# keyword.py - $Keyword$ expansion for Mercurial

#

# Copyright 2007-2015 Christian Ebert <blacktrash@gmx.net>

#

# This software may be used and distributed according to the terms of the

# GNU General Public License version 2 or any later version.

#

# $Id$

#

# Keyword expansion hack against the grain of a Distributed SCM

#

# There are many good reasons why this is not needed in a distributed

# SCM, still it may be useful in very small projects based on single

# files (like LaTeX packages), that are mostly addressed to an

# audience not running a version control system.

#

# For in-depth discussion refer to

# <https://mercurial-scm.org/wiki/KeywordPlan>.

#

# Keyword expansion is based on Mercurial's changeset template mappings.

#

# Binary files are not touched.

#

# Files to act upon/ignore are specified in the [keyword] section.

# Customized keyword template mappings in the [keywordmaps] section.

#

# Run 'hg help keyword' and 'hg kwdemo' to get info on configuration.

'''expand keywords in tracked files

This extension expands RCS/CVS-like or self-customized $Keywords$ in

tracked text files selected by your configuration.

Keywords are only expanded in local repositories and not stored in the

change history. The mechanism can be regarded as a convenience for the

current user or for archive distribution.

Keywords expand to the changeset data pertaining to the latest change

relative to the working directory parent of each file.

Configuration is done in the [keyword], [keywordset] and [keywordmaps]

sections of hgrc files.

Example::

    [keyword]

    # expand keywords in every python file except those matching "x*"

    **.py =

    x*    = ignore

    [keywordset]

    # prefer svn- over cvs-like default keywordmaps

    svn = True

.. note::

   The more specific you are in your filename patterns the less you

   lose speed in huge repositories.

For [keywordmaps] template mapping and expansion demonstration and

control run :hg:`kwdemo`. See :hg:`help templates` for a list of

available templates and filters.

Three additional date template filters are provided:

:``utcdate``:    "2006/09/18 15:13:13"

:``svnutcdate``: "2006-09-18 15:13:13Z"

:``svnisodate``: "2006-09-18 08:13:13 -700 (Mon, 18 Sep 2006)"

The default template mappings (view with :hg:`kwdemo -d`) can be

replaced with customized keywords and templates. Again, run

:hg:`kwdemo` to control the results of your configuration changes.

Before changing/disabling active keywords, you must run :hg:`kwshrink`

to avoid storing expanded keywords in the change history.

To force expansion after enabling it, or a configuration change, run

:hg:`kwexpand`.

Expansions spanning more than one line and incremental expansions,

like CVS' $Log$, are not supported. A keyword template map "Log =

{desc}" expands to the first line of the changeset description.

'''

from __future__ import absolute_import

import os

import re

import weakref

from mercurial.i18n import _

from mercurial.pycompat import getattr

from mercurial.hgweb import webcommands

from mercurial import (

    cmdutil,

    context,

    dispatch,

    error,

    extensions,

    filelog,

    localrepo,

    logcmdutil,

    match,

    patch,

    pathutil,

    pycompat,

    registrar,

    scmutil,

    templatefilters,

    templateutil,

    util,

)

from mercurial.utils import (

    dateutil,

    stringutil,

)

cmdtable = {}

command = registrar.command(cmdtable)

# Note for extension authors: ONLY specify testedwith = 'ships-with-hg-core' for

# extensions which SHIP WITH MERCURIAL. Non-mainline extensions should

# be specifying the version(s) of Mercurial they are tested with, or

# leave the attribute unspecified.

testedwith = b'ships-with-hg-core'

# hg commands that do not act on keywords

nokwcommands = (

    b'add addremove annotate bundle export grep incoming init log'

    b' outgoing push tip verify convert email glog'

)

# webcommands that do not act on keywords

nokwwebcommands = b'annotate changeset rev filediff diff comparison'

# hg commands that trigger expansion only when writing to working dir,

# not when reading filelog, and unexpand when reading from working dir

restricted = (

    b'merge kwexpand kwshrink record qrecord resolve transplant'

    b' unshelve rebase graft backout histedit fetch'

)

# names of extensions using dorecord

recordextensions = b'record'

colortable = {

    b'kwfiles.enabled': b'green bold',

    b'kwfiles.deleted': b'cyan bold underline',

    b'kwfiles.enabledunknown': b'green',

    b'kwfiles.ignored': b'bold',

    b'kwfiles.ignoredunknown': b'none',

}

templatefilter = registrar.templatefilter()

configtable = {}

configitem = registrar.configitem(configtable)

configitem(

    b'keywordset',

    b'svn',

    default=False,

)

# date like in cvs' $Date

@templatefilter(b'utcdate', intype=templateutil.date)

def utcdate(date):

    """Date. Returns a UTC-date in this format: "2009/08/18 11:00:13"."""

    dateformat = b'%Y/%m/%d %H:%M:%S'

    return dateutil.datestr((date[0], 0), dateformat)

# date like in svn's $Date

@templatefilter(b'svnisodate', intype=templateutil.date)

def svnisodate(date):

    """Date. Returns a date in this format: "2009-08-18 13:00:13

    +0200 (Tue, 18 Aug 2009)".

    """

    return dateutil.datestr(date, b'%Y-%m-%d %H:%M:%S %1%2 (%a, %d %b %Y)')

# date like in svn's $Id

@templatefilter(b'svnutcdate', intype=templateutil.date)

def svnutcdate(date):

    """Date. Returns a UTC-date in this format: "2009-08-18

    11:00:13Z".

    """

    dateformat = b'%Y-%m-%d %H:%M:%SZ'

    return dateutil.datestr((date[0], 0), dateformat)

# make keyword tools accessible

kwtools = {b'hgcmd': b''}

def _defaultkwmaps(ui):

    '''Returns default keywordmaps according to keywordset configuration.'''

    templates = {

        b'Revision': b'{node|short}',

        b'Author': b'{author|user}',

    }

    kwsets = (

        {

            b'Date': b'{date|utcdate}',

            b'RCSfile': b'{file|basename},v',

            b'RCSFile': b'{file|basename},v',  # kept for backwards compatibility

            # with hg-keyword

            b'Source': b'{root}/{file},v',

            b'Id': b'{file|basename},v {node|short} {date|utcdate} {author|user}',

            b'Header': b'{root}/{file},v {node|short} {date|utcdate} {author|user}',

        },

        {

            b'Date': b'{date|svnisodate}',

            b'Id': b'{file|basename},v {node|short} {date|svnutcdate} {author|user}',

            b'LastChangedRevision': b'{node|short}',

            b'LastChangedBy': b'{author|user}',

            b'LastChangedDate': b'{date|svnisodate}',

        },

    )

    templates.update(kwsets[ui.configbool(b'keywordset', b'svn')])

    return templates

def _shrinktext(text, subfunc):

    """Helper for keyword expansion removal in text.

    Depending on subfunc also returns number of substitutions."""

    return subfunc(br'$\1$', text)

def _preselect(wstatus, changed):

    """Retrieves modified and added files from a working directory state

    and returns the subset of each contained in given changed files

    retrieved from a change context."""

    modified = [f for f in wstatus.modified if f in changed]

    added = [f for f in wstatus.added if f in changed]

    return modified, added

class kwtemplater(object):

    """

    Sets up keyword templates, corresponding keyword regex, and

    provides keyword substitution functions.

    """

    def __init__(self, ui, repo, inc, exc):

        self.ui = ui

        self._repo = weakref.ref(repo)

        self.match = match.match(repo.root, b'', [], inc, exc)

        self.restrict = kwtools[b'hgcmd'] in restricted.split()

        self.postcommit = False

        kwmaps = self.ui.configitems(b'keywordmaps')

        if kwmaps:  # override default templates

            self.templates = dict(kwmaps)

        else:

            self.templates = _defaultkwmaps(self.ui)

    @property

    def repo(self):

        return self._repo()

    @util.propertycache

    def escape(self):

        '''Returns bar-separated and escaped keywords.'''

        return b'|'.join(map(stringutil.reescape, self.templates.keys()))

    @util.propertycache

    def rekw(self):

        '''Returns regex for unexpanded keywords.'''

        return re.compile(br'\$(%s)\$' % self.escape)

    @util.propertycache

    def rekwexp(self):

        '''Returns regex for expanded keywords.'''

        return re.compile(br'\$(%s): [^$\n\r]*? \$' % self.escape)

    def substitute(self, data, path, ctx, subfunc):

        '''Replaces keywords in data with expanded template.'''

        def kwsub(mobj):

            kw = mobj.group(1)

            ct = logcmdutil.maketemplater(

                self.ui, self.repo, self.templates[kw]

            )

            self.ui.pushbuffer()

            ct.show(ctx, root=self.repo.root, file=path)

            ekw = templatefilters.firstline(self.ui.popbuffer())

            return b'$%s: %s $' % (kw, ekw)

        return subfunc(kwsub, data)

    def linkctx(self, path, fileid):

        '''Similar to filelog.linkrev, but returns a changectx.'''

        return self.repo.filectx(path, fileid=fileid).changectx()

    def expand(self, path, node, data):

        '''Returns data with keywords expanded.'''

        if (

            not self.restrict

            and self.match(path)

            and not stringutil.binary(data)

        ):

            ctx = self.linkctx(path, node)

            return self.substitute(data, path, ctx, self.rekw.sub)

        return data

    def iskwfile(self, cand, ctx):

        """Returns subset of candidates which are configured for keyword

        expansion but are not symbolic links."""

        return [f for f in cand if self.match(f) and b'l' not in ctx.flags(f)]

    def overwrite(self, ctx, candidates, lookup, expand, rekw=False):

        '''Overwrites selected files expanding/shrinking keywords.'''

        if self.restrict or lookup or self.postcommit:  # exclude kw_copy

            candidates = self.iskwfile(candidates, ctx)

        if not candidates:

            return

        kwcmd = self.restrict and lookup  # kwexpand/kwshrink

        if self.restrict or expand and lookup:

            mf = ctx.manifest()

        if self.restrict or rekw:

            re_kw = self.rekw

        else:

            re_kw = self.rekwexp

        if expand:

            msg = _(b'overwriting %s expanding keywords\n')

        else:

            msg = _(b'overwriting %s shrinking keywords\n')

        for f in candidates:

            if self.restrict:

                data = self.repo.file(f).read(mf[f])

            else:

                data = self.repo.wread(f)

            if stringutil.binary(data):

                continue

            if expand:

                parents = ctx.parents()

                if lookup:

                    ctx = self.linkctx(f, mf[f])

                elif self.restrict and len(parents) > 1:

                    # merge commit

                    # in case of conflict f is in modified state during

                    # merge, even if f does not differ from f in parent

                    for p in parents:

                        if f in p and not p[f].cmp(ctx[f]):

                            ctx = p[f].changectx()

                            break

                data, found = self.substitute(data, f, ctx, re_kw.subn)

            elif self.restrict:

                found = re_kw.search(data)

            else:

                data, found = _shrinktext(data, re_kw.subn)

            if found:

                self.ui.note(msg % f)

                fp = self.repo.wvfs(f, b"wb", atomictemp=True)

                fp.write(data)

                fp.close()

                if kwcmd:

                    self.repo.dirstate.normal(f)

                elif self.postcommit:

                    self.repo.dirstate.normallookup(f)

    def shrink(self, fname, text):

        '''Returns text with all keyword substitutions removed.'''

        if self.match(fname) and not stringutil.binary(text):

            return _shrinktext(text, self.rekwexp.sub)

        return text

    def shrinklines(self, fname, lines):

        '''Returns lines with keyword substitutions removed.'''

        if self.match(fname):

            text = b''.join(lines)

            if not stringutil.binary(text):

                return _shrinktext(text, self.rekwexp.sub).splitlines(True)

        return lines

    def wread(self, fname, data):

        """If in restricted mode returns data read from wdir with

        keyword substitutions removed."""

        if self.restrict:

            return self.shrink(fname, data)

        return data

class kwfilelog(filelog.filelog):

    """

    Subclass of filelog to hook into its read, add, cmp methods.

    Keywords are "stored" unexpanded, and processed on reading.

    """

    def __init__(self, opener, kwt, path):

        super(kwfilelog, self).__init__(opener, path)