hgkw/keyword.py
author Christian Ebert <blacktrash@gmx.net>
Wed, 18 Jul 2007 22:45:50 +0200
changeset 193 7a775a8f6fb9
parent 192 0d2a6c9f8343
child 201 e826c3cdc52d
permissions -rw-r--r--
Discard kwmap-templates branch; keep only early forked off default

# 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}).
#
# 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 or archive distribution.

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

Example:
    [extensions]
    hgext.keyword =

    [keyword]
    # expand keywords in every python file except those matching "x*"
    **.py =
    x* = ignore

Note: the more specific you are in your [keyword] filename patterns
      the less you lose speed in huge repos.

For a [keywordmaps] template mapping and expansion demonstration
run "hg kwdemo".

An additional date template filter {date|utcdate} is provided.

You can replace the default template mappings with customized keywords
and templates of your choice.
Again, run "hg kwdemo" to control the results of your config changes.

When you change keyword configuration, especially the active keywords,
and do not want to store expanded keywords in change history, run
"hg kwshrink", and then change configuration.

Caveat: "hg import" fails if the patch context contains an active
        keyword. In that case run "hg kwshrink", reimport, and then
        "hg kwexpand".
        Or, better, use bundle/unbundle to share changes.
'''

from mercurial import commands, cmdutil, context, fancyopts
from mercurial import filelog, localrepo, templater, util, hg
from mercurial.i18n import gettext as _
# 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, shutil, sys, tempfile, time

commands.optionalrepo += ' kwdemo'

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]

def keywordmatcher(ui, repo):
    '''Collects include/exclude filename patterns for expansion
    candidates of current configuration. Returns filename matching
    function if include patterns exist, None otherwise.'''
    inc, exc = [], ['.hg*']
    for pat, opt in ui.configitems('keyword'):
        if opt != 'ignore':
            inc.append(pat)
        else:
            exc.append(pat)
    if not inc:
        return None
    return util.matcher(repo.root, inc=inc, exc=exc)[1]

class kwtemplater(object):
    '''
    Sets up keyword templates, corresponding keyword regex, and
    provides keyword substitution functions.
    '''
    def __init__(self, ui, repo, path='', node=None, expand=True):
        self.ui = ui
        self.repo = repo
        self.path = path
        self.node = node
        templates = dict(ui.configitems('keywordmaps'))
        if templates:
            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))
        if expand:
            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)
        else:
            self.t = None

    def ctxnode(self, node):
        '''Obtains missing node from file context.'''
        if not self.node:
            c = context.filectx(self.repo, self.path, fileid=node)
            self.node = c.node()

    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 keywords expanded.'''
        if util.binary(data):
            return data
        self.ctxnode(node)
        return self.re_kw.sub(self.kwsub, data)

    def process(self, node, data):
        '''Returns a tuple: data, count.
        Count is number of keywords/keyword substitutions.
        Keywords in data are expanded, if templater was initialized.'''
        if util.binary(data):
            return data, None
        if self.t:
            self.ctxnode(node)
            return self.re_kw.subn(self.kwsub, data)
        return data, self.re_kw.search(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, man, commit=True):
        '''Overwrites files in working directory if keywords are detected.
        Keywords are expanded if keyword templater is initialized,
        otherwise their substitution is removed.'''
        expand = self.t is not None
        action = ('shrinking', 'expanding')[expand]
        notify = (self.ui.note, self.ui.debug)[commit]
        files = []
        for f in candidates:
            fp = self.repo.file(f, kwcnt=True, kwexp=expand)
            data, cnt = fp.read(man[f])
            if cnt:
                notify(_('overwriting %s %s keywords\n') % (f, action))
                try:
                    self.repo.wwrite(f, data, man.flags(f))
                except AttributeError:
                    # older versions want file descriptor as 3. optional arg
                    self.repo.wwrite(f, 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 processed on reading.
    '''
    def __init__(self, opener, path, kwtemplater, kwcnt):
        super(kwfilelog, self).__init__(opener, path)
        self.kwtemplater = kwtemplater
        self.kwcnt = kwcnt

    def read(self, node):
        '''Passes data through kwemplater methods for
        either unconditional keyword expansion
        or counting of keywords and substitution method
        set by the calling overwrite function.'''
        data = super(kwfilelog, self).read(node)
        if not self.kwcnt:
            return self.kwtemplater.expand(node, data)
        return self.kwtemplater.process(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)
        if self.renamed(node):
            t2 = super(kwfilelog, self).read(node)
            return t2 != text
        return super(kwfilelog, self).cmp(node, text)

def overwrite(ui, repo, files=None, expand=True):
    '''Expands/shrinks keywords in working directory.'''
    wlock = repo.wlock()
    try:
        ctx = repo.changectx()
        if not ctx:
            raise hg.RepoError(_('no changeset found'))
        for changed in repo.status()[:4]:
            if changed:
                raise util.Abort(_('local changes detected'))
        kwfmatcher = keywordmatcher(ui, repo)
        if kwfmatcher is None:
            ui.warn(_('no files configured for keyword expansion\n'))
            return
        m = ctx.manifest()
        if files:
            files = [f for f in files if f in m.keys()]
        else:
            files = m.keys()
        files = [f for f in files if kwfmatcher(f) and not os.path.islink(f)]
        if not files:
            ui.warn(_('given files not tracked or '
                      'not configured for expansion\n'))
            return
        kwt = kwtemplater(ui, repo, node=ctx.node(), expand=expand)
        kwt.overwrite(files, m, commit=False)
    finally:
        wlock.release()


def shrink(ui, repo, *args):
    '''revert expanded keywords in working directory

    run before:
               disabling keyword expansion
               changing keyword expansion configuration
    or if you experience problems with "hg import"
    '''
    overwrite(ui, repo, files=args, expand=False)

def expand(ui, repo, *args):
    '''expand keywords in working directory

    run after (re)enabling keyword expansion
    '''
    overwrite(ui, repo, files=args)

def demo(ui, repo, *args, **opts):
    '''print [keywordmaps] configuration and an expansion example

    show current, custom, or default keyword template maps and their expansion
    '''
    msg = 'hg keyword config and expansion example'
    kwstatus = 'current'
    fn = 'demo.txt'
    tmpdir = tempfile.mkdtemp('', 'kwdemo.')
    ui.note(_('creating temporary repo at %s\n') % tmpdir)
    _repo = localrepo.localrepository(ui, path=tmpdir, create=True)
    # for backwards compatibility
    ui = _repo.ui
    ui.setconfig('keyword', fn, '')
    if opts['default']:
        kwstatus = 'default'
        kwmaps = deftemplates
    else:
        if args or opts['rcfile']:
            kwstatus = 'custom'
        for tmap in args:
            k, v = tmap.split('=', 1)
            ui.setconfig('keywordmaps', k.strip(), v.strip())
        if opts['rcfile']:
            ui.readconfig(opts['rcfile'])
        kwmaps = dict(ui.configitems('keywordmaps')) or deftemplates
    if ui.configitems('keywordmaps'):
        for k, v in kwmaps.items():
            ui.setconfig('keywordmaps', k, v)
    reposetup(ui, _repo)
    ui.status(_('config with %s keyword template maps:\n') % kwstatus)
    ui.write('[keyword]\n%s =\n[keywordmaps]\n' % fn)
    for k, v in kwmaps.items():
        ui.write('%s = %s\n' % (k, v))
    path = _repo.wjoin(fn)
    keywords = '$' + '$\n$'.join(kwmaps.keys()) + '$\n'
    _repo.wopener(fn, 'w').write(keywords)
    _repo.add([fn])
    ui.note(_('\n%s keywords written to %s:\n') % (kwstatus, path))
    ui.note(keywords)
    ui.note(_("\nhg --repository '%s' commit\n") % tmpdir)
    _repo.commit(text=msg)
    pathinfo = ('', ' in %s' % path)[ui.verbose]
    ui.status(_('\n%s keywords expanded%s:\n') % (kwstatus, pathinfo))
    ui.write(_repo.wread(fn))
    ui.debug(_('\nremoving temporary repo %s\n') % tmpdir)
    shutil.rmtree(tmpdir)


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.'''

    # for backwards compatibility
    ui = repo.ui

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

    kwfmatcher = keywordmatcher(ui, repo)
    if kwfmatcher is None:
        return

    class kwrepo(repo.__class__):
        def file(self, f, kwcnt=False, kwexp=True):
            if f[0] == '/':
                f = f[1:]
            if kwfmatcher(f):
                kwt = kwtemplater(ui, self, path=f, expand=kwexp)
                return kwfilelog(self.sopener, f, kwt, kwcnt)
            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={}):
            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

                cl = self.changelog.read(node)
                candidates = [f for f in cl[3] if kwfmatcher(f)
                              and f not in removed
                              and not os.path.islink(self.wjoin(f))]
                if candidates:
                    m = self.manifest.read(cl[0])
                    kwt = kwtemplater(ui, self, node=node)
                    kwt.overwrite(candidates, m)
                return node
            finally:
                if wrelease:
                    wlock.release()

    repo.__class__ = kwrepo


cmdtable = {
    'kwdemo':
        (demo,
         [('d', 'default', None, _('show default keyword template maps')),
          ('f', 'rcfile', [], _('read maps from RCFILE'))],
         _('hg kwdemo [-d || [-f RCFILE] TEMPLATEMAP ...]')),
    'kwshrink': (shrink, [], _('hg kwshrink [NAME] ...')),
    'kwexpand': (expand, [], _('hg kwexpand [NAME] ...')),
}