# 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.

The exansion works in 2 modes:

    1) working mode: substitution takes place on every commit and

       update of the working repository.

    2) archive mode: substitution is only triggered by "hg archive".

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

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

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

     ** = archive     ## keywords in all textfiles are expanded

                      ## when creating a distribution

     y* = noarchive   ## keywords in files matching "y*" are not expanded

                      ## on archive creation

     ...

     [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

'''

try:

    from mercurial.demandload import * # stable

    from mercurial.i18n import gettext as _

    demandload(globals(), 'mercurial:commands,fancyopts,templater,util')

    demandload(globals(), 'mercurial:cmdutil,context,filelog')

    demandload(globals(), 'os re sys time')

except ImportError:                    # demandimport

    from mercurial.i18n import _

    from mercurial import commands, fancyopts, templater, util

    from mercurial import cmdutil, context, filelog

    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}',

        }

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 = commands.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):

        self.ui = ui

        self.repo = repo

        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

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

                '|'.join([re.escape(k) for k in self.templates.keys()]))

        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, path, node):

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

        kw = mobj.group(1)

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

        self.ui.pushbuffer()

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

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

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

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

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

        if util.binary(data):

            return data

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

        cnode = c.node()

        return self.re_kw.sub(lambda m: self.kwsub(m, path, cnode), 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, node):

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

        files = []

        for f in candidates:

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

            if not util.binary(data):

                data, kwct = self.re_kw.subn(lambda m:

                        self.kwsub(m, f, node), 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):

    '''

    Superclass over filelog to customize 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.path = path

        self.kwtemplater = kwtemplater

    def read(self, node):

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

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

        return self.kwtemplater.expand(self.path, node, self, 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():

        return

    archivemode = (getcmd(repo.ui) == 'archive')

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

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

        if opt == 'archive':

            archive.append(pat)

        elif opt == 'noarchive':

            noarchive.append(pat)

        elif opt == 'ignore':

            exc.append(pat)

        else:

            inc.append(pat)

    if archivemode:

        inc, exc = archive, noarchive

    if not inc:

        return

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

    class kwrepo(repo.__class__):

        def file(self, f):

            if f[0] == '/':

                f = f[1:]

            # only use kwfilelog when needed

            if self.kwfmatcher(f):

                kwt = kwtemplater(self.ui, self)

                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()[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]

                if not candidates:

                    return node

                kwt = kwtemplater(self.ui, self)

                kwt.overwrite(candidates, node)

                return node

            finally:

                if wrelease:

                    wlock.release()

    repo.__class__ = kwrepo