# keyword.py - keyword expansion for Mercurial

#

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

#

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

# of the GNU General Public License, incorporated herein by reference.

#

# $Id$

#

# Keyword expansion hack against the grain of a DSCM

#

# 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

# <http://www.selenic.com/mercurial/wiki/index.cgi/KeywordPlan>.

#

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

# The extension provides an additional UTC-date filter ({date|utcdate}).

#

# The user has the choice either to create his own keywords and their

# expansions or to use the CVS-like default ones.

#

# Expansions spanning more than one line are truncated to their first line.

# Incremental expansion (like CVS' $Log$) is not supported.

#

# Binary files are not touched.

#

# Setup in hgrc:

#

#     # enable extension

#     keyword = /full/path/to/keyword.py

#     # or, if script in hgext folder:

#     # hgext.keyword =

'''keyword expansion in local repositories

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

the text files selected by your configuration.

Keywords are only expanded in local repositories and not logged by

Mercurial internally. The mechanism can be regarded as a convenience

for the current user and may be turned off anytime.

Caveat: "hg import" might fail if the patches were exported from a

repo with a different/no keyword setup, whereas "hg unbundle" is

safe.

Configuration is done in the [keyword] and [keywordmaps] sections of

hgrc files.

Example:

     [keyword]

     # filename patterns for expansion are configured in this section

     # files matching patterns with value 'ignore' are ignored

     **.py =          ## expand keywords in all python files

     x* = ignore      ## but ignore files matching "x*"

     ...

     [keywordmaps]

     # custom hg template maps _replace_ the CVS-like default ones

     HGdate = {date|rfc822date}

     lastlog = {desc} ## same as {desc|firstline} in this context

     checked in by = {author}

     ...

If no [keywordmaps] are configured the extension falls back on the

following defaults:

     Revision: changeset id

     Author: username

     Date: %Y/%m/%d %H:%M:%S      ## [UTC]

     RCSFile: basename,v

     Source: /path/to/basename,v

     Id: basename,v csetid %Y/%m/%d %H:%M:%S username

     Header: /path/to/basename,v csetid %Y/%m/%d %H:%M:%S username

'''

from mercurial.i18n import gettext as _

from mercurial import commands, fancyopts, templater, util

from mercurial import cmdutil, context, filelog

# findcmd might be in cmdutil or commands

# depending on mercurial version

if hasattr(cmdutil, "findcmd"):

    findcmd = cmdutil.findcmd

else:

    findcmd = commands.findcmd

import os, re, sys, time

deftemplates = {

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

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

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

        'RCSFile': '{file|basename},v',

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

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

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

        }

nokwcommands = ('add', 'addremove', 'bundle', 'clone', 'copy', 'export',

                'incoming', 'outgoing', 'push', 'remove', 'rename', 'rollback')

def utcdate(date):

    '''Returns hgdate in cvs-like UTC format.'''

    return time.strftime('%Y/%m/%d %H:%M:%S', time.gmtime(date[0]))

def getcmd(ui):

    '''Returns current hg command.'''

    # commands.parse(ui, sys.argv[1:])[0] breaks "hg diff -r"

    try:

        args = fancyopts.fancyopts(sys.argv[1:], commands.globalopts, {})

    except fancyopts.getopt.GetoptError, inst:

        raise commands.ParseError(None, inst)

    if args:

        cmd = args[0]

        aliases, i = findcmd(ui, cmd)

        return aliases[0]

class kwtemplater(object):

    '''

    Sets up keyword templates, corresponding keyword regex, and

    provides keyword substitution functions.

    '''

    def __init__(self, ui, repo, path='', node=None):

        self.ui = ui

        self.repo = repo

        self.path = path

        self.node = node

        templates = dict(ui.configitems('keywordmaps'))

        if templates:

            # parse templates here for less overhead in kwsub matchfunc

            for k in templates.keys():

                templates[k] = templater.parsestring(templates[k],

                                                     quoted=False)

        self.templates = templates or deftemplates

        escaped = [re.escape(k) for k in self.templates.keys()]

        self.re_kw = re.compile(r'\$(%s)[^$]*?\$' % '|'.join(escaped))

        templater.common_filters['utcdate'] = utcdate

        try:

            self.t = cmdutil.changeset_templater(ui, repo, False, '', False)

        except TypeError:

            # depending on hg rev changeset_templater has extra "brinfo" arg

            self.t = cmdutil.changeset_templater(ui, repo,

                                                 False, None, '', False)

    def kwsub(self, mobj):

        '''Substitutes keyword using corresponding template.'''

        kw = mobj.group(1)

        self.t.use_template(self.templates[kw])

        self.ui.pushbuffer()

        self.t.show(changenode=self.node, root=self.repo.root, file=self.path)

        keywordsub = templater.firstline(self.ui.popbuffer())

        return '$%s: %s $' % (kw, keywordsub)

    def expand(self, node, data):

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

        if util.binary(data):

            return data

        c = context.filectx(self.repo, self.path, fileid=node)

        self.node = c.node()

        return self.re_kw.sub(self.kwsub, data)

    def shrink(self, text):

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

        if util.binary(text):

            return text

        return self.re_kw.sub(r'$\1$', text)

    def overwrite(self, candidates):

        '''Overwrites candidates in working dir expanding keywords.'''

        files = []

        for f in candidates:

            data = self.repo.wfile(f).read()

            if not util.binary(data):

                self.path = f

                data, kwct = self.re_kw.subn(self.kwsub, data)

                if kwct:

                    self.ui.debug(_('overwriting %s expanding keywords\n') % f)

                    self.repo.wfile(f, 'w').write(data)

                    files.append(f)

        if files:

            self.repo.dirstate.update(files, 'n')

class kwfilelog(filelog.filelog):

    '''

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

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

    '''

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

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

        self.kwtemplater = kwtemplater

    def read(self, node):

        '''Substitutes keywords when reading filelog.'''

        data = super(kwfilelog, self).read(node)

        return self.kwtemplater.expand(node, data)

    def add(self, text, meta, tr, link, p1=None, p2=None):

        '''Removes keyword substitutions when adding to filelog.'''

        text = self.kwtemplater.shrink(text)

        return super(kwfilelog, self).add(text, meta, tr, link, p1=p1, p2=p2)

    def cmp(self, node, text):

        '''Removes keyword substitutions for comparison.'''

        text = self.kwtemplater.shrink(text)

        return super(kwfilelog, self).cmp(node, text)

def reposetup(ui, repo):

    '''Sets up repo as kwrepo for keyword substitution.

    Overrides file method to return kwfilelog instead of filelog

    if file matches user configuration.

    Wraps commit to overwrite configured files with updated

    keyword substitutions.

    This is done for local repos only, and only if there are

    files configured at all for keyword substitution.'''

    if not repo.local() or getcmd(repo.ui) in nokwcommands:

        return

    inc, exc = [], ['.hg*']

    for pat, opt in repo.ui.configitems('keyword'):

        if opt != 'ignore':

            inc.append(pat)

        else:

            exc.append(pat)

    if not inc:

        return

    kwfmatcher = util.matcher(repo.root, inc=inc, exc=exc)[1]

    class kwrepo(repo.__class__):

        '''

        Subclass of repo's class attribute to wrap its file and commit methods.

        '''

        def file(self, f):

            '''Opens kwfilelog instead of filelog if needed.'''

            if f[0] == '/':

                f = f[1:]

            if kwfmatcher(f):

                kwt = kwtemplater(repo.ui, self, path=f)

                return kwfilelog(self.sopener, f, kwt)

            else:

                return filelog.filelog(self.sopener, f)

        def commit(self, files=None, text="", user=None, date=None,

                   match=util.always, force=False, lock=None, wlock=None,

                   force_editor=False, p1=None, p2=None, extra={}):

            '''Wraps commit, expanding keywords of committed and

            configured files in working directory.'''

            wrelease = False

            if not wlock:

                wlock = self.wlock()

                wrelease = True

            try:

                removed = self.status(node1=p1, node2=p2, files=files,

                                      match=match, wlock=wlock)[2]

                node = super(kwrepo,

                             self).commit(files=files, text=text, user=user,

                                          date=date, match=match, force=force,

                                          lock=lock, wlock=wlock,

                                          force_editor=force_editor,

                                          p1=p1, p2=p2, extra=extra)

                if node is None:

                    return node

                candidates = self.changelog.read(node)[3]

                candidates = [f for f in candidates if kwfmatcher(f)

                              and f not in removed

                              and not os.path.islink(self.wjoin(f))]

                if not candidates:

                    return node

                kwt = kwtemplater(repo.ui, self, node=node)

                kwt.overwrite(candidates)

                return node

            finally:

                if wrelease:

                    wlock.release()

    repo.__class__ = kwrepo