Before this patch, the shell alias causes failure when it takes its
specific (= unknown for "hg") options in the command line, because
"_parse()" can't accept them.
This is the regression introduced by 7849ac1dbc57.
It fixed the issue that ambiguity between shell aliases and commands
defined by extensions was ignored. But it also caused that ambiguous
shell alias is handled in "_parse()" even if it takes specific options
in the command line.
To avoid such failure, this patch checks shell alias again after
loading extensions.
All aliases and commands (including ones defined by extensions) are
completely defined before the 2nd (= newly added in this patch)
"_checkshellalias()" invocation, and "cmdutil.findcmd(strict=False)"
can detect ambiguity between them correctly.
For efficiency, this patch does:
- omit the 2nd "_checkshellalias()" invocation if "[ui] strict= True"
it causes "cmdutil.findcmd(strict=True)", of which result should
be equal to one of the 1st invocation before adding aliases
- avoid removing the 1st "_checkshellalias()" invocation
it causes "cmdutil.findcmd(strict=True)" invocation preventing
shell alias execution from loading extensions uselessly
To reduce changes in the subsequent patch fixing issue4355, this patch
makes "_checkshellalias" reusable regardless of adding aliases.
In this patch, alias definitions are added and restored, only when
"precheck=True".
Old behavior:
hg help x hg x -h hg help -e x hg help -c x
config topic topic (!) - cmd
showconfig cmd topic (!) - cmd
rebase cmd cmd ext cmd
New behavior:
hg help x hg x -h hg help -e x hg help -c x
config topic cmd - cmd
showconfig cmd cmd - cmd
rebase cmd cmd ext cmd
Added "unexpected leading whitespace" message to parse error
when .hgrc has a line that starts with whitespace.
Helps new users unfamiliar with syntax of rc file.
Before this patch, there was no way to pass in all the positional parameters as
separate words down to another command.
(1) $@ (without quotes) would expand to all the parameters separated by a space.
This would work fine for arguments without spaces, but arguments with spaces
in them would be split up by POSIX shells into separate words.
(2) '$@' (in single quotes) would expand to all the parameters within a pair of
single quotes. POSIX shells would then treat the entire list of arguments
as one word.
(3) "$@" (in double quotes) would expand similarly to (2).
With this patch, we expand "$@" (in double quotes) as all positional
parameters, quoted individually with util.shellquote, and separated by spaces.
Under standard field-splitting conditions, POSIX shells will tokenize each
argument into exactly one word.
This is a backwards-incompatible change, but the old behavior was arguably a
bug: Bourne-derived shells have expanded "$@" as a tokenized list of positional
parameters for a very long time. I could find this behavior specified in IEEE
Std 1003.1-2001, and this probably goes back to much further before that.
No command should fail with ValueError just because there is unparseable
alias definition.
It returns 1 like other badalias handlers, but should be changed to 255 in
a later version because we use 255 for general command error.
This also includes test for shell aliases. It avoid using "false" command
because "man false" does not say "exit with 1" but "exit with a status code
indicating failure."
When having ui.debugger=somedebugger in one's ~/.hgrc, this then
somedebugger would be imported for every hg command. With this patch,
this import only happens if the --debugger parameter is passed.
Before this patch, shell alias may be executed by abbreviated command
name unexpectedly, even if abbreviated command name matches also
against the command provided by extension.
For example, "rebate" shell alias is executed by "hg reba", even if
rebase extension (= "rebase" command) is enabled. In this case, "hg
reba" should be aborted because of command name ambiguity.
This patch makes "_checkshellalias()" invoke "cmdutil.findcmd()"
always with "strict=True" (default value).
If abbreviated command name matches against only one shell alias even
after loading extensions, such shell alias will be executed via
"_parse()".
This patch doesn't remove "_checkshellalias()" invocation itself,
because it may prevent shell alias from loading extensions uselessly.
Some debuggers, such as ipdb, load escape codes and color codes even when later
turned off. This will affect scripts that do simple parsing and can't handle
escape codes. Therefore, we only load a custom debugger if ui.plain() is false.
This adds the ability to specify a config option, ui.debugger, to a custom pdb
module, such as ipdb, and have mercurial use that as its debugger. As long as
the value of ui.debugger is a loadable module with the set_trace and
post_mortem functions, then dispatch will be able to use the custom module.
Debugging _parseconfig is still available in the case of an error since it will
be caught with a default the value of pdb.post_mortem.
Previously, command line parsing of --config arguments was done in
_dispatch. This means that it takes place after activating the debugger. In an
upcoming patch, we will add a ui.debugger setting so we need to have this
parsing done before _runcatch.
Previously the blackbox wrapped runcommand, but this failed to see the error
codes that were created if an exception occurred. I moved that logging to now
wrap _runcatch, so it can observe and log the actual error code (such as when
a user ctrl+c's during a command).
Updated the tests as well. Tested the change by running all the tests with the
blackbox extension enabled and verifying nothing broke (aside from things that
printed what extensions were enabeld).
The progress tests are affected by calls to time.time() so they needed to be
updated to pass.
When running commands like 'hg export -o mypatch-%N.patch', the blackbox
would throw an exception because it tried to format %N. This change
prevents it from trying to format the command string.
As mentioned in bug 2043, --config is also not supported in an alias. So report
this the same way as the other "early" options.
Example with alias.broken = stat --config a.config=1
Before:
$ hg broken
abort: Option --config may not be abbreviated!
After:
$ hg broken
error in definition for alias 'broken': --config may only be given on the command line
User 'timeless' in irc mentioned that having the blackbox be
translated would result in logs that:
- may be mixed language, if multiple users use the same repo
- are not google searchable (since searching for english gives more
results)
- might not be readable by an admin if the employee is using hg in
his native language
And therefore we should log everything in english.
Uses ui.log to log which commands are run, their exit code, the time taken,
and any unhandled exceptions thrown.
Example log lines:
2013/02/09 08:35:19 durham> add foo
2013/02/09 08:35:19 durham> add exited 0 after 0.02 seconds
Updates the progress tests because they use a mocked time.time() which these
changes affect.
The number of output lines was hardcoded to 30.
There was a 'nested' configuration options that controlled something else
related to counting the number of output lines.
This introduces the profiling.limit configuration option for controlling the
number of profiling output to show.
We ensure all repositores created through `mercurial.hg.repository`
are "hidden" filtered. This is an even stronger enforcement than
86530c899687.
Citing Matt's response to changeset 86530c899687 installing filtering
in dispatch:
> Unfortunately, this means that code that doesn't go through dispatch (ie all
> those crazy misguided people using Mercurial as a library) are going to see
> these hidden changesets.
>
> Might be better to instead install the filter in localrepo construction by
> default and disable it in dispatch.
The dispatch code now enables filtering of "hidden" changesets globally. The
filter is installed before command and extension invocation. The `--hidden`
switch is now global and disables this filtering for any command.
Code in log dedicated to changeset exclusion is removed as this global filtering
has the same effect.
Mercurial would sometimes exit with:
abort: No such file or directory
where str of the actual OSError exception was the more helpful:
[Errno 2] No such file or directory: ''
The exception will now always show the filename and quote it:
abort: No such file or directory: ''
When extensions had an empty `testedwith` attribute the code tried to parse it
and failed. As a result the actual error were shallowed by a This crash.
We now treat empty strip as 'unknown'
Maintain a whitelist of commands to infer the repo for instead. The whitelist
contains those commands that take file(s) in the working dir as arguments.
The "worst" extension still is the one tested with the lowest tested version
below the current version of Mercurial, but if an extension with was only
tested with newer versions, it is considered a candidate for a bad extension,
too. In this case extensions which have been tested with higher versions of
Mercurial are considered better. This allows finding the oldest extension if
ct can't be calculated correctly and therefore defaults to an empty tuple, and
it involves less changes to the comparison logic during the current code
freeze.
When developing, we may see non-standard version strings of the form
5d64306f39bb+20120525
which caused tuplever() to raise
ValueError: invalid literal for int() with base 10: '5d64306f39bb'
and shadowing the real traceback.
Extension authors should explicitly declare their supported hg
versions and include a buglink attribute in their extension. In the
event that a traceback occurs, we'll identify the
least-recently-tested extensionas the most likely source of the defect
and suggest the user disable that extension.
Packagers should make every effort to ship hg versions from exact
tags, or with as few modifications as possible so that the versioning
can work appropriately.
Commands working without a repository, like "init", are listed in
commands.norepo. Commands optionally using a repository, like "showconfig", are
listed in commands.optionalrepo. Command aliases were inheriting the former but
not the latter.
This can be selected using the config variable profiling.type or
the environment variable HGPROF ("ls" for the default, "stat" for
statprof). The only tuneable is the frequency, profiling.freq,
which defaults to 1000 Hz.
If statprof is not available, a warning is printed.
The exit code returned from a program to the shell is unsigned 8-bit, but
Mercurial would sometimes try to exit with negative numbers or None. sys.exit
on Unix will convert that to 8-bit exit codes, but on Windows negative values
showed up as 0.
The exit code is now explicitly converted to unsigned 8-bit.
Previously, if you set an alias for "ci", it'd also shadow "commit"
even though you didn't specify that. This occurred for all commands
with explicit short variations.
Previously aliases that overrode existing commands would wrap the old alias
on every call to dispatch() (twice actually), which is an obvious re-entrancy
issue for things like the command server or TortoiseHG.
Older clients will still print the provided error message and not much else:
over ssh, this will be each line prefixed with 'remote: ' in addition to an
"abort: unexpected response: '\n'"; over http, this will be the '---%<---'
banners in addition to the 'does not appear to be a repository' message.
Currently, clients with this patch will display 'abort: remote error:\n' and
the provided error text, but it is trivial to style the error text however is
deemed appropriate.
Closing here means we've closed the repo passed to us in the request,
which is not our responsibility.
This is essential for bundlerepo, and possibly other localrepository
subclasses who do something in their close().
and check if we got one before creating.
note that the contents of the ui object might change after
dispatch() returns (by options passed through --config for example),
to ensure it doesn't, pass a copy() of it.
Example
$ hg clone --jump foo bar
hg clone: option --jump not recognized
hg clone [OPTION]... SOURCE [DEST]
make a copy of an existing repository
options:
-U --noupdate the clone will include an empty working copy (only a
repository)
-u --updaterev REV revision, tag or branch to check out
-r --rev REV [+] include the specified changeset
-b --branch BRANCH [+] clone only the specified branch
--pull use pull protocol to copy metadata
--uncompressed use uncompressed transfer (fast over LAN)
-e --ssh CMD specify ssh command to use
--remotecmd CMD specify hg command to run on the remote side
--insecure do not verify server certificate (ignoring
web.cacerts config)
[+] marked option can be specified multiple times
use "hg help clone" to show the full help text
Motivation for this change
If the user already has specified the command, he probably already knows
the command to some extent. Apparently, he has a problem with the options,
so we show him just the synopsis with the short help and the details about
the options, with a hint on the last line how to get the full help text.
Why is Mercurial better with this change?
Experts who just forgot about the details of an option don't get that
much text thrown at them, while the newbies still get a hint on the last
line how to get the full help text.
This improves the misleading error message
$ hg identify
abort: there is no Mercurial repository here (.hg not found)!
to the more explicit
$ hg identify
abort: requirement 'fake' not supported!
for all commands in commands.optionalrepo, which includes the identify
and serve commands in particular.
This is for the case when a new entry in .hg/requires will be defined
in a future Mercurial release.
Add empty repository.close() and call it in dispatch.
Remove bundlerepository.__del__(), merging it into bundlerepository.close(),
which overrides repository.close().
http://docs.python.org/reference/datamodel.html says:
"It is not guaranteed that __del__() methods are called for objects that
still exist when the interpreter exits."
If --insecure specified, it behaves in the same way as no web.cacerts
configured.
Also shows hint for --insecure option when _verifycert() failed. But currently
the hint isn't displayed on SSLError, because it needs a certain level of
changes.
This patch modifies the check for shell aliases to prevent crashing when an invalid
global option is given.
When an invalid global option is given the check will simply return and let the
normal error handling for this case happen.
This patch refactors the dispatch code to change how arguments to shell aliases
are handled.
A separate "pass" to determine whether a command is a shell alias has been
added. The rough steps dispatch now performs when a command is given are these:
* Parse all arguments up to the command name.
* If any arguments such as --repository or --cwd are given (which could change
the config file used, and therefore the definition of aliases), they are
taken into account.
* We determine whether the command is a shell alias.
* If so, execute the alias. The --repo and --cwd arguments are still in effect.
Any arguments *after* the command name are passed unchanged through to the
shell command (and interpolated as normal.
* If the command is *not* a shell alias, the dispatching is effectively "reset"
and reparsed as normal in its entirety.
The net effect of this patch is to make shell alias commands behave as you
would expect.
Any arguments you give to a shell alias *after* the alias name are passed
through unchanged. This lets you do something like the following:
[alias]
filereleased = !$HG log -r 'descendants(adds("$1")) and tagged()' -l1 $2 $3 $4 $5
$ hg filereleased hgext/bookmarks.py --style compact
Previously the `--style compact` part would fail because Mercurial would
interpret those arguments as arguments to the alias command itself (which
doesn't take any arguments).
Also: running something like `hg -R ~/src/hg-crew filereleased
hgext/bookmarks.py` when `filereleased` is only defined in that repo's config
will now work.
These global arguments can *only* be given to a shell alias *before* the alias
name. For example, this will *not* work in the above situation:
$ hg filereleased -R ~/src/hg-crew hgext/bookmarks.py
The reason for this is that you may want to pass arguments like --repository to
the alias (or, more likely, their short versions like -R):
[alias]
own = !chown $@ `$HG root`
$ hg own steve
$ hg own -R steve
Currently, given an alias like the following:
[alias]
summary = summary --remote
The alias might be executed - or it might not - depending on the order
of the cmdtable dict.
This happens because cmdalias gets assigned back to the cmdtable like so:
cmdtable['summary'] = ...
Yet '^summary|sum' is still in the table, so which one cmdutil.findcmd()
chooses isn't deterministic.
This patch makes cmdalias assign back to '^summary|sum'. It uses the same
cmdtable key lookup that extensions.wrapcommand() does.
This patch changes the functionality of shell aliases to add more powerful
options for working with shell alias arguments.
First: the alias name + arguments to a shell alias are set as an HG_ARGS
environment variable, delimited by spaces. This matches the behavior of hooks.
Second: any occurrences of "$@" (without quotes) are replaced with the
arguments, separated by spaces. This happens *before* the alias gets to the shell.
Third: any positive numeric variables ("$1", "$2", etc) are replaced with the
appropriate argument, indexed from 1. "$0" is replaced with the name of the
alias. Any "extra" numeric variables are replaced with an empty string. This
happens *before* the alias gets to the shell.
These changes allow for more flexible shell aliases:
[alias]
echo = !echo $@
count = !hg log -r "$@" --template='.' | wc -c | sed -e 's/ //g'
qqueuemv = !mv "`hg root`/.hg/patches-$1" "`hg root`/.hg/patches-$2"
In action:
$ hg echo foo
foo
$ hg count 'branch(default)'
901
$ hg count 'branch(stable) and keyword(fixes)'
102
$ hg qqueuemv myfeature somefeature
The logic pre-emptively checks for -R, --repo, --repository and --cwd
in order to give the user a more helpful error message. In addition,
each option is handled invididually, which avoids listing them all in
the error.
Before:
% hg --config alias.broken='push --cwd /dev/null' broken
abort: Option --cwd may not be abbreviated!
After:
% hg --config alias.broken='push --cwd /dev/null' broken
error in definition for alias 'broken': --cwd may only be given on the command line
Aliased commands that received bad arguments would raise TypeError instead of
SignatureError. This only affected commands that weren't wrapped by extensions.
Using util.checksignature() in cmdalias.__call__() ensures SignatureError is
raised correctly.
Previous behavior wasn't very helpful:
$ hg st foo
abort: No such file or directory
Now we tell more about what failed:
abort: error getting current working directory: No such file or directory
This patch adds git-style "shell aliases" to Mercurial.
Any alias with a definition beginning with a '!' will be treated as a shell
alias. For example:
[alias]
echo = !echo
qempty = !hg qrefresh -X "`hg root`" ; echo Emptied patch "`hg qtop`"
$ hg echo foo
foo
$ hg qempty
Emptied patch foo
$
This resolves the issue of hg cmd --mq not being colorized. This was due
to color wrapping only the instance of ui passed to dispatch._runcommand(),
which isn't the same ui object that mq.mqcommand() receives. After dispatch
calls extensions.loadall(), it makes sure any changes to ui.__class__ in
uisetup are propagated.
progress is updated to wrap ui in the same manner because wrapfunction
doesn't play well when ui.__class__ has been replaced by another extension
(orig will point to the old class method instead of color's).
python hooks are passed two new keyword arguments:
- opts: a dict of options; unsepcified options are set to their default
- pats: a list of arguments
shell hooks receive two new variables containing string representations
of the above data:
- $HG_OPTS
- $HG_PATS
for example, the opts and pats for 'hg -f v1.1' would be:
{'force': True, 'message': '', 'rev': '', 'user': '', 'date': '', 'local': None, 'remove': None, 'mq': None}
['v1.1']
Previously, Mercurial assumed that the last word of the string
representation was the name of the moduled that was imported. This
assmption is incorrect, despite being true for the common case of an
exception raised by the Python VM.
For example, hgsubversion raises an ImportError with a helpful message
if the Subversion bindings were not found. The final word of this
message is not meaningful on its own, and is never the name of a
module.
This patch changes the output printed to be a simple stringification
of the exception instance. In most cases, this will be `abort: No
module named X!' rather than `abort: could not import module X!'.
No functionality change; all tests pass.
__doc__ of aliased command shouldn't cointain non-ASCII characters,
because it'll be gettext-ed later by commands.help_().
Here gettext can raise UnicodeDecodeError.
Once concatenated two translatable strings into one, it become untranslatable.
So this patch moves 'alias for:' from dispatch.cmdalias to commands.help_,
where help texts are translated.
'alias for:' was introduced by 027d5c280eda.
Before a command is declared unknown, each extension in hgext is searched,
starting with hgext.<cmdname>. If there's a matching command, a help message
suggests the appropriate extension and how to enable it.
Every extension could potentially be imported, but for cases like rebase,
relink, etc. only one extension is imported.
For the case of "hg help disabledext", if the extension is in hgext, the
extension description is read and a similar help suggestion is printed.
No extension import occurs.
previously uisetup() was invoked by extensions.loadall(), but
extsetup() was by _dispatch().
there's no need to split them because we have nothing to do
between uisetup() and extsetup().
this fixes issue1824 indirectly.
Extensions are now loaded with a call-graph like this:
dispatch._dispatch
extensions.loadall
extensions.load
# add foo module to extensions._extensions
extensions.load
# add bar module to extensions._extensions
foo.uisetup(ui)
bar.uisetup(ui)
foo.extsetup()
bar.extsetup()
commands.table.update(foo.cmdtable)
commands.table.update(bar.cmdtable)
hg.repository
foo.reposetup(ui, repo)
bar.reposetup(ui, repo)
The uisetup calls could easily be moved out to dispatch._dispatch, but
have been kept in extensions.loadall since at least TortoiseHg calls
extensions.loadall and expects it to call uisetup.
The extensions.load function called uisetup. It now has an unused ui
argument which has been kept for backwards compatibility.
Allows defining other output formats for profiling.
If an invalid format is given, output a warning and ignore it.
For now, only the standard 'text' value is supported.
hotshot was an experimental module, which is broken for Python < 2.5
And even for Python >= 2.5 users, hotshot usage is discouraged: cProfile
(formerly lsprof) should be used instead.
This pulls the pre-command hook/command/post-command hook workflow out of
the method it is in and puts it into its own method so that it potentially
could be exposed for extensions to wrap.
This changes the behavior of qguard in the case of setting negative guards, as -- will now always be required.
Fixes issue1402.
Doc fixes for mq by mpm.