

chg: limit reconnect attempts Some users may have hg as a wrapper script which sets sensitive environment variables (like setting up virtualenv). This will make chg redirect forever because the environment variables are never considered up to date. This patch adds a limit (10) for reconnect attempts and warn the user with a possible solution if the limit is exceeded.

chg: use validate to make sure the server is up to date This patch uses the newly added validate method to make sure the server has loaded the up-to-date config and extensions. If the server cannot validate itself, the client will receive instructions and follow them to try to reach another server that is more likely to validate itself. The instructions can be a redirect (connect to another server address) and/or an unlink (stops an out-dated server).

chg: implement validate in hgclient This patch implements the corresponding validate method in hgclient. It will return instruction strings as is without taking any real action.

hgclient: use absolute_import and print_function

fixpax: use absolute_import and print_function

debugcmdserver: use absolute_import and print_function

check-config: use absolute_import and print_function

casesmash: use absolute_import

chgserver: implement validate command validate will load the repo config and check if the server has up-to-date config to continue serve the client. In case it does not, the server will send instructions to the client about what to do next, including to retry with a different address or to unlink an outdated socket file to stop an old server.

templater: handle exception when applying map operator to non-iterable object Before this, "{noniterable % template}" raised an exception. This tries to provide a better indication for the common case, where a left-hand-side expression is a keyword.

templater: factor out thin helper that evaluates argument as string This function is trivial, but it can make it clear that args[i] is a (func, data) pair.

color: make label() template function stringify "expr" argument first Since label() should return a string (or a thunk to be evaluated to a string), this change is okay. This helps porting to evalstring() helper. See the next patch for details.

templater: fix shortest() to evaluate int argument and handle error

templater: fix pad() to evaluate int argument and handle error

templater: drop redundant type conversion when evaluating integer argument A function argument may be an integer. In this case, it isn't necessary to convert a value to string and back to integer. Because an argument may be an arbitrary object (e.g. date tuple), TypeError should be caught as well.

templater: factor out function that evaluates argument as integer We have more bare int()s that may raise ValueError or TypeError. This function will be used to fix them.

chgserver: use basename for socket symlink Previously we use full path and the symlink may point to outside (unsafe) world if the directory is moved. This patch fixes it by only linking to basename of the target. Therefore the symbolic link and socket files will always stay in the same directory.

patch: when importing from email, RFC2047-decode From/Subject headers Reported at https://bugs.debian.org/737498

histedit: reword message when a changeset produces no changes There are various ways to use histedit such that an item in the list of things to perform will not result in a change relative to the previous repository state. When that happens, histedit does not keep the commit/message. This changes the note to try to explain to the user that it will not be present in their history.

tests: replace cat.py with cat in test-histedit-fold-non-commute.t

tests: Solaris grep doesn't add a trailing newline when it's missing The bad-extension tests emits a list of not-loaded extensions, and pipes that output through grep. On Solaris, the test-output gets "(no-eol)" appended because although the message has no trailing newline, GNU grep adds it. If we simply add the newline to the message, the problem goes away for both versions of grep.

tests: Solaris cmp complains about empty files, even with -s When you compare an empty file, such as /dev/null, with a non-empty file, Solaris cmp complains on stderr with "cmp: EOF on /dev/null", even if the -s argument is present. GNU cmp makes the complaint, but silences it with -s. We can change the pdiff utility to simply redirect stderr to /dev/null so that we don't have to worry about this difference in the test files.

tests: Solaris sed does not support "\n" meaning newline in the RHS of s/// The blackbox test rewrites a copy of test-dispatch.py on the fly, and adds a couple of lines with the s/// command. GNU sed supports the use of the \n escape to represent a newline, but not Solaris sed. Using a literal newline, prefixed by a backslash, works with both versions of the utility.

tests: Solaris cp doesn't support the -T option The treemanifest tests use the -T option to cp in order to ensure that the two directories named on the commandline are treated as peers, rather than the usual behavior when the final argument is a directory. GNU cp has this option, but other implementations may not. Thankfully, there's no pressing reason to use it. We can simply copy the contents of the first directory into the target directory, since we know that the target directory already exists.

templater: make date() use helper function to evaluate argument A date argument should never be a generator, but using evalfuncarg() should be good for consistency.

templater: fix revset() to evaluate format arguments eagerly See the previous patch for why. This patch also removes redundant list() construction from a list.

templater: fix ifcontains() to evaluate items argument eagerly See the previous patch for why. An "items" argument may be a string, a generator, or an arbitrary container object.

templater: fix get() to evaluate arguments eagerly If a key is constructed from a template expression, it may be a generator. In that case, a key have to be stringified. A dictarg should never be a generator, but this patch also changes it to call evalfuncarg() for consistency.

import-checker: report local with stdlib late warning Without this, developers have to figure it out on their own

tests: update test output for test written on stable branch (issue5104) The changed test lines were added in d493d64757eb, which landed on stable. When this changeset merged to the default branch, the test failed because a4692267bc2d (on default but not stable) changed the order of working directory update when performing a share-based clone from pooled storage. The changes in this patch are similar to test changes in a4692267bc2d.

chgserver: pass hashstate and base server address to chgcmdserver In order to detect a hash change from a request handler, chg must know the original hashstate. It also needs the base server address to figure out redirect addresses.

chg: drop progress.assume-tty config It was necessary to go through progress.uisetup() to set up the progressui wrapper. Since the progress extension has got into the core, progress.assume-tty is no longer necessary.

chgserver: mangle server address to include confighash Before this patch, chgserver will use the address provided by the client. The new design is one server per confighash. This patch appends "-$confighash" to the address the client provides. To maintain the compatibility and make sure the client can connect to the server, a symbolic link is created at the original address pointing to the new address. The address is intentionally mangled at the server, instead of being pre- calculated by some other process (eg. a previous server). In this way, we can avoid file system race conditions.

chgserver: update docs Update the docstring to reflect the latest changes

merge with stable

commands: use absolute_import All mercurial.* modules are now using absolute_import \o/

cmdutil: use absolute_import Now that @command doesn't write back into commands when it is being executed during the loading of commands.py itself, we are unblocked from converting cmdutil to absolute_import.

keyword: use absolute_import

pager: use absolute_import

pager: add tests We've never had tests for pager, and I want to start experimenting with ways to clean up the implementation and make it a bit more graceful.

tests: flag Windows specific lines about background closing as optional

run-tests: defer leftover (?) cleanup until after all output is exhausted Previously, after matching a single line, any contiguous subsequent lines ending with (?) would be added to the output and removed from the expected output. This is a problem if the subsequent test output would have matched the consumed (?) line, because it kept the optional line and then added a duplicate without the (?) [1]. Instead, wait until there is nothing more to match before handling the leftovers. [1] https://www.mercurial-scm.org/pipermail/mercurial-devel/2016-February/080197.html

test-run-tests: pad the failure test to preserve the run order Test size seems to dictate the order in which the tests are run, and the next patch will add to test-success.t.