hgkw/keyword.py
author Christian Ebert <blacktrash@gmx.net>
Wed, 14 Feb 2007 02:50:06 +0100
branchkwmap-templates
changeset 143 3485b0ef99c4
parent 142 aedce5bcc9eb
child 144 19b3d1de3db0
permissions -rw-r--r--
Add wlock during overwrite Message-ID: <20070213102701.GC30656@cecm.usp.br>

# 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.
#
# Default $keywords$ and their $keyword: substition $ are:
#     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
#
# Expansions spanning more than one line are truncated to their first line.
# Incremental expansion (like CVS' $Log$) is not supported.
#
# Simple setup in hgrc:
#
#     # enable extension
#     keyword = /full/path/to/keyword.py
#     # or, if script in hgext folder:
#     # hgext.keyword =
#
#     # filename patterns for expansion are configured in this section
#     # files matching patterns with value 'ignore' are ignored
#     [keyword]
#     **.py =
#     x* = ignore
#     ...
#     # in case you prefer your own keyword maps over the cvs-like defaults:
#     [keywordmaps]
#     HGdate = {date|rfc822date}
#     lastlog = {desc} ## same as {desc|firstline} in this context
#     checked in by = {author}
#     ...

'''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.

Substitution takes place on every commit and update of the working
repository.

Configuration is done in the [keyword] and [keywordmaps] sections of
hgrc files.
'''

try:
    from mercurial.demandload import * # stable
    from mercurial.i18n import gettext as _
    demandload(globals(), 'mercurial:cmdutil,templater,util')
    demandload(globals(), 'mercurial:context,filelog,revlog')
    demandload(globals(), 're time')
except ImportError: # demandimport
    from mercurial.i18n import _
    from mercurial import cmdutil, templater, util
    from mercurial import context, filelog, revlog
    import re, 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]))

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')


def reposetup(ui, repo):
    '''Sets up repo, and filelog especially, as kwrepo and kwfilelog
    for keyword substitution. This is done for local repos only.'''

    if not repo.local():
        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

    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:]
            return filelog.filelog(self.sopener, f, self, self.revlogversion)

        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.'''

            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]
            candidates = [f for f in candidates if f not in removed
                    and not self._link(f) and self.kwfmatcher(f)]
            if not candidates:
                return node

            wrelease = False
            if not wlock:
                wlock = self.wlock()
                wrelease = True
            try:
                kwt = kwtemplater(self.ui, self)
                kwt.overwrite(candidates, node)
            finally:
                if wrelease:
                    wlock.release()
            return node

    class kwfilelog(filelog.filelog):
        '''
        Superclass over filelog to customize it's read, add, cmp methods.
        Keywords are "stored" unexpanded, and expanded on reading.
        '''
        def __init__(self, opener, path, repo=None,
                     defversion=revlog.REVLOG_DEFAULT_VERSION):
            super(kwfilelog, self).__init__(opener, path, defversion)
            self._repo = repo
            self._path = path
            # only init kwtemplater if needed
            if hasattr(repo, 'kwfmatcher') and repo.kwfmatcher(path):
                self.kwt = kwtemplater(repo.ui, repo)
            else:
                self.kwt = None

        def read(self, node):
            '''Substitutes keywords when reading filelog.'''
            data = super(kwfilelog, self).read(node)
            if self.kwt:
                data = self.kwt.expand(self._path, node, self, data)
            return data

        def add(self, text, meta, tr, link, p1=None, p2=None):
            '''Removes keyword substitutions when adding to filelog.'''
            if self.kwt:
                text = self.kwt.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.'''
            if self.kwt:
                text = self.kwt.shrink(text)
            return super(kwfilelog, self).cmp(node, text)

    filelog.filelog = kwfilelog
    repo.__class__ = kwrepo