hgweb: port archive command to modern response API

Well, I tried to go with PEP 3333's recommendations and only allow
our WSGI application to emit data via a response generator.
Unfortunately, the "archive" command calls into the zipfile and
tarfile modules and these operator on file objects and must send
their data to an object with write(). There's no easy way turn
these write() calls into a generator.

So, we teach our response type how to expose a file object like
object that can be used to write() output. We try to keep the API
consistent with how things work currently: callers must call a
setbody*(), then sendresponse() to trigger sending of headers,
and only then can they get a handle on the object to perform
writing.

This required overloading the return value of @webcommand functions
even more. Fortunately, we're almost completely ported off the
legacy API. So we should be able to simplify matters in the near
future.

A test relying on this functionality has also been updated to use
the new API.

Differential Revision: https://phab.mercurial-scm.org/D2792

hgext/keyword.py

@@ -1,814 +1,817 b''
 # keyword.py - $Keyword$ expansion for Mercurial
 #
 # Copyright 2007-2015 Christian Ebert <[e-mail]>
 #
 # 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 tempfile
 import weakref
 
 from mercurial.i18n import _
 from mercurial.hgweb import webcommands
 
 from mercurial import (
     cmdutil,
     context,
     dispatch,
     error,
     extensions,
     filelog,
     localrepo,
     logcmdutil,
     match,
     patch,
     pathutil,
     pycompat,
     registrar,
     scmutil,
     templatefilters,
     util,
 )
 from mercurial.utils import dateutil
 
 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 = 'ships-with-hg-core'
 
 # hg commands that do not act on keywords
 nokwcommands = ('add addremove annotate bundle export grep incoming init log'
                 ' outgoing push tip verify convert email glog')
 
 # webcommands that do not act on keywords
 nokwwebcommands = ('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 = ('merge kwexpand kwshrink record qrecord resolve transplant'
               ' unshelve rebase graft backout histedit fetch')
 
 # names of extensions using dorecord
 recordextensions = 'record'
 
 colortable = {
     'kwfiles.enabled': 'green bold',
     'kwfiles.deleted': 'cyan bold underline',
     'kwfiles.enabledunknown': 'green',
     'kwfiles.ignored': 'bold',
     'kwfiles.ignoredunknown': 'none'
 }
 
 templatefilter = registrar.templatefilter()
 
 configtable = {}
 configitem = registrar.configitem(configtable)
 
 configitem('keywordset', 'svn',
     default=False,
 )
 # date like in cvs' $Date
 @templatefilter('utcdate')
 def utcdate(text):
     '''Date. Returns a UTC-date in this format: "2009/08/18 11:00:13".
     '''
     dateformat = '%Y/%m/%d %H:%M:%S'
     return dateutil.datestr((dateutil.parsedate(text)[0], 0), dateformat)
 # date like in svn's $Date
 @templatefilter('svnisodate')
 def svnisodate(text):
     '''Date. Returns a date in this format: "2009-08-18 13:00:13
     +0200 (Tue, 18 Aug 2009)".
     '''
     return dateutil.datestr(text, '%Y-%m-%d %H:%M:%S %1%2 (%a, %d %b %Y)')
 # date like in svn's $Id
 @templatefilter('svnutcdate')
 def svnutcdate(text):
     '''Date. Returns a UTC-date in this format: "2009-08-18
     11:00:13Z".
     '''
     dateformat = '%Y-%m-%d %H:%M:%SZ'
     return dateutil.datestr((dateutil.parsedate(text)[0], 0), dateformat)
 
 # make keyword tools accessible
 kwtools = {'hgcmd': ''}
 
 def _defaultkwmaps(ui):
     '''Returns default keywordmaps according to keywordset configuration.'''
     templates = {
         'Revision': '{node|short}',
         'Author': '{author|user}',
     }
     kwsets = ({
         'Date': '{date|utcdate}',
         'RCSfile': '{file|basename},v',
         'RCSFile': '{file|basename},v', # kept for backwards compatibility
                                         # with hg-keyword
         'Source': '{root}/{file},v',
         'Id': '{file|basename},v {node|short} {date|utcdate} {author|user}',
         'Header': '{root}/{file},v {node|short} {date|utcdate} {author|user}',
     }, {
         'Date': '{date|svnisodate}',
         'Id': '{file|basename},v {node|short} {date|svnutcdate} {author|user}',
         'LastChangedRevision': '{node|short}',
         'LastChangedBy': '{author|user}',
         'LastChangedDate': '{date|svnisodate}',
     })
     templates.update(kwsets[ui.configbool('keywordset', 'svn')])
     return templates
 
 def _shrinktext(text, subfunc):
     '''Helper for keyword expansion removal in text.
     Depending on subfunc also returns number of substitutions.'''
     return subfunc(r'$\1$', 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, '', [], inc, exc)
         self.restrict = kwtools['hgcmd'] in restricted.split()
         self.postcommit = False
 
         kwmaps = self.ui.configitems('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 '|'.join(map(re.escape, self.templates.keys()))
 
     @util.propertycache
     def rekw(self):
         '''Returns regex for unexpanded keywords.'''
         return re.compile(r'\$(%s)\$' % self.escape)
 
     @util.propertycache
     def rekwexp(self):
         '''Returns regex for expanded keywords.'''
         return re.compile(r'\$(%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 '$%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 util.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 '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 = _('overwriting %s expanding keywords\n')
         else:
             msg = _('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 util.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, "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 util.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 = ''.join(lines)
             if not util.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)
         self.kwt = kwt
         self.path = path
 
     def read(self, node):
         '''Expands keywords when reading filelog.'''
         data = super(kwfilelog, self).read(node)
         if self.renamed(node):
             return data
         return self.kwt.expand(self.path, node, data)
 
     def add(self, text, meta, tr, link, p1=None, p2=None):
         '''Removes keyword substitutions when adding to filelog.'''
         text = self.kwt.shrink(self.path, text)
         return super(kwfilelog, self).add(text, meta, tr, link, p1, p2)
 
     def cmp(self, node, text):
         '''Removes keyword substitutions for comparison.'''
         text = self.kwt.shrink(self.path, text)
         return super(kwfilelog, self).cmp(node, text)
 
 def _status(ui, repo, wctx, kwt, *pats, **opts):
     '''Bails out if [keyword] configuration is not active.
     Returns status of working directory.'''
     if kwt:
         opts = pycompat.byteskwargs(opts)
         return repo.status(match=scmutil.match(wctx, pats, opts), clean=True,
                            unknown=opts.get('unknown') or opts.get('all'))
     if ui.configitems('keyword'):
         raise error.Abort(_('[keyword] patterns cannot match'))
     raise error.Abort(_('no [keyword] patterns configured'))
 
 def _kwfwrite(ui, repo, expand, *pats, **opts):
     '''Selects files and passes them to kwtemplater.overwrite.'''
     wctx = repo[None]
     if len(wctx.parents()) > 1:
         raise error.Abort(_('outstanding uncommitted merge'))
     kwt = getattr(repo, '_keywordkwt', None)
     with repo.wlock():
         status = _status(ui, repo, wctx, kwt, *pats, **opts)
         if status.modified or status.added or status.removed or status.deleted:
             raise error.Abort(_('outstanding uncommitted changes'))
         kwt.overwrite(wctx, status.clean, True, expand)
 
 @command('kwdemo',
          [('d', 'default', None, _('show default keyword template maps')),
           ('f', 'rcfile', '',
            _('read maps from rcfile'), _('FILE'))],
          _('hg kwdemo [-d] [-f RCFILE] [TEMPLATEMAP]...'),
          optionalrepo=True)
 def demo(ui, repo, *args, **opts):
     '''print [keywordmaps] configuration and an expansion example
 
     Show current, custom, or default keyword template maps and their
     expansions.
 
     Extend the current configuration by specifying maps as arguments
     and using -f/--rcfile to source an external hgrc file.
 
     Use -d/--default to disable current configuration.
 
     See :hg:`help templates` for information on templates and filters.
     '''
     def demoitems(section, items):
         ui.write('[%s]\n' % section)
         for k, v in sorted(items):
             ui.write('%s = %s\n' % (k, v))
 
     fn = 'demo.txt'
     tmpdir = tempfile.mkdtemp('', 'kwdemo.')
     ui.note(_('creating temporary repository at %s\n') % tmpdir)
     if repo is None:
         baseui = ui
     else:
         baseui = repo.baseui
     repo = localrepo.localrepository(baseui, tmpdir, True)
     ui.setconfig('keyword', fn, '', 'keyword')
     svn = ui.configbool('keywordset', 'svn')
     # explicitly set keywordset for demo output
     ui.setconfig('keywordset', 'svn', svn, 'keyword')
 
     uikwmaps = ui.configitems('keywordmaps')
     if args or opts.get(r'rcfile'):
         ui.status(_('\n\tconfiguration using custom keyword template maps\n'))
         if uikwmaps:
             ui.status(_('\textending current template maps\n'))
         if opts.get(r'default') or not uikwmaps:
             if svn:
                 ui.status(_('\toverriding default svn keywordset\n'))
             else:
                 ui.status(_('\toverriding default cvs keywordset\n'))
         if opts.get(r'rcfile'):
             ui.readconfig(opts.get('rcfile'))
         if args:
             # simulate hgrc parsing
             rcmaps = '[keywordmaps]\n%s\n' % '\n'.join(args)
             repo.vfs.write('hgrc', rcmaps)
             ui.readconfig(repo.vfs.join('hgrc'))
         kwmaps = dict(ui.configitems('keywordmaps'))
     elif opts.get(r'default'):
         if svn:
             ui.status(_('\n\tconfiguration using default svn keywordset\n'))
         else:
             ui.status(_('\n\tconfiguration using default cvs keywordset\n'))
         kwmaps = _defaultkwmaps(ui)
         if uikwmaps:
             ui.status(_('\tdisabling current template maps\n'))
             for k, v in kwmaps.iteritems():
                 ui.setconfig('keywordmaps', k, v, 'keyword')
     else:
         ui.status(_('\n\tconfiguration using current keyword template maps\n'))
         if uikwmaps:
             kwmaps = dict(uikwmaps)
         else:
             kwmaps = _defaultkwmaps(ui)
 
     uisetup(ui)
     reposetup(ui, repo)
     ui.write(('[extensions]\nkeyword =\n'))
     demoitems('keyword', ui.configitems('keyword'))
     demoitems('keywordset', ui.configitems('keywordset'))
     demoitems('keywordmaps', kwmaps.iteritems())
     keywords = '$' + '$\n$'.join(sorted(kwmaps.keys())) + '$\n'
     repo.wvfs.write(fn, keywords)
     repo[None].add([fn])
     ui.note(_('\nkeywords written to %s:\n') % fn)
     ui.note(keywords)
     with repo.wlock():
         repo.dirstate.setbranch('demobranch')
     for name, cmd in ui.configitems('hooks'):
         if name.split('.', 1)[0].find('commit') > -1:
             repo.ui.setconfig('hooks', name, '', 'keyword')
     msg = _('hg keyword configuration and expansion example')
     ui.note(("hg ci -m '%s'\n" % msg))
     repo.commit(text=msg)
     ui.status(_('\n\tkeywords expanded\n'))
     ui.write(repo.wread(fn))
     repo.wvfs.rmtree(repo.root)
 
 @command('kwexpand',
     cmdutil.walkopts,
     _('hg kwexpand [OPTION]... [FILE]...'),
     inferrepo=True)
 def expand(ui, repo, *pats, **opts):
     '''expand keywords in the working directory
 
     Run after (re)enabling keyword expansion.
 
     kwexpand refuses to run if given files contain local changes.
     '''
     # 3rd argument sets expansion to True
     _kwfwrite(ui, repo, True, *pats, **opts)
 
 @command('kwfiles',
          [('A', 'all', None, _('show keyword status flags of all files')),
           ('i', 'ignore', None, _('show files excluded from expansion')),
           ('u', 'unknown', None, _('only show unknown (not tracked) files')),
          ] + cmdutil.walkopts,
          _('hg kwfiles [OPTION]... [FILE]...'),
          inferrepo=True)
 def files(ui, repo, *pats, **opts):
     '''show files configured for keyword expansion
 
     List which files in the working directory are matched by the
     [keyword] configuration patterns.
 
     Useful to prevent inadvertent keyword expansion and to speed up
     execution by including only files that are actual candidates for
     expansion.
 
     See :hg:`help keyword` on how to construct patterns both for
     inclusion and exclusion of files.
 
     With -A/--all and -v/--verbose the codes used to show the status
     of files are::
 
       K = keyword expansion candidate
       k = keyword expansion candidate (not tracked)
       I = ignored
       i = ignored (not tracked)
     '''
     kwt = getattr(repo, '_keywordkwt', None)
     wctx = repo[None]
     status = _status(ui, repo, wctx, kwt, *pats, **opts)
     if pats:
         cwd = repo.getcwd()
     else:
         cwd = ''
     files = []
     opts = pycompat.byteskwargs(opts)
     if not opts.get('unknown') or opts.get('all'):
         files = sorted(status.modified + status.added + status.clean)
     kwfiles = kwt.iskwfile(files, wctx)
     kwdeleted = kwt.iskwfile(status.deleted, wctx)
     kwunknown = kwt.iskwfile(status.unknown, wctx)
     if not opts.get('ignore') or opts.get('all'):
         showfiles = kwfiles, kwdeleted, kwunknown
     else:
         showfiles = [], [], []
     if opts.get('all') or opts.get('ignore'):
         showfiles += ([f for f in files if f not in kwfiles],
                       [f for f in status.unknown if f not in kwunknown])
     kwlabels = 'enabled deleted enabledunknown ignored ignoredunknown'.split()
     kwstates = zip(kwlabels, 'K!kIi', showfiles)
     fm = ui.formatter('kwfiles', opts)
     fmt = '%.0s%s\n'
     if opts.get('all') or ui.verbose:
         fmt = '%s %s\n'
     for kwstate, char, filenames in kwstates:
         label = 'kwfiles.' + kwstate
         for f in filenames:
             fm.startitem()
             fm.write('kwstatus path', fmt, char,
                      repo.pathto(f, cwd), label=label)
     fm.end()
 
 @command('kwshrink',
     cmdutil.walkopts,
     _('hg kwshrink [OPTION]... [FILE]...'),
     inferrepo=True)
 def shrink(ui, repo, *pats, **opts):
     '''revert expanded keywords in the working directory
 
     Must be run before changing/disabling active keywords.
 
     kwshrink refuses to run if given files contain local changes.
     '''
     # 3rd argument sets expansion to False
     _kwfwrite(ui, repo, False, *pats, **opts)
 
 # monkeypatches
 
 def kwpatchfile_init(orig, self, ui, gp, backend, store, eolmode=None):
     '''Monkeypatch/wrap patch.patchfile.__init__ to avoid
     rejects or conflicts due to expanded keywords in working dir.'''
     orig(self, ui, gp, backend, store, eolmode)
     kwt = getattr(getattr(backend, 'repo', None), '_keywordkwt', None)
     if kwt:
         # shrink keywords read from working dir
         self.lines = kwt.shrinklines(self.fname, self.lines)
 
 def kwdiff(orig, repo, *args, **kwargs):
     '''Monkeypatch patch.diff to avoid expansion.'''
     kwt = getattr(repo, '_keywordkwt', None)
     if kwt:
         restrict = kwt.restrict
         kwt.restrict = True
     try:
         for chunk in orig(repo, *args, **kwargs):
             yield chunk
     finally:
         if kwt:
             kwt.restrict = restrict
 
 def kwweb_skip(orig, web, req, tmpl):
     '''Wraps webcommands.x turning off keyword expansion.'''
     kwt = getattr(web.repo, '_keywordkwt', None)
     if kwt:
         origmatch = kwt.match
         kwt.match = util.never
     try:
         res = orig(web, req, tmpl)
         if res is web.res:
             res = res.sendresponse()
+        elif res is True:
+            return
+
         for chunk in res:
             yield chunk
     finally:
         if kwt:
             kwt.match = origmatch
 
 def kw_amend(orig, ui, repo, old, extra, pats, opts):
     '''Wraps cmdutil.amend expanding keywords after amend.'''
     kwt = getattr(repo, '_keywordkwt', None)
     if kwt is None:
         return orig(ui, repo, old, extra, pats, opts)
     with repo.wlock():
         kwt.postcommit = True
         newid = orig(ui, repo, old, extra, pats, opts)
         if newid != old.node():
             ctx = repo[newid]
             kwt.restrict = True
             kwt.overwrite(ctx, ctx.files(), False, True)
             kwt.restrict = False
         return newid
 
 def kw_copy(orig, ui, repo, pats, opts, rename=False):
     '''Wraps cmdutil.copy so that copy/rename destinations do not
     contain expanded keywords.
     Note that the source of a regular file destination may also be a
     symlink:
     hg cp sym x                -> x is symlink
     cp sym x; hg cp -A sym x   -> x is file (maybe expanded keywords)
     For the latter we have to follow the symlink to find out whether its
     target is configured for expansion and we therefore must unexpand the
     keywords in the destination.'''
     kwt = getattr(repo, '_keywordkwt', None)
     if kwt is None:
         return orig(ui, repo, pats, opts, rename)
     with repo.wlock():
         orig(ui, repo, pats, opts, rename)
         if opts.get('dry_run'):
             return
         wctx = repo[None]
         cwd = repo.getcwd()
 
         def haskwsource(dest):
             '''Returns true if dest is a regular file and configured for
             expansion or a symlink which points to a file configured for
             expansion. '''
             source = repo.dirstate.copied(dest)
             if 'l' in wctx.flags(source):
                 source = pathutil.canonpath(repo.root, cwd,
                                            os.path.realpath(source))
             return kwt.match(source)
 
         candidates = [f for f in repo.dirstate.copies() if
                       'l' not in wctx.flags(f) and haskwsource(f)]
         kwt.overwrite(wctx, candidates, False, False)
 
 def kw_dorecord(orig, ui, repo, commitfunc, *pats, **opts):
     '''Wraps record.dorecord expanding keywords after recording.'''
     kwt = getattr(repo, '_keywordkwt', None)
     if kwt is None:
         return orig(ui, repo, commitfunc, *pats, **opts)
     with repo.wlock():
         # record returns 0 even when nothing has changed
         # therefore compare nodes before and after
         kwt.postcommit = True
         ctx = repo['.']
         wstatus = ctx.status()
         ret = orig(ui, repo, commitfunc, *pats, **opts)
         recctx = repo['.']
         if ctx != recctx:
             modified, added = _preselect(wstatus, recctx.files())
             kwt.restrict = False
             kwt.overwrite(recctx, modified, False, True)
             kwt.overwrite(recctx, added, False, True, True)
             kwt.restrict = True
         return ret
 
 def kwfilectx_cmp(orig, self, fctx):
     if fctx._customcmp:
         return fctx.cmp(self)
     kwt = getattr(self._repo, '_keywordkwt', None)
     if kwt is None:
         return orig(self, fctx)
     # keyword affects data size, comparing wdir and filelog size does
     # not make sense
     if (fctx._filenode is None and
         (self._repo._encodefilterpats or
          kwt.match(fctx.path()) and 'l' not in fctx.flags() or
          self.size() - 4 == fctx.size()) or
         self.size() == fctx.size()):
         return self._filelog.cmp(self._filenode, fctx.data())
     return True
 
 def uisetup(ui):
     ''' Monkeypatches dispatch._parse to retrieve user command.
     Overrides file method to return kwfilelog instead of filelog
     if file matches user configuration.
     Wraps commit to overwrite configured files with updated
     keyword substitutions.
     Monkeypatches patch and webcommands.'''
 
     def kwdispatch_parse(orig, ui, args):
         '''Monkeypatch dispatch._parse to obtain running hg command.'''
         cmd, func, args, options, cmdoptions = orig(ui, args)
         kwtools['hgcmd'] = cmd
         return cmd, func, args, options, cmdoptions
 
     extensions.wrapfunction(dispatch, '_parse', kwdispatch_parse)
 
     extensions.wrapfunction(context.filectx, 'cmp', kwfilectx_cmp)
     extensions.wrapfunction(patch.patchfile, '__init__', kwpatchfile_init)
     extensions.wrapfunction(patch, 'diff', kwdiff)
     extensions.wrapfunction(cmdutil, 'amend', kw_amend)
     extensions.wrapfunction(cmdutil, 'copy', kw_copy)
     extensions.wrapfunction(cmdutil, 'dorecord', kw_dorecord)
     for c in nokwwebcommands.split():
         extensions.wrapfunction(webcommands, c, kwweb_skip)
 
 def reposetup(ui, repo):
     '''Sets up repo as kwrepo for keyword substitution.'''
 
     try:
         if (not repo.local() or kwtools['hgcmd'] in nokwcommands.split()
             or '.hg' in util.splitpath(repo.root)
             or repo._url.startswith('bundle:')):
             return
     except AttributeError:
         pass
 
     inc, exc = [], ['.hg*']
     for pat, opt in ui.configitems('keyword'):
         if opt != 'ignore':
             inc.append(pat)
         else:
             exc.append(pat)
     if not inc:
         return
 
     kwt = kwtemplater(ui, repo, inc, exc)
 
     class kwrepo(repo.__class__):
         def file(self, f):
             if f[0] == '/':
                 f = f[1:]
             return kwfilelog(self.svfs, kwt, f)
 
         def wread(self, filename):
             data = super(kwrepo, self).wread(filename)
             return kwt.wread(filename, data)
 
         def commit(self, *args, **opts):
             # use custom commitctx for user commands
             # other extensions can still wrap repo.commitctx directly
             self.commitctx = self.kwcommitctx
             try:
                 return super(kwrepo, self).commit(*args, **opts)
             finally:
                 del self.commitctx
 
         def kwcommitctx(self, ctx, error=False):
             n = super(kwrepo, self).commitctx(ctx, error)
             # no lock needed, only called from repo.commit() which already locks
             if not kwt.postcommit:
                 restrict = kwt.restrict
                 kwt.restrict = True
                 kwt.overwrite(self[n], sorted(ctx.added() + ctx.modified()),
                               False, True)
                 kwt.restrict = restrict
             return n
 
         def rollback(self, dryrun=False, force=False):
             with self.wlock():
                 origrestrict = kwt.restrict
                 try:
                     if not dryrun:
                         changed = self['.'].files()
                     ret = super(kwrepo, self).rollback(dryrun, force)
                     if not dryrun:
                         ctx = self['.']
                         modified, added = _preselect(ctx.status(), changed)
                         kwt.restrict = False
                         kwt.overwrite(ctx, modified, True, True)
                         kwt.overwrite(ctx, added, True, False)
                     return ret
                 finally:
                     kwt.restrict = origrestrict
 
     repo.__class__ = kwrepo
     repo._keywordkwt = kwt

mercurial/hgweb/hgweb_mod.py

@@ -1,455 +1,456 b''
 # hgweb/hgweb_mod.py - Web interface for a repository.
 #
 # Copyright 21 May 2005 - (c) 2005 Jake Edge <[e-mail]>
 # Copyright 2005-2007 Matt Mackall <[e-mail]>
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 from __future__ import absolute_import
 
 import contextlib
 import os
 
 from .common import (
     ErrorResponse,
     HTTP_BAD_REQUEST,
     HTTP_NOT_FOUND,
     HTTP_NOT_MODIFIED,
     HTTP_OK,
     HTTP_SERVER_ERROR,
     cspvalues,
     permhooks,
 )
 
 from .. import (
     encoding,
     error,
     formatter,
     hg,
     hook,
     profiling,
     pycompat,
     repoview,
     templatefilters,
     templater,
     ui as uimod,
     util,
     wireprotoserver,
 )
 
 from . import (
     request as requestmod,
     webcommands,
     webutil,
     wsgicgi,
 )
 
 archivespecs = util.sortdict((
     ('zip', ('application/zip', 'zip', '.zip', None)),
     ('gz', ('application/x-gzip', 'tgz', '.tar.gz', None)),
     ('bz2', ('application/x-bzip2', 'tbz2', '.tar.bz2', None)),
 ))
 
 def getstyle(req, configfn, templatepath):
     styles = (
         req.qsparams.get('style', None),
         configfn('web', 'style'),
         'paper',
     )
     return styles, templater.stylemap(styles, templatepath)
 
 def makebreadcrumb(url, prefix=''):
     '''Return a 'URL breadcrumb' list
 
     A 'URL breadcrumb' is a list of URL-name pairs,
     corresponding to each of the path items on a URL.
     This can be used to create path navigation entries.
     '''
     if url.endswith('/'):
         url = url[:-1]
     if prefix:
         url = '/' + prefix + url
     relpath = url
     if relpath.startswith('/'):
         relpath = relpath[1:]
 
     breadcrumb = []
     urlel = url
     pathitems = [''] + relpath.split('/')
     for pathel in reversed(pathitems):
         if not pathel or not urlel:
             break
         breadcrumb.append({'url': urlel, 'name': pathel})
         urlel = os.path.dirname(urlel)
     return reversed(breadcrumb)
 
 class requestcontext(object):
     """Holds state/context for an individual request.
 
     Servers can be multi-threaded. Holding state on the WSGI application
     is prone to race conditions. Instances of this class exist to hold
     mutable and race-free state for requests.
     """
     def __init__(self, app, repo, req, res):
         self.repo = repo
         self.reponame = app.reponame
         self.req = req
         self.res = res
 
         self.archivespecs = archivespecs
 
         self.maxchanges = self.configint('web', 'maxchanges')
         self.stripecount = self.configint('web', 'stripes')
         self.maxshortchanges = self.configint('web', 'maxshortchanges')
         self.maxfiles = self.configint('web', 'maxfiles')
         self.allowpull = self.configbool('web', 'allow-pull')
 
         # we use untrusted=False to prevent a repo owner from using
         # web.templates in .hg/hgrc to get access to any file readable
         # by the user running the CGI script
         self.templatepath = self.config('web', 'templates', untrusted=False)
 
         # This object is more expensive to build than simple config values.
         # It is shared across requests. The app will replace the object
         # if it is updated. Since this is a reference and nothing should
         # modify the underlying object, it should be constant for the lifetime
         # of the request.
         self.websubtable = app.websubtable
 
         self.csp, self.nonce = cspvalues(self.repo.ui)
 
     # Trust the settings from the .hg/hgrc files by default.
     def config(self, section, name, default=uimod._unset, untrusted=True):
         return self.repo.ui.config(section, name, default,
                                    untrusted=untrusted)
 
     def configbool(self, section, name, default=uimod._unset, untrusted=True):
         return self.repo.ui.configbool(section, name, default,
                                        untrusted=untrusted)
 
     def configint(self, section, name, default=uimod._unset, untrusted=True):
         return self.repo.ui.configint(section, name, default,
                                       untrusted=untrusted)
 
     def configlist(self, section, name, default=uimod._unset, untrusted=True):
         return self.repo.ui.configlist(section, name, default,
                                        untrusted=untrusted)
 
     def archivelist(self, nodeid):
         allowed = self.configlist('web', 'allow_archive')
         for typ, spec in self.archivespecs.iteritems():
             if typ in allowed or self.configbool('web', 'allow%s' % typ):
                 yield {'type': typ, 'extension': spec[2], 'node': nodeid}
 
     def templater(self, req):
         # determine scheme, port and server name
         # this is needed to create absolute urls
         logourl = self.config('web', 'logourl')
         logoimg = self.config('web', 'logoimg')
         staticurl = (self.config('web', 'staticurl')
                      or req.apppath + '/static/')
         if not staticurl.endswith('/'):
             staticurl += '/'
 
         # some functions for the templater
 
         def motd(**map):
             yield self.config('web', 'motd')
 
         # figure out which style to use
 
         vars = {}
         styles, (style, mapfile) = getstyle(req, self.config,
                                             self.templatepath)
         if style == styles[0]:
             vars['style'] = style
 
         sessionvars = webutil.sessionvars(vars, '?')
 
         if not self.reponame:
             self.reponame = (self.config('web', 'name', '')
                              or req.reponame
                              or req.apppath
                              or self.repo.root)
 
         def websubfilter(text):
             return templatefilters.websub(text, self.websubtable)
 
         # create the templater
         # TODO: export all keywords: defaults = templatekw.keywords.copy()
         defaults = {
             'url': req.apppath + '/',
             'logourl': logourl,
             'logoimg': logoimg,
             'staticurl': staticurl,
             'urlbase': req.advertisedbaseurl,
             'repo': self.reponame,
             'encoding': encoding.encoding,
             'motd': motd,
             'sessionvars': sessionvars,
             'pathdef': makebreadcrumb(req.apppath),
             'style': style,
             'nonce': self.nonce,
         }
         tres = formatter.templateresources(self.repo.ui, self.repo)
         tmpl = templater.templater.frommapfile(mapfile,
                                                filters={'websub': websubfilter},
                                                defaults=defaults,
                                                resources=tres)
         return tmpl
 
 
 class hgweb(object):
     """HTTP server for individual repositories.
 
     Instances of this class serve HTTP responses for a particular
     repository.
 
     Instances are typically used as WSGI applications.
 
     Some servers are multi-threaded. On these servers, there may
     be multiple active threads inside __call__.
     """
     def __init__(self, repo, name=None, baseui=None):
         if isinstance(repo, str):
             if baseui:
                 u = baseui.copy()
             else:
                 u = uimod.ui.load()
             r = hg.repository(u, repo)
         else:
             # we trust caller to give us a private copy
             r = repo
 
         r.ui.setconfig('ui', 'report_untrusted', 'off', 'hgweb')
         r.baseui.setconfig('ui', 'report_untrusted', 'off', 'hgweb')
         r.ui.setconfig('ui', 'nontty', 'true', 'hgweb')
         r.baseui.setconfig('ui', 'nontty', 'true', 'hgweb')
         # resolve file patterns relative to repo root
         r.ui.setconfig('ui', 'forcecwd', r.root, 'hgweb')
         r.baseui.setconfig('ui', 'forcecwd', r.root, 'hgweb')
         # displaying bundling progress bar while serving feel wrong and may
         # break some wsgi implementation.
         r.ui.setconfig('progress', 'disable', 'true', 'hgweb')
         r.baseui.setconfig('progress', 'disable', 'true', 'hgweb')
         self._repos = [hg.cachedlocalrepo(self._webifyrepo(r))]
         self._lastrepo = self._repos[0]
         hook.redirect(True)
         self.reponame = name
 
     def _webifyrepo(self, repo):
         repo = getwebview(repo)
         self.websubtable = webutil.getwebsubs(repo)
         return repo
 
     @contextlib.contextmanager
     def _obtainrepo(self):
         """Obtain a repo unique to the caller.
 
         Internally we maintain a stack of cachedlocalrepo instances
         to be handed out. If one is available, we pop it and return it,
         ensuring it is up to date in the process. If one is not available,
         we clone the most recently used repo instance and return it.
 
         It is currently possible for the stack to grow without bounds
         if the server allows infinite threads. However, servers should
         have a thread limit, thus establishing our limit.
         """
         if self._repos:
             cached = self._repos.pop()
             r, created = cached.fetch()
         else:
             cached = self._lastrepo.copy()
             r, created = cached.fetch()
         if created:
             r = self._webifyrepo(r)
 
         self._lastrepo = cached
         self.mtime = cached.mtime
         try:
             yield r
         finally:
             self._repos.append(cached)
 
     def run(self):
         """Start a server from CGI environment.
 
         Modern servers should be using WSGI and should avoid this
         method, if possible.
         """
         if not encoding.environ.get('GATEWAY_INTERFACE',
                                     '').startswith("CGI/1."):
             raise RuntimeError("This function is only intended to be "
                                "called while running as a CGI script.")
         wsgicgi.launch(self)
 
     def __call__(self, env, respond):
         """Run the WSGI application.
 
         This may be called by multiple threads.
         """
         req = requestmod.wsgirequest(env, respond)
         return self.run_wsgi(req)
 
     def run_wsgi(self, wsgireq):
         """Internal method to run the WSGI application.
 
         This is typically only called by Mercurial. External consumers
         should be using instances of this class as the WSGI application.
         """
         with self._obtainrepo() as repo:
             profile = repo.ui.configbool('profiling', 'enabled')
             with profiling.profile(repo.ui, enabled=profile):
                 for r in self._runwsgi(wsgireq, repo):
                     yield r
 
     def _runwsgi(self, wsgireq, repo):
         req = wsgireq.req
         res = wsgireq.res
         rctx = requestcontext(self, repo, req, res)
 
         # This state is global across all threads.
         encoding.encoding = rctx.config('web', 'encoding')
         rctx.repo.ui.environ = wsgireq.env
 
         if rctx.csp:
             # hgwebdir may have added CSP header. Since we generate our own,
             # replace it.
             wsgireq.headers = [h for h in wsgireq.headers
                                if h[0] != 'Content-Security-Policy']
             wsgireq.headers.append(('Content-Security-Policy', rctx.csp))
             res.headers['Content-Security-Policy'] = rctx.csp
 
         handled = wireprotoserver.handlewsgirequest(
             rctx, wsgireq, req, res, self.check_perm)
         if handled:
             return res.sendresponse()
 
         if req.havepathinfo:
             query = req.dispatchpath
         else:
             query = req.querystring.partition('&')[0].partition(';')[0]
 
         # translate user-visible url structure to internal structure
 
         args = query.split('/', 2)
         if 'cmd' not in req.qsparams and args and args[0]:
             cmd = args.pop(0)
             style = cmd.rfind('-')
             if style != -1:
                 req.qsparams['style'] = cmd[:style]
                 cmd = cmd[style + 1:]
 
             # avoid accepting e.g. style parameter as command
             if util.safehasattr(webcommands, cmd):
                 req.qsparams['cmd'] = cmd
 
             if cmd == 'static':
                 req.qsparams['file'] = '/'.join(args)
             else:
                 if args and args[0]:
                     node = args.pop(0).replace('%2F', '/')
                     req.qsparams['node'] = node
                 if args:
                     if 'file' in req.qsparams:
                         del req.qsparams['file']
                     for a in args:
                         req.qsparams.add('file', a)
 
             ua = req.headers.get('User-Agent', '')
             if cmd == 'rev' and 'mercurial' in ua:
                 req.qsparams['style'] = 'raw'
 
             if cmd == 'archive':
                 fn = req.qsparams['node']
                 for type_, spec in rctx.archivespecs.iteritems():
                     ext = spec[2]
                     if fn.endswith(ext):
                         req.qsparams['node'] = fn[:-len(ext)]
                         req.qsparams['type'] = type_
         else:
             cmd = req.qsparams.get('cmd', '')
 
         # process the web interface request
 
         try:
             tmpl = rctx.templater(req)
             ctype = tmpl('mimetype', encoding=encoding.encoding)
             ctype = templater.stringify(ctype)
 
             # check read permissions non-static content
             if cmd != 'static':
                 self.check_perm(rctx, wsgireq, None)
 
             if cmd == '':
                 req.qsparams['cmd'] = tmpl.cache['default']
                 cmd = req.qsparams['cmd']
 
             # Don't enable caching if using a CSP nonce because then it wouldn't
             # be a nonce.
             if rctx.configbool('web', 'cache') and not rctx.nonce:
                 tag = 'W/"%d"' % self.mtime
                 if req.headers.get('If-None-Match') == tag:
                     raise ErrorResponse(HTTP_NOT_MODIFIED)
 
                 wsgireq.headers.append((r'ETag', pycompat.sysstr(tag)))
                 res.headers['ETag'] = tag
 
             if cmd not in webcommands.__all__:
                 msg = 'no such method: %s' % cmd
                 raise ErrorResponse(HTTP_BAD_REQUEST, msg)
             else:
                 # Set some globals appropriate for web handlers. Commands can
                 # override easily enough.
                 res.status = '200 Script output follows'
                 res.headers['Content-Type'] = ctype
                 content = getattr(webcommands, cmd)(rctx, wsgireq, tmpl)
 
                 if content is res:
                     return res.sendresponse()
-
-                wsgireq.respond(HTTP_OK, ctype)
-
-            return content
+                elif content is True:
+                    return []
+                else:
+                    wsgireq.respond(HTTP_OK, ctype)
+                    return content
 
         except (error.LookupError, error.RepoLookupError) as err:
             wsgireq.respond(HTTP_NOT_FOUND, ctype)
             msg = pycompat.bytestr(err)
             if (util.safehasattr(err, 'name') and
                 not isinstance(err,  error.ManifestLookupError)):
                 msg = 'revision not found: %s' % err.name
             return tmpl('error', error=msg)
         except (error.RepoError, error.RevlogError) as inst:
             wsgireq.respond(HTTP_SERVER_ERROR, ctype)
             return tmpl('error', error=pycompat.bytestr(inst))
         except ErrorResponse as inst:
             wsgireq.respond(inst, ctype)
             if inst.code == HTTP_NOT_MODIFIED:
                 # Not allowed to return a body on a 304
                 return ['']
             return tmpl('error', error=pycompat.bytestr(inst))
 
     def check_perm(self, rctx, req, op):
         for permhook in permhooks:
             permhook(rctx, req, op)
 
 def getwebview(repo):
     """The 'web.view' config controls changeset filter to hgweb. Possible
     values are ``served``, ``visible`` and ``all``. Default is ``served``.
     The ``served`` filter only shows changesets that can be pulled from the
     hgweb instance.  The``visible`` filter includes secret changesets but
     still excludes "hidden" one.
 
     See the repoview module for details.
 
     The option has been around undocumented since Mercurial 2.5, but no
     user ever asked about it. So we better keep it undocumented for now."""
     # experimental config: web.view
     viewconfig = repo.ui.config('web', 'view', untrusted=True)
     if viewconfig == 'all':
         return repo.unfiltered()
     elif viewconfig in repoview.filtertable:
         return repo.filtered(viewconfig)
     else:
         return repo.filtered('served')

mercurial/hgweb/request.py

@@ -1,581 +1,627 b''
 # hgweb/request.py - An http request from either CGI or the standalone server.
 #
 # Copyright 21 May 2005 - (c) 2005 Jake Edge <[e-mail]>
 # Copyright 2005, 2006 Matt Mackall <[e-mail]>
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 from __future__ import absolute_import
 
 import errno
 import socket
 import wsgiref.headers as wsgiheaders
 #import wsgiref.validate
 
 from .common import (
     ErrorResponse,
     HTTP_NOT_MODIFIED,
     statusmessage,
 )
 
 from ..thirdparty import (
     attr,
 )
 from .. import (
     error,
     pycompat,
     util,
 )
 
 class multidict(object):
     """A dict like object that can store multiple values for a key.
 
     Used to store parsed request parameters.
 
     This is inspired by WebOb's class of the same name.
     """
     def __init__(self):
         # Stores (key, value) 2-tuples. This isn't the most efficient. But we
         # don't rely on parameters that much, so it shouldn't be a perf issue.
         # we can always add dict for fast lookups.
         self._items = []
 
     def __getitem__(self, key):
         """Returns the last set value for a key."""
         for k, v in reversed(self._items):
             if k == key:
                 return v
 
         raise KeyError(key)
 
     def __setitem__(self, key, value):
         """Replace a values for a key with a new value."""
         try:
             del self[key]
         except KeyError:
             pass
 
         self._items.append((key, value))
 
     def __delitem__(self, key):
         """Delete all values for a key."""
         oldlen = len(self._items)
 
         self._items[:] = [(k, v) for k, v in self._items if k != key]
 
         if oldlen == len(self._items):
             raise KeyError(key)
 
     def __contains__(self, key):
         return any(k == key for k, v in self._items)
 
     def __len__(self):
         return len(self._items)
 
     def get(self, key, default=None):
         try:
             return self.__getitem__(key)
         except KeyError:
             return default
 
     def add(self, key, value):
         """Add a new value for a key. Does not replace existing values."""
         self._items.append((key, value))
 
     def getall(self, key):
         """Obtains all values for a key."""
         return [v for k, v in self._items if k == key]
 
     def getone(self, key):
         """Obtain a single value for a key.
 
         Raises KeyError if key not defined or it has multiple values set.
         """
         vals = self.getall(key)
 
         if not vals:
             raise KeyError(key)
 
         if len(vals) > 1:
             raise KeyError('multiple values for %r' % key)
 
         return vals[0]
 
     def asdictoflists(self):
         d = {}
         for k, v in self._items:
             if k in d:
                 d[k].append(v)
             else:
                 d[k] = [v]
 
         return d
 
 @attr.s(frozen=True)
 class parsedrequest(object):
     """Represents a parsed WSGI request.
 
     Contains both parsed parameters as well as a handle on the input stream.
     """
 
     # Request method.
     method = attr.ib()
     # Full URL for this request.
     url = attr.ib()
     # URL without any path components. Just <proto>://<host><port>.
     baseurl = attr.ib()
     # Advertised URL. Like ``url`` and ``baseurl`` but uses SERVER_NAME instead
     # of HTTP: Host header for hostname. This is likely what clients used.
     advertisedurl = attr.ib()
     advertisedbaseurl = attr.ib()
     # URL scheme (part before ``://``). e.g. ``http`` or ``https``.
     urlscheme = attr.ib()
     # Value of REMOTE_USER, if set, or None.
     remoteuser = attr.ib()
     # Value of REMOTE_HOST, if set, or None.
     remotehost = attr.ib()
     # WSGI application path.
     apppath = attr.ib()
     # List of path parts to be used for dispatch.
     dispatchparts = attr.ib()
     # URL path component (no query string) used for dispatch.
     dispatchpath = attr.ib()
     # Whether there is a path component to this request. This can be true
     # when ``dispatchpath`` is empty due to REPO_NAME muckery.
     havepathinfo = attr.ib()
     # The name of the repository being accessed.
     reponame = attr.ib()
     # Raw query string (part after "?" in URL).
     querystring = attr.ib()
     # multidict of query string parameters.
     qsparams = attr.ib()
     # wsgiref.headers.Headers instance. Operates like a dict with case
     # insensitive keys.
     headers = attr.ib()
     # Request body input stream.
     bodyfh = attr.ib()
 
 def parserequestfromenv(env, bodyfh):
     """Parse URL components from environment variables.
 
     WSGI defines request attributes via environment variables. This function
     parses the environment variables into a data structure.
     """
     # PEP-0333 defines the WSGI spec and is a useful reference for this code.
 
     # We first validate that the incoming object conforms with the WSGI spec.
     # We only want to be dealing with spec-conforming WSGI implementations.
     # TODO enable this once we fix internal violations.
     #wsgiref.validate.check_environ(env)
 
     # PEP-0333 states that environment keys and values are native strings
     # (bytes on Python 2 and str on Python 3). The code points for the Unicode
     # strings on Python 3 must be between \00000-\000FF. We deal with bytes
     # in Mercurial, so mass convert string keys and values to bytes.
     if pycompat.ispy3:
         env = {k.encode('latin-1'): v for k, v in env.iteritems()}
         env = {k: v.encode('latin-1') if isinstance(v, str) else v
                for k, v in env.iteritems()}
 
     # https://www.python.org/dev/peps/pep-0333/#environ-variables defines
     # the environment variables.
     # https://www.python.org/dev/peps/pep-0333/#url-reconstruction defines
     # how URLs are reconstructed.
     fullurl = env['wsgi.url_scheme'] + '://'
     advertisedfullurl = fullurl
 
     def addport(s):
         if env['wsgi.url_scheme'] == 'https':
             if env['SERVER_PORT'] != '443':
                 s += ':' + env['SERVER_PORT']
         else:
             if env['SERVER_PORT'] != '80':
                 s += ':' + env['SERVER_PORT']
 
         return s
 
     if env.get('HTTP_HOST'):
         fullurl += env['HTTP_HOST']
     else:
         fullurl += env['SERVER_NAME']
         fullurl = addport(fullurl)
 
     advertisedfullurl += env['SERVER_NAME']
     advertisedfullurl = addport(advertisedfullurl)
 
     baseurl = fullurl
     advertisedbaseurl = advertisedfullurl
 
     fullurl += util.urlreq.quote(env.get('SCRIPT_NAME', ''))
     advertisedfullurl += util.urlreq.quote(env.get('SCRIPT_NAME', ''))
     fullurl += util.urlreq.quote(env.get('PATH_INFO', ''))
     advertisedfullurl += util.urlreq.quote(env.get('PATH_INFO', ''))
 
     if env.get('QUERY_STRING'):
         fullurl += '?' + env['QUERY_STRING']
         advertisedfullurl += '?' + env['QUERY_STRING']
 
     # When dispatching requests, we look at the URL components (PATH_INFO
     # and QUERY_STRING) after the application root (SCRIPT_NAME). But hgwebdir
     # has the concept of "virtual" repositories. This is defined via REPO_NAME.
     # If REPO_NAME is defined, we append it to SCRIPT_NAME to form a new app
     # root. We also exclude its path components from PATH_INFO when resolving
     # the dispatch path.
 
     apppath = env['SCRIPT_NAME']
 
     if env.get('REPO_NAME'):
         if not apppath.endswith('/'):
             apppath += '/'
 
         apppath += env.get('REPO_NAME')
 
     if 'PATH_INFO' in env:
         dispatchparts = env['PATH_INFO'].strip('/').split('/')
 
         # Strip out repo parts.
         repoparts = env.get('REPO_NAME', '').split('/')
         if dispatchparts[:len(repoparts)] == repoparts:
             dispatchparts = dispatchparts[len(repoparts):]
     else:
         dispatchparts = []
 
     dispatchpath = '/'.join(dispatchparts)
 
     querystring = env.get('QUERY_STRING', '')
 
     # We store as a list so we have ordering information. We also store as
     # a dict to facilitate fast lookup.
     qsparams = multidict()
     for k, v in util.urlreq.parseqsl(querystring, keep_blank_values=True):
         qsparams.add(k, v)
 
     # HTTP_* keys contain HTTP request headers. The Headers structure should
     # perform case normalization for us. We just rewrite underscore to dash
     # so keys match what likely went over the wire.
     headers = []
     for k, v in env.iteritems():
         if k.startswith('HTTP_'):
             headers.append((k[len('HTTP_'):].replace('_', '-'), v))
 
     headers = wsgiheaders.Headers(headers)
 
     # This is kind of a lie because the HTTP header wasn't explicitly
     # sent. But for all intents and purposes it should be OK to lie about
     # this, since a consumer will either either value to determine how many
     # bytes are available to read.
     if 'CONTENT_LENGTH' in env and 'HTTP_CONTENT_LENGTH' not in env:
         headers['Content-Length'] = env['CONTENT_LENGTH']
 
     # TODO do this once we remove wsgirequest.inp, otherwise we could have
     # multiple readers from the underlying input stream.
     #bodyfh = env['wsgi.input']
     #if 'Content-Length' in headers:
     #    bodyfh = util.cappedreader(bodyfh, int(headers['Content-Length']))
 
     return parsedrequest(method=env['REQUEST_METHOD'],
                          url=fullurl, baseurl=baseurl,
                          advertisedurl=advertisedfullurl,
                          advertisedbaseurl=advertisedbaseurl,
                          urlscheme=env['wsgi.url_scheme'],
                          remoteuser=env.get('REMOTE_USER'),
                          remotehost=env.get('REMOTE_HOST'),
                          apppath=apppath,
                          dispatchparts=dispatchparts, dispatchpath=dispatchpath,
                          havepathinfo='PATH_INFO' in env,
                          reponame=env.get('REPO_NAME'),
                          querystring=querystring,
                          qsparams=qsparams,
                          headers=headers,
                          bodyfh=bodyfh)
 
 class offsettrackingwriter(object):
     """A file object like object that is append only and tracks write count.
 
     Instances are bound to a callable. This callable is called with data
     whenever a ``write()`` is attempted.
 
     Instances track the amount of written data so they can answer ``tell()``
     requests.
 
     The intent of this class is to wrap the ``write()`` function returned by
     a WSGI ``start_response()`` function. Since ``write()`` is a callable and
     not a file object, it doesn't implement other file object methods.
     """
     def __init__(self, writefn):
         self._write = writefn
         self._offset = 0
 
     def write(self, s):
         res = self._write(s)
         # Some Python objects don't report the number of bytes written.
         if res is None:
             self._offset += len(s)
         else:
             self._offset += res
 
     def flush(self):
         pass
 
     def tell(self):
         return self._offset
 
 class wsgiresponse(object):
     """Represents a response to a WSGI request.
 
     A response consists of a status line, headers, and a body.
 
     Consumers must populate the ``status`` and ``headers`` fields and
     make a call to a ``setbody*()`` method before the response can be
     issued.
 
     When it is time to start sending the response over the wire,
     ``sendresponse()`` is called. It handles emitting the header portion
     of the response message. It then yields chunks of body data to be
     written to the peer. Typically, the WSGI application itself calls
     and returns the value from ``sendresponse()``.
     """
 
     def __init__(self, req, startresponse):
         """Create an empty response tied to a specific request.
 
         ``req`` is a ``parsedrequest``. ``startresponse`` is the
         ``start_response`` function passed to the WSGI application.
         """
         self._req = req
         self._startresponse = startresponse
 
         self.status = None
         self.headers = wsgiheaders.Headers([])
 
         self._bodybytes = None
         self._bodygen = None
+        self._bodywillwrite = False
         self._started = False
+        self._bodywritefn = None
+
+    def _verifybody(self):
+        if (self._bodybytes is not None or self._bodygen is not None
+            or self._bodywillwrite):
+            raise error.ProgrammingError('cannot define body multiple times')
 
     def setbodybytes(self, b):
         """Define the response body as static bytes."""
-        if self._bodybytes is not None or self._bodygen is not None:
-            raise error.ProgrammingError('cannot define body multiple times')
-
+        self._verifybody()
         self._bodybytes = b
         self.headers['Content-Length'] = '%d' % len(b)
 
     def setbodygen(self, gen):
         """Define the response body as a generator of bytes."""
-        if self._bodybytes is not None or self._bodygen is not None:
-            raise error.ProgrammingError('cannot define body multiple times')
+        self._verifybody()
+        self._bodygen = gen
+
+    def setbodywillwrite(self):
+        """Signal an intent to use write() to emit the response body.
+
+        **This is the least preferred way to send a body.**
 
-        self._bodygen = gen
+        It is preferred for WSGI applications to emit a generator of chunks
+        constituting the response body. However, some consumers can't emit
+        data this way. So, WSGI provides a way to obtain a ``write(data)``
+        function that can be used to synchronously perform an unbuffered
+        write.
+
+        Calling this function signals an intent to produce the body in this
+        manner.
+        """
+        self._verifybody()
+        self._bodywillwrite = True
 
     def sendresponse(self):
         """Send the generated response to the client.
 
         Before this is called, ``status`` must be set and one of
         ``setbodybytes()`` or ``setbodygen()`` must be called.
 
         Calling this method multiple times is not allowed.
         """
         if self._started:
             raise error.ProgrammingError('sendresponse() called multiple times')
 
         self._started = True
 
         if not self.status:
             raise error.ProgrammingError('status line not defined')
 
-        if self._bodybytes is None and self._bodygen is None:
+        if (self._bodybytes is None and self._bodygen is None
+            and not self._bodywillwrite):
             raise error.ProgrammingError('response body not defined')
 
         # Various HTTP clients (notably httplib) won't read the HTTP response
         # until the HTTP request has been sent in full. If servers (us) send a
         # response before the HTTP request has been fully sent, the connection
         # may deadlock because neither end is reading.
         #
         # We work around this by "draining" the request data before
         # sending any response in some conditions.
         drain = False
         close = False
 
         # If the client sent Expect: 100-continue, we assume it is smart enough
         # to deal with the server sending a response before reading the request.
         # (httplib doesn't do this.)
         if self._req.headers.get('Expect', '').lower() == '100-continue':
             pass
         # Only tend to request methods that have bodies. Strictly speaking,
         # we should sniff for a body. But this is fine for our existing
         # WSGI applications.
         elif self._req.method not in ('POST', 'PUT'):
             pass
         else:
             # If we don't know how much data to read, there's no guarantee
             # that we can drain the request responsibly. The WSGI
             # specification only says that servers *should* ensure the
             # input stream doesn't overrun the actual request. So there's
             # no guarantee that reading until EOF won't corrupt the stream
             # state.
             if not isinstance(self._req.bodyfh, util.cappedreader):
                 close = True
             else:
                 # We /could/ only drain certain HTTP response codes. But 200 and
                 # non-200 wire protocol responses both require draining. Since
                 # we have a capped reader in place for all situations where we
                 # drain, it is safe to read from that stream. We'll either do
                 # a drain or no-op if we're already at EOF.
                 drain = True
 
         if close:
             self.headers['Connection'] = 'Close'
 
         if drain:
             assert isinstance(self._req.bodyfh, util.cappedreader)
             while True:
                 chunk = self._req.bodyfh.read(32768)
                 if not chunk:
                     break
 
-        self._startresponse(pycompat.sysstr(self.status), self.headers.items())
+        write = self._startresponse(pycompat.sysstr(self.status),
+                                    self.headers.items())
+
         if self._bodybytes:
             yield self._bodybytes
         elif self._bodygen:
             for chunk in self._bodygen:
                 yield chunk
+        elif self._bodywillwrite:
+            self._bodywritefn = write
         else:
             error.ProgrammingError('do not know how to send body')
 
+    def getbodyfile(self):
+        """Obtain a file object like object representing the response body.
+
+        For this to work, you must call ``setbodywillwrite()`` and then
+        ``sendresponse()`` first. ``sendresponse()`` is a generator and the
+        function won't run to completion unless the generator is advanced. The
+        generator yields not items. The easiest way to consume it is with
+        ``list(res.sendresponse())``, which should resolve to an empty list -
+        ``[]``.
+        """
+        if not self._bodywillwrite:
+            raise error.ProgrammingError('must call setbodywillwrite() first')
+
+        if not self._started:
+            raise error.ProgrammingError('must call sendresponse() first; did '
+                                         'you remember to consume it since it '
+                                         'is a generator?')
+
+        assert self._bodywritefn
+        return offsettrackingwriter(self._bodywritefn)
+
 class wsgirequest(object):
     """Higher-level API for a WSGI request.
 
     WSGI applications are invoked with 2 arguments. They are used to
     instantiate instances of this class, which provides higher-level APIs
     for obtaining request parameters, writing HTTP output, etc.
     """
     def __init__(self, wsgienv, start_response):
         version = wsgienv[r'wsgi.version']
         if (version < (1, 0)) or (version >= (2, 0)):
             raise RuntimeError("Unknown and unsupported WSGI version %d.%d"
                                % version)
 
         inp = wsgienv[r'wsgi.input']
 
         if r'HTTP_CONTENT_LENGTH' in wsgienv:
             inp = util.cappedreader(inp, int(wsgienv[r'HTTP_CONTENT_LENGTH']))
         elif r'CONTENT_LENGTH' in wsgienv:
             inp = util.cappedreader(inp, int(wsgienv[r'CONTENT_LENGTH']))
 
         self.err = wsgienv[r'wsgi.errors']
         self.threaded = wsgienv[r'wsgi.multithread']
         self.multiprocess = wsgienv[r'wsgi.multiprocess']
         self.run_once = wsgienv[r'wsgi.run_once']
         self.env = wsgienv
         self.req = parserequestfromenv(wsgienv, inp)
         self.res = wsgiresponse(self.req, start_response)
         self._start_response = start_response
         self.server_write = None
         self.headers = []
 
     def respond(self, status, type, filename=None, body=None):
         if not isinstance(type, str):
             type = pycompat.sysstr(type)
         if self._start_response is not None:
             self.headers.append((r'Content-Type', type))
             if filename:
                 filename = (filename.rpartition('/')[-1]
                             .replace('\\', '\\\\').replace('"', '\\"'))
                 self.headers.append(('Content-Disposition',
                                      'inline; filename="%s"' % filename))
             if body is not None:
                 self.headers.append((r'Content-Length', str(len(body))))
 
             for k, v in self.headers:
                 if not isinstance(v, str):
                     raise TypeError('header value must be string: %r' % (v,))
 
             if isinstance(status, ErrorResponse):
                 self.headers.extend(status.headers)
                 if status.code == HTTP_NOT_MODIFIED:
                     # RFC 2616 Section 10.3.5: 304 Not Modified has cases where
                     # it MUST NOT include any headers other than these and no
                     # body
                     self.headers = [(k, v) for (k, v) in self.headers if
                                     k in ('Date', 'ETag', 'Expires',
                                           'Cache-Control', 'Vary')]
                 status = statusmessage(status.code, pycompat.bytestr(status))
             elif status == 200:
                 status = '200 Script output follows'
             elif isinstance(status, int):
                 status = statusmessage(status)
 
             # Various HTTP clients (notably httplib) won't read the HTTP
             # response until the HTTP request has been sent in full. If servers
             # (us) send a response before the HTTP request has been fully sent,
             # the connection may deadlock because neither end is reading.
             #
             # We work around this by "draining" the request data before
             # sending any response in some conditions.
             drain = False
             close = False
 
             # If the client sent Expect: 100-continue, we assume it is smart
             # enough to deal with the server sending a response before reading
             # the request. (httplib doesn't do this.)
             if self.env.get(r'HTTP_EXPECT', r'').lower() == r'100-continue':
                 pass
             # Only tend to request methods that have bodies. Strictly speaking,
             # we should sniff for a body. But this is fine for our existing
             # WSGI applications.
             elif self.env[r'REQUEST_METHOD'] not in (r'POST', r'PUT'):
                 pass
             else:
                 # If we don't know how much data to read, there's no guarantee
                 # that we can drain the request responsibly. The WSGI
                 # specification only says that servers *should* ensure the
                 # input stream doesn't overrun the actual request. So there's
                 # no guarantee that reading until EOF won't corrupt the stream
                 # state.
                 if not isinstance(self.req.bodyfh, util.cappedreader):
                     close = True
                 else:
                     # We /could/ only drain certain HTTP response codes. But 200
                     # and non-200 wire protocol responses both require draining.
                     # Since we have a capped reader in place for all situations
                     # where we drain, it is safe to read from that stream. We'll
                     # either do a drain or no-op if we're already at EOF.
                     drain = True
 
             if close:
                 self.headers.append((r'Connection', r'Close'))
 
             if drain:
                 assert isinstance(self.req.bodyfh, util.cappedreader)
                 while True:
                     chunk = self.req.bodyfh.read(32768)
                     if not chunk:
                         break
 
             self.server_write = self._start_response(
                 pycompat.sysstr(status), self.headers)
             self._start_response = None
             self.headers = []
         if body is not None:
             self.write(body)
             self.server_write = None
 
     def write(self, thing):
         if thing:
             try:
                 self.server_write(thing)
             except socket.error as inst:
                 if inst[0] != errno.ECONNRESET:
                     raise
 
     def flush(self):
         return None
 
 def wsgiapplication(app_maker):
     '''For compatibility with old CGI scripts. A plain hgweb() or hgwebdir()
     can and should now be used as a WSGI application.'''
     application = app_maker()
     def run_wsgi(env, respond):
         return application(env, respond)
     return run_wsgi

mercurial/hgweb/webcommands.py

@@ -1,1511 +1,1512 b''
 #
 # Copyright 21 May 2005 - (c) 2005 Jake Edge <[e-mail]>
 # Copyright 2005-2007 Matt Mackall <[e-mail]>
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 from __future__ import absolute_import
 
 import copy
 import mimetypes
 import os
 import re
 
 from ..i18n import _
 from ..node import hex, nullid, short
 
 from .common import (
     ErrorResponse,
     HTTP_FORBIDDEN,
     HTTP_NOT_FOUND,
-    HTTP_OK,
     get_contact,
     paritygen,
     staticfile,
 )
-from . import (
-    request as requestmod,
-)
 
 from .. import (
     archival,
     dagop,
     encoding,
     error,
     graphmod,
     pycompat,
     revset,
     revsetlang,
     scmutil,
     smartset,
     templater,
     util,
 )
 
 from . import (
     webutil,
 )
 
 __all__ = []
 commands = {}
 
 class webcommand(object):
     """Decorator used to register a web command handler.
 
     The decorator takes as its positional arguments the name/path the
     command should be accessible under.
 
     When called, functions receive as arguments a ``requestcontext``,
     ``wsgirequest``, and a templater instance for generatoring output.
     The functions should populate the ``rctx.res`` object with details
     about the HTTP response.
 
     The function can return the ``requestcontext.res`` instance to signal
     that it wants to use this object to generate the response. If an iterable
     is returned, the ``wsgirequest`` instance will be used and the returned
-    content will constitute the response body.
+    content will constitute the response body. ``True`` can be returned to
+    indicate that the function already sent output and the caller doesn't
+    need to do anything more to send the response.
 
     Usage:
 
     @webcommand('mycommand')
     def mycommand(web, req, tmpl):
         pass
     """
 
     def __init__(self, name):
         self.name = name
 
     def __call__(self, func):
         __all__.append(self.name)
         commands[self.name] = func
         return func
 
 @webcommand('log')
 def log(web, req, tmpl):
     """
     /log[/{revision}[/{path}]]
     --------------------------
 
     Show repository or file history.
 
     For URLs of the form ``/log/{revision}``, a list of changesets starting at
     the specified changeset identifier is shown. If ``{revision}`` is not
     defined, the default is ``tip``. This form is equivalent to the
     ``changelog`` handler.
 
     For URLs of the form ``/log/{revision}/{file}``, the history for a specific
     file will be shown. This form is equivalent to the ``filelog`` handler.
     """
 
     if req.req.qsparams.get('file'):
         return filelog(web, req, tmpl)
     else:
         return changelog(web, req, tmpl)
 
 @webcommand('rawfile')
 def rawfile(web, req, tmpl):
     guessmime = web.configbool('web', 'guessmime')
 
     path = webutil.cleanpath(web.repo, req.req.qsparams.get('file', ''))
     if not path:
         return manifest(web, req, tmpl)
 
     try:
         fctx = webutil.filectx(web.repo, req)
     except error.LookupError as inst:
         try:
             return manifest(web, req, tmpl)
         except ErrorResponse:
             raise inst
 
     path = fctx.path()
     text = fctx.data()
     mt = 'application/binary'
     if guessmime:
         mt = mimetypes.guess_type(path)[0]
         if mt is None:
             if util.binary(text):
                 mt = 'application/binary'
             else:
                 mt = 'text/plain'
     if mt.startswith('text/'):
         mt += '; charset="%s"' % encoding.encoding
 
     web.res.headers['Content-Type'] = mt
     filename = (path.rpartition('/')[-1]
                 .replace('\\', '\\\\').replace('"', '\\"'))
     web.res.headers['Content-Disposition'] = 'inline; filename="%s"' % filename
     web.res.setbodybytes(text)
     return web.res
 
 def _filerevision(web, req, tmpl, fctx):
     f = fctx.path()
     text = fctx.data()
     parity = paritygen(web.stripecount)
     ishead = fctx.filerev() in fctx.filelog().headrevs()
 
     if util.binary(text):
         mt = mimetypes.guess_type(f)[0] or 'application/octet-stream'
         text = '(binary:%s)' % mt
 
     def lines():
         for lineno, t in enumerate(text.splitlines(True)):
             yield {"line": t,
                    "lineid": "l%d" % (lineno + 1),
                    "linenumber": "% 6d" % (lineno + 1),
                    "parity": next(parity)}
 
     web.res.setbodygen(tmpl(
         'filerevision',
         file=f,
         path=webutil.up(f),
         text=lines(),
         symrev=webutil.symrevorshortnode(req, fctx),
         rename=webutil.renamelink(fctx),
         permissions=fctx.manifest().flags(f),
         ishead=int(ishead),
         **pycompat.strkwargs(webutil.commonentry(web.repo, fctx))))
 
     return web.res
 
 @webcommand('file')
 def file(web, req, tmpl):
     """
     /file/{revision}[/{path}]
     -------------------------
 
     Show information about a directory or file in the repository.
 
     Info about the ``path`` given as a URL parameter will be rendered.
 
     If ``path`` is a directory, information about the entries in that
     directory will be rendered. This form is equivalent to the ``manifest``
     handler.
 
     If ``path`` is a file, information about that file will be shown via
     the ``filerevision`` template.
 
     If ``path`` is not defined, information about the root directory will
     be rendered.
     """
     if web.req.qsparams.get('style') == 'raw':
         return rawfile(web, req, tmpl)
 
     path = webutil.cleanpath(web.repo, req.req.qsparams.get('file', ''))
     if not path:
         return manifest(web, req, tmpl)
     try:
         return _filerevision(web, req, tmpl, webutil.filectx(web.repo, req))
     except error.LookupError as inst:
         try:
             return manifest(web, req, tmpl)
         except ErrorResponse:
             raise inst
 
 def _search(web, req, tmpl):
     MODE_REVISION = 'rev'
     MODE_KEYWORD = 'keyword'
     MODE_REVSET = 'revset'
 
     def revsearch(ctx):
         yield ctx
 
     def keywordsearch(query):
         lower = encoding.lower
         qw = lower(query).split()
 
         def revgen():
             cl = web.repo.changelog
             for i in xrange(len(web.repo) - 1, 0, -100):
                 l = []
                 for j in cl.revs(max(0, i - 99), i):
                     ctx = web.repo[j]
                     l.append(ctx)
                 l.reverse()
                 for e in l:
                     yield e
 
         for ctx in revgen():
             miss = 0
             for q in qw:
                 if not (q in lower(ctx.user()) or
                         q in lower(ctx.description()) or
                         q in lower(" ".join(ctx.files()))):
                     miss = 1
                     break
             if miss:
                 continue
 
             yield ctx
 
     def revsetsearch(revs):
         for r in revs:
             yield web.repo[r]
 
     searchfuncs = {
         MODE_REVISION: (revsearch, 'exact revision search'),
         MODE_KEYWORD: (keywordsearch, 'literal keyword search'),
         MODE_REVSET: (revsetsearch, 'revset expression search'),
     }
 
     def getsearchmode(query):
         try:
             ctx = web.repo[query]
         except (error.RepoError, error.LookupError):
             # query is not an exact revision pointer, need to
             # decide if it's a revset expression or keywords
             pass
         else:
             return MODE_REVISION, ctx
 
         revdef = 'reverse(%s)' % query
         try:
             tree = revsetlang.parse(revdef)
         except error.ParseError:
             # can't parse to a revset tree
             return MODE_KEYWORD, query
 
         if revsetlang.depth(tree) <= 2:
             # no revset syntax used
             return MODE_KEYWORD, query
 
         if any((token, (value or '')[:3]) == ('string', 're:')
                for token, value, pos in revsetlang.tokenize(revdef)):
             return MODE_KEYWORD, query
 
         funcsused = revsetlang.funcsused(tree)
         if not funcsused.issubset(revset.safesymbols):
             return MODE_KEYWORD, query
 
         mfunc = revset.match(web.repo.ui, revdef, repo=web.repo)
         try:
             revs = mfunc(web.repo)
             return MODE_REVSET, revs
             # ParseError: wrongly placed tokens, wrongs arguments, etc
             # RepoLookupError: no such revision, e.g. in 'revision:'
             # Abort: bookmark/tag not exists
             # LookupError: ambiguous identifier, e.g. in '(bc)' on a large repo
         except (error.ParseError, error.RepoLookupError, error.Abort,
                 LookupError):
             return MODE_KEYWORD, query
 
     def changelist(**map):
         count = 0
 
         for ctx in searchfunc[0](funcarg):
             count += 1
             n = ctx.node()
             showtags = webutil.showtag(web.repo, tmpl, 'changelogtag', n)
             files = webutil.listfilediffs(tmpl, ctx.files(), n, web.maxfiles)
 
             yield tmpl('searchentry',
                        parity=next(parity),
                        changelogtag=showtags,
                        files=files,
                        **pycompat.strkwargs(webutil.commonentry(web.repo, ctx)))
 
             if count >= revcount:
                 break
 
     query = req.req.qsparams['rev']
     revcount = web.maxchanges
     if 'revcount' in req.req.qsparams:
         try:
             revcount = int(req.req.qsparams.get('revcount', revcount))
             revcount = max(revcount, 1)
             tmpl.defaults['sessionvars']['revcount'] = revcount
         except ValueError:
             pass
 
     lessvars = copy.copy(tmpl.defaults['sessionvars'])
     lessvars['revcount'] = max(revcount // 2, 1)
     lessvars['rev'] = query
     morevars = copy.copy(tmpl.defaults['sessionvars'])
     morevars['revcount'] = revcount * 2
     morevars['rev'] = query
 
     mode, funcarg = getsearchmode(query)
 
     if 'forcekw' in req.req.qsparams:
         showforcekw = ''
         showunforcekw = searchfuncs[mode][1]
         mode = MODE_KEYWORD
         funcarg = query
     else:
         if mode != MODE_KEYWORD:
             showforcekw = searchfuncs[MODE_KEYWORD][1]
         else:
             showforcekw = ''
         showunforcekw = ''
 
     searchfunc = searchfuncs[mode]
 
     tip = web.repo['tip']
     parity = paritygen(web.stripecount)
 
     web.res.setbodygen(tmpl(
         'search',
         query=query,
         node=tip.hex(),
         symrev='tip',
         entries=changelist,
         archives=web.archivelist('tip'),
         morevars=morevars,
         lessvars=lessvars,
         modedesc=searchfunc[1],
         showforcekw=showforcekw,
         showunforcekw=showunforcekw))
 
     return web.res
 
 @webcommand('changelog')
 def changelog(web, req, tmpl, shortlog=False):
     """
     /changelog[/{revision}]
     -----------------------
 
     Show information about multiple changesets.
 
     If the optional ``revision`` URL argument is absent, information about
     all changesets starting at ``tip`` will be rendered. If the ``revision``
     argument is present, changesets will be shown starting from the specified
     revision.
 
     If ``revision`` is absent, the ``rev`` query string argument may be
     defined. This will perform a search for changesets.
 
     The argument for ``rev`` can be a single revision, a revision set,
     or a literal keyword to search for in changeset data (equivalent to
     :hg:`log -k`).
 
     The ``revcount`` query string argument defines the maximum numbers of
     changesets to render.
 
     For non-searches, the ``changelog`` template will be rendered.
     """
 
     query = ''
     if 'node' in req.req.qsparams:
         ctx = webutil.changectx(web.repo, req)
         symrev = webutil.symrevorshortnode(req, ctx)
     elif 'rev' in req.req.qsparams:
         return _search(web, req, tmpl)
     else:
         ctx = web.repo['tip']
         symrev = 'tip'
 
     def changelist():
         revs = []
         if pos != -1:
             revs = web.repo.changelog.revs(pos, 0)
         curcount = 0
         for rev in revs:
             curcount += 1
             if curcount > revcount + 1:
                 break
 
             entry = webutil.changelistentry(web, web.repo[rev], tmpl)
             entry['parity'] = next(parity)
             yield entry
 
     if shortlog:
         revcount = web.maxshortchanges
     else:
         revcount = web.maxchanges
 
     if 'revcount' in req.req.qsparams:
         try:
             revcount = int(req.req.qsparams.get('revcount', revcount))
             revcount = max(revcount, 1)
             tmpl.defaults['sessionvars']['revcount'] = revcount
         except ValueError:
             pass
 
     lessvars = copy.copy(tmpl.defaults['sessionvars'])
     lessvars['revcount'] = max(revcount // 2, 1)
     morevars = copy.copy(tmpl.defaults['sessionvars'])
     morevars['revcount'] = revcount * 2
 
     count = len(web.repo)
     pos = ctx.rev()
     parity = paritygen(web.stripecount)
 
     changenav = webutil.revnav(web.repo).gen(pos, revcount, count)
 
     entries = list(changelist())
     latestentry = entries[:1]
     if len(entries) > revcount:
         nextentry = entries[-1:]
         entries = entries[:-1]
     else:
         nextentry = []
 
     web.res.setbodygen(tmpl(
         'shortlog' if shortlog else 'changelog',
         changenav=changenav,
         node=ctx.hex(),
         rev=pos,
         symrev=symrev,
         changesets=count,
         entries=entries,
         latestentry=latestentry,
         nextentry=nextentry,
         archives=web.archivelist('tip'),
         revcount=revcount,
         morevars=morevars,
         lessvars=lessvars,
         query=query))
 
     return web.res
 
 @webcommand('shortlog')
 def shortlog(web, req, tmpl):
     """
     /shortlog
     ---------
 
     Show basic information about a set of changesets.
 
     This accepts the same parameters as the ``changelog`` handler. The only
     difference is the ``shortlog`` template will be rendered instead of the
     ``changelog`` template.
     """
     return changelog(web, req, tmpl, shortlog=True)
 
 @webcommand('changeset')
 def changeset(web, req, tmpl):
     """
     /changeset[/{revision}]
     -----------------------
 
     Show information about a single changeset.
 
     A URL path argument is the changeset identifier to show. See ``hg help
     revisions`` for possible values. If not defined, the ``tip`` changeset
     will be shown.
 
     The ``changeset`` template is rendered. Contents of the ``changesettag``,
     ``changesetbookmark``, ``filenodelink``, ``filenolink``, and the many
     templates related to diffs may all be used to produce the output.
     """
     ctx = webutil.changectx(web.repo, req)
     web.res.setbodygen(tmpl('changeset',
                             **webutil.changesetentry(web, req, tmpl, ctx)))
     return web.res
 
 rev = webcommand('rev')(changeset)
 
 def decodepath(path):
     """Hook for mapping a path in the repository to a path in the
     working copy.
 
     Extensions (e.g., largefiles) can override this to remap files in
     the virtual file system presented by the manifest command below."""
     return path
 
 @webcommand('manifest')
 def manifest(web, req, tmpl):
     """
     /manifest[/{revision}[/{path}]]
     -------------------------------
 
     Show information about a directory.
 
     If the URL path arguments are omitted, information about the root
     directory for the ``tip`` changeset will be shown.
 
     Because this handler can only show information for directories, it
     is recommended to use the ``file`` handler instead, as it can handle both
     directories and files.
 
     The ``manifest`` template will be rendered for this handler.
     """
     if 'node' in req.req.qsparams:
         ctx = webutil.changectx(web.repo, req)
         symrev = webutil.symrevorshortnode(req, ctx)
     else:
         ctx = web.repo['tip']
         symrev = 'tip'
     path = webutil.cleanpath(web.repo, req.req.qsparams.get('file', ''))
     mf = ctx.manifest()
     node = ctx.node()
 
     files = {}
     dirs = {}
     parity = paritygen(web.stripecount)
 
     if path and path[-1:] != "/":
         path += "/"
     l = len(path)
     abspath = "/" + path
 
     for full, n in mf.iteritems():
         # the virtual path (working copy path) used for the full
         # (repository) path
         f = decodepath(full)
 
         if f[:l] != path:
             continue
         remain = f[l:]
         elements = remain.split('/')
         if len(elements) == 1:
             files[remain] = full
         else:
             h = dirs # need to retain ref to dirs (root)
             for elem in elements[0:-1]:
                 if elem not in h:
                     h[elem] = {}
                 h = h[elem]
                 if len(h) > 1:
                     break
             h[None] = None # denotes files present
 
     if mf and not files and not dirs:
         raise ErrorResponse(HTTP_NOT_FOUND, 'path not found: ' + path)
 
     def filelist(**map):
         for f in sorted(files):
             full = files[f]
 
             fctx = ctx.filectx(full)
             yield {"file": full,
                    "parity": next(parity),
                    "basename": f,
                    "date": fctx.date(),
                    "size": fctx.size(),
                    "permissions": mf.flags(full)}
 
     def dirlist(**map):
         for d in sorted(dirs):
 
             emptydirs = []
             h = dirs[d]
             while isinstance(h, dict) and len(h) == 1:
                 k, v = next(iter(h.items()))
                 if v:
                     emptydirs.append(k)
                 h = v
 
             path = "%s%s" % (abspath, d)
             yield {"parity": next(parity),
                    "path": path,
                    "emptydirs": "/".join(emptydirs),
                    "basename": d}
 
     web.res.setbodygen(tmpl(
         'manifest',
         symrev=symrev,
         path=abspath,
         up=webutil.up(abspath),
         upparity=next(parity),
         fentries=filelist,
         dentries=dirlist,
         archives=web.archivelist(hex(node)),
         **pycompat.strkwargs(webutil.commonentry(web.repo, ctx))))
 
     return web.res
 
 @webcommand('tags')
 def tags(web, req, tmpl):
     """
     /tags
     -----
 
     Show information about tags.
 
     No arguments are accepted.
 
     The ``tags`` template is rendered.
     """
     i = list(reversed(web.repo.tagslist()))
     parity = paritygen(web.stripecount)
 
     def entries(notip, latestonly, **map):
         t = i
         if notip:
             t = [(k, n) for k, n in i if k != "tip"]
         if latestonly:
             t = t[:1]
         for k, n in t:
             yield {"parity": next(parity),
                    "tag": k,
                    "date": web.repo[n].date(),
                    "node": hex(n)}
 
     web.res.setbodygen(tmpl(
         'tags',
         node=hex(web.repo.changelog.tip()),
         entries=lambda **x: entries(False, False, **x),
         entriesnotip=lambda **x: entries(True, False, **x),
         latestentry=lambda **x: entries(True, True, **x)))
 
     return web.res
 
 @webcommand('bookmarks')
 def bookmarks(web, req, tmpl):
     """
     /bookmarks
     ----------
 
     Show information about bookmarks.
 
     No arguments are accepted.
 
     The ``bookmarks`` template is rendered.
     """
     i = [b for b in web.repo._bookmarks.items() if b[1] in web.repo]
     sortkey = lambda b: (web.repo[b[1]].rev(), b[0])
     i = sorted(i, key=sortkey, reverse=True)
     parity = paritygen(web.stripecount)
 
     def entries(latestonly, **map):
         t = i
         if latestonly:
             t = i[:1]
         for k, n in t:
             yield {"parity": next(parity),
                    "bookmark": k,
                    "date": web.repo[n].date(),
                    "node": hex(n)}
 
     if i:
         latestrev = i[0][1]
     else:
         latestrev = -1
 
     web.res.setbodygen(tmpl(
         'bookmarks',
         node=hex(web.repo.changelog.tip()),
         lastchange=[{'date': web.repo[latestrev].date()}],
         entries=lambda **x: entries(latestonly=False, **x),
         latestentry=lambda **x: entries(latestonly=True, **x)))
 
     return web.res
 
 @webcommand('branches')
 def branches(web, req, tmpl):
     """
     /branches
     ---------
 
     Show information about branches.
 
     All known branches are contained in the output, even closed branches.
 
     No arguments are accepted.
 
     The ``branches`` template is rendered.
     """
     entries = webutil.branchentries(web.repo, web.stripecount)
     latestentry = webutil.branchentries(web.repo, web.stripecount, 1)
 
     web.res.setbodygen(tmpl(
         'branches',
         node=hex(web.repo.changelog.tip()),
         entries=entries,
         latestentry=latestentry))
 
     return web.res
 
 @webcommand('summary')
 def summary(web, req, tmpl):
     """
     /summary
     --------
 
     Show a summary of repository state.
 
     Information about the latest changesets, bookmarks, tags, and branches
     is captured by this handler.
 
     The ``summary`` template is rendered.
     """
     i = reversed(web.repo.tagslist())
 
     def tagentries(**map):
         parity = paritygen(web.stripecount)
         count = 0
         for k, n in i:
             if k == "tip": # skip tip
                 continue
 
             count += 1
             if count > 10: # limit to 10 tags
                 break
 
             yield tmpl("tagentry",
                        parity=next(parity),
                        tag=k,
                        node=hex(n),
                        date=web.repo[n].date())
 
     def bookmarks(**map):
         parity = paritygen(web.stripecount)
         marks = [b for b in web.repo._bookmarks.items() if b[1] in web.repo]
         sortkey = lambda b: (web.repo[b[1]].rev(), b[0])
         marks = sorted(marks, key=sortkey, reverse=True)
         for k, n in marks[:10]:  # limit to 10 bookmarks
             yield {'parity': next(parity),
                    'bookmark': k,
                    'date': web.repo[n].date(),
                    'node': hex(n)}
 
     def changelist(**map):
         parity = paritygen(web.stripecount, offset=start - end)
         l = [] # build a list in forward order for efficiency
         revs = []
         if start < end:
             revs = web.repo.changelog.revs(start, end - 1)
         for i in revs:
             ctx = web.repo[i]
 
             l.append(tmpl(
                 'shortlogentry',
                 parity=next(parity),
                 **pycompat.strkwargs(webutil.commonentry(web.repo, ctx))))
 
         for entry in reversed(l):
             yield entry
 
     tip = web.repo['tip']
     count = len(web.repo)
     start = max(0, count - web.maxchanges)
     end = min(count, start + web.maxchanges)
 
     desc = web.config("web", "description")
     if not desc:
         desc = 'unknown'
 
     web.res.setbodygen(tmpl(
         'summary',
         desc=desc,
         owner=get_contact(web.config) or 'unknown',
         lastchange=tip.date(),
         tags=tagentries,
         bookmarks=bookmarks,
         branches=webutil.branchentries(web.repo, web.stripecount, 10),
         shortlog=changelist,
         node=tip.hex(),
         symrev='tip',
         archives=web.archivelist('tip'),
         labels=web.configlist('web', 'labels')))
 
     return web.res
 
 @webcommand('filediff')
 def filediff(web, req, tmpl):
     """
     /diff/{revision}/{path}
     -----------------------
 
     Show how a file changed in a particular commit.
 
     The ``filediff`` template is rendered.
 
     This handler is registered under both the ``/diff`` and ``/filediff``
     paths. ``/diff`` is used in modern code.
     """
     fctx, ctx = None, None
     try:
         fctx = webutil.filectx(web.repo, req)
     except LookupError:
         ctx = webutil.changectx(web.repo, req)
         path = webutil.cleanpath(web.repo, req.req.qsparams['file'])
         if path not in ctx.files():
             raise
 
     if fctx is not None:
         path = fctx.path()
         ctx = fctx.changectx()
     basectx = ctx.p1()
 
     style = web.config('web', 'style')
     if 'style' in req.req.qsparams:
         style = req.req.qsparams['style']
 
     diffs = webutil.diffs(web, tmpl, ctx, basectx, [path], style)
     if fctx is not None:
         rename = webutil.renamelink(fctx)
         ctx = fctx
     else:
         rename = []
         ctx = ctx
 
     web.res.setbodygen(tmpl(
         'filediff',
         file=path,
         symrev=webutil.symrevorshortnode(req, ctx),
         rename=rename,
         diff=diffs,
         **pycompat.strkwargs(webutil.commonentry(web.repo, ctx))))
 
     return web.res
 
 diff = webcommand('diff')(filediff)
 
 @webcommand('comparison')
 def comparison(web, req, tmpl):
     """
     /comparison/{revision}/{path}
     -----------------------------
 
     Show a comparison between the old and new versions of a file from changes
     made on a particular revision.
 
     This is similar to the ``diff`` handler. However, this form features
     a split or side-by-side diff rather than a unified diff.
 
     The ``context`` query string argument can be used to control the lines of
     context in the diff.
 
     The ``filecomparison`` template is rendered.
     """
     ctx = webutil.changectx(web.repo, req)
     if 'file' not in req.req.qsparams:
         raise ErrorResponse(HTTP_NOT_FOUND, 'file not given')
     path = webutil.cleanpath(web.repo, req.req.qsparams['file'])
 
     parsecontext = lambda v: v == 'full' and -1 or int(v)
     if 'context' in req.req.qsparams:
         context = parsecontext(req.req.qsparams['context'])
     else:
         context = parsecontext(web.config('web', 'comparisoncontext', '5'))
 
     def filelines(f):
         if f.isbinary():
             mt = mimetypes.guess_type(f.path())[0]
             if not mt:
                 mt = 'application/octet-stream'
             return [_('(binary file %s, hash: %s)') % (mt, hex(f.filenode()))]
         return f.data().splitlines()
 
     fctx = None
     parent = ctx.p1()
     leftrev = parent.rev()
     leftnode = parent.node()
     rightrev = ctx.rev()
     rightnode = ctx.node()
     if path in ctx:
         fctx = ctx[path]
         rightlines = filelines(fctx)
         if path not in parent:
             leftlines = ()
         else:
             pfctx = parent[path]
             leftlines = filelines(pfctx)
     else:
         rightlines = ()
         pfctx = ctx.parents()[0][path]
         leftlines = filelines(pfctx)
 
     comparison = webutil.compare(tmpl, context, leftlines, rightlines)
     if fctx is not None:
         rename = webutil.renamelink(fctx)
         ctx = fctx
     else:
         rename = []
         ctx = ctx
 
     web.res.setbodygen(tmpl(
         'filecomparison',
         file=path,
         symrev=webutil.symrevorshortnode(req, ctx),
         rename=rename,
         leftrev=leftrev,
         leftnode=hex(leftnode),
         rightrev=rightrev,
         rightnode=hex(rightnode),
         comparison=comparison,
         **pycompat.strkwargs(webutil.commonentry(web.repo, ctx))))
 
     return web.res
 
 @webcommand('annotate')
 def annotate(web, req, tmpl):
     """
     /annotate/{revision}/{path}
     ---------------------------
 
     Show changeset information for each line in a file.
 
     The ``ignorews``, ``ignorewsamount``, ``ignorewseol``, and
     ``ignoreblanklines`` query string arguments have the same meaning as
     their ``[annotate]`` config equivalents. It uses the hgrc boolean
     parsing logic to interpret the value. e.g. ``0`` and ``false`` are
     false and ``1`` and ``true`` are true. If not defined, the server
     default settings are used.
 
     The ``fileannotate`` template is rendered.
     """
     fctx = webutil.filectx(web.repo, req)
     f = fctx.path()
     parity = paritygen(web.stripecount)
     ishead = fctx.filerev() in fctx.filelog().headrevs()
 
     # parents() is called once per line and several lines likely belong to
     # same revision. So it is worth caching.
     # TODO there are still redundant operations within basefilectx.parents()
     # and from the fctx.annotate() call itself that could be cached.
     parentscache = {}
     def parents(f):
         rev = f.rev()
         if rev not in parentscache:
             parentscache[rev] = []
             for p in f.parents():
                 entry = {
                     'node': p.hex(),
                     'rev': p.rev(),
                 }
                 parentscache[rev].append(entry)
 
         for p in parentscache[rev]:
             yield p
 
     def annotate(**map):
         if fctx.isbinary():
             mt = (mimetypes.guess_type(fctx.path())[0]
                   or 'application/octet-stream')
             lines = [((fctx.filectx(fctx.filerev()), 1), '(binary:%s)' % mt)]
         else:
             lines = webutil.annotate(req, fctx, web.repo.ui)
 
         previousrev = None
         blockparitygen = paritygen(1)
         for lineno, (aline, l) in enumerate(lines):
             f = aline.fctx
             rev = f.rev()
             if rev != previousrev:
                 blockhead = True
                 blockparity = next(blockparitygen)
             else:
                 blockhead = None
             previousrev = rev
             yield {"parity": next(parity),
                    "node": f.hex(),
                    "rev": rev,
                    "author": f.user(),
                    "parents": parents(f),
                    "desc": f.description(),
                    "extra": f.extra(),
                    "file": f.path(),
                    "blockhead": blockhead,
                    "blockparity": blockparity,
                    "targetline": aline.lineno,
                    "line": l,
                    "lineno": lineno + 1,
                    "lineid": "l%d" % (lineno + 1),
                    "linenumber": "% 6d" % (lineno + 1),
                    "revdate": f.date()}
 
     diffopts = webutil.difffeatureopts(req, web.repo.ui, 'annotate')
     diffopts = {k: getattr(diffopts, k) for k in diffopts.defaults}
 
     web.res.setbodygen(tmpl(
         'fileannotate',
         file=f,
         annotate=annotate,
         path=webutil.up(f),
         symrev=webutil.symrevorshortnode(req, fctx),
         rename=webutil.renamelink(fctx),
         permissions=fctx.manifest().flags(f),
         ishead=int(ishead),
         diffopts=diffopts,
         **pycompat.strkwargs(webutil.commonentry(web.repo, fctx))))
 
     return web.res
 
 @webcommand('filelog')
 def filelog(web, req, tmpl):
     """
     /filelog/{revision}/{path}
     --------------------------
 
     Show information about the history of a file in the repository.
 
     The ``revcount`` query string argument can be defined to control the
     maximum number of entries to show.
 
     The ``filelog`` template will be rendered.
     """
 
     try:
         fctx = webutil.filectx(web.repo, req)
         f = fctx.path()
         fl = fctx.filelog()
     except error.LookupError:
         f = webutil.cleanpath(web.repo, req.req.qsparams['file'])
         fl = web.repo.file(f)
         numrevs = len(fl)
         if not numrevs: # file doesn't exist at all
             raise
         rev = webutil.changectx(web.repo, req).rev()
         first = fl.linkrev(0)
         if rev < first: # current rev is from before file existed
             raise
         frev = numrevs - 1
         while fl.linkrev(frev) > rev:
             frev -= 1
         fctx = web.repo.filectx(f, fl.linkrev(frev))
 
     revcount = web.maxshortchanges
     if 'revcount' in req.req.qsparams:
         try:
             revcount = int(req.req.qsparams.get('revcount', revcount))
             revcount = max(revcount, 1)
             tmpl.defaults['sessionvars']['revcount'] = revcount
         except ValueError:
             pass
 
     lrange = webutil.linerange(req)
 
     lessvars = copy.copy(tmpl.defaults['sessionvars'])
     lessvars['revcount'] = max(revcount // 2, 1)
     morevars = copy.copy(tmpl.defaults['sessionvars'])
     morevars['revcount'] = revcount * 2
 
     patch = 'patch' in req.req.qsparams
     if patch:
         lessvars['patch'] = morevars['patch'] = req.req.qsparams['patch']
     descend = 'descend' in req.req.qsparams
     if descend:
         lessvars['descend'] = morevars['descend'] = req.req.qsparams['descend']
 
     count = fctx.filerev() + 1
     start = max(0, count - revcount) # first rev on this page
     end = min(count, start + revcount) # last rev on this page
     parity = paritygen(web.stripecount, offset=start - end)
 
     repo = web.repo
     revs = fctx.filelog().revs(start, end - 1)
     entries = []
 
     diffstyle = web.config('web', 'style')
     if 'style' in req.req.qsparams:
         diffstyle = req.req.qsparams['style']
 
     def diff(fctx, linerange=None):
         ctx = fctx.changectx()
         basectx = ctx.p1()
         path = fctx.path()
         return webutil.diffs(web, tmpl, ctx, basectx, [path], diffstyle,
                              linerange=linerange,
                              lineidprefix='%s-' % ctx.hex()[:12])
 
     linerange = None
     if lrange is not None:
         linerange = webutil.formatlinerange(*lrange)
         # deactivate numeric nav links when linerange is specified as this
         # would required a dedicated "revnav" class
         nav = None
         if descend:
             it = dagop.blockdescendants(fctx, *lrange)
         else:
             it = dagop.blockancestors(fctx, *lrange)
         for i, (c, lr) in enumerate(it, 1):
             diffs = None
             if patch:
                 diffs = diff(c, linerange=lr)
             # follow renames accross filtered (not in range) revisions
             path = c.path()
             entries.append(dict(
                 parity=next(parity),
                 filerev=c.rev(),
                 file=path,
                 diff=diffs,
                 linerange=webutil.formatlinerange(*lr),
                 **pycompat.strkwargs(webutil.commonentry(repo, c))))
             if i == revcount:
                 break
         lessvars['linerange'] = webutil.formatlinerange(*lrange)
         morevars['linerange'] = lessvars['linerange']
     else:
         for i in revs:
             iterfctx = fctx.filectx(i)
             diffs = None
             if patch:
                 diffs = diff(iterfctx)
             entries.append(dict(
                 parity=next(parity),
                 filerev=i,
                 file=f,
                 diff=diffs,
                 rename=webutil.renamelink(iterfctx),
                 **pycompat.strkwargs(webutil.commonentry(repo, iterfctx))))
         entries.reverse()
         revnav = webutil.filerevnav(web.repo, fctx.path())
         nav = revnav.gen(end - 1, revcount, count)
 
     latestentry = entries[:1]
 
     web.res.setbodygen(tmpl(
         'filelog',
         file=f,
         nav=nav,
         symrev=webutil.symrevorshortnode(req, fctx),
         entries=entries,
         descend=descend,
         patch=patch,
         latestentry=latestentry,
         linerange=linerange,
         revcount=revcount,
         morevars=morevars,
         lessvars=lessvars,
         **pycompat.strkwargs(webutil.commonentry(web.repo, fctx))))
 
     return web.res
 
 @webcommand('archive')
 def archive(web, req, tmpl):
     """
     /archive/{revision}.{format}[/{path}]
     -------------------------------------
 
     Obtain an archive of repository content.
 
     The content and type of the archive is defined by a URL path parameter.
     ``format`` is the file extension of the archive type to be generated. e.g.
     ``zip`` or ``tar.bz2``. Not all archive types may be allowed by your
     server configuration.
 
     The optional ``path`` URL parameter controls content to include in the
     archive. If omitted, every file in the specified revision is present in the
     archive. If included, only the specified file or contents of the specified
     directory will be included in the archive.
 
     No template is used for this handler. Raw, binary content is generated.
     """
 
     type_ = req.req.qsparams.get('type')
     allowed = web.configlist("web", "allow_archive")
     key = req.req.qsparams['node']
 
     if type_ not in web.archivespecs:
         msg = 'Unsupported archive type: %s' % type_
         raise ErrorResponse(HTTP_NOT_FOUND, msg)
 
     if not ((type_ in allowed or
              web.configbool("web", "allow" + type_))):
         msg = 'Archive type not allowed: %s' % type_
         raise ErrorResponse(HTTP_FORBIDDEN, msg)
 
     reponame = re.sub(br"\W+", "-", os.path.basename(web.reponame))
     cnode = web.repo.lookup(key)
     arch_version = key
     if cnode == key or key == 'tip':
         arch_version = short(cnode)
     name = "%s-%s" % (reponame, arch_version)
 
     ctx = webutil.changectx(web.repo, req)
     pats = []
     match = scmutil.match(ctx, [])
     file = req.req.qsparams.get('file')
     if file:
         pats = ['path:' + file]
         match = scmutil.match(ctx, pats, default='path')
         if pats:
             files = [f for f in ctx.manifest().keys() if match(f)]
             if not files:
                 raise ErrorResponse(HTTP_NOT_FOUND,
                     'file(s) not found: %s' % file)
 
     mimetype, artype, extension, encoding = web.archivespecs[type_]
-    headers = [
-        ('Content-Disposition', 'attachment; filename=%s%s' % (name, extension))
-        ]
+
+    web.res.headers['Content-Type'] = mimetype
+    web.res.headers['Content-Disposition'] = 'attachment; filename=%s%s' % (
+        name, extension)
+
     if encoding:
-        headers.append(('Content-Encoding', encoding))
-    req.headers.extend(headers)
-    req.respond(HTTP_OK, mimetype)
+        web.res.headers['Content-Encoding'] = encoding
 
-    bodyfh = requestmod.offsettrackingwriter(req.write)
+    web.res.setbodywillwrite()
+    assert list(web.res.sendresponse()) == []
+
+    bodyfh = web.res.getbodyfile()
 
     archival.archive(web.repo, bodyfh, cnode, artype, prefix=name,
                      matchfn=match,
                      subrepos=web.configbool("web", "archivesubrepos"))
-    return []
 
+    return True
 
 @webcommand('static')
 def static(web, req, tmpl):
     fname = req.req.qsparams['file']
     # a repo owner may set web.static in .hg/hgrc to get any file
     # readable by the user running the CGI script
     static = web.config("web", "static", None, untrusted=False)
     if not static:
         tp = web.templatepath or templater.templatepaths()
         if isinstance(tp, str):
             tp = [tp]
         static = [os.path.join(p, 'static') for p in tp]
 
     staticfile(static, fname, web.res)
     return web.res
 
 @webcommand('graph')
 def graph(web, req, tmpl):
     """
     /graph[/{revision}]
     -------------------
 
     Show information about the graphical topology of the repository.
 
     Information rendered by this handler can be used to create visual
     representations of repository topology.
 
     The ``revision`` URL parameter controls the starting changeset. If it's
     absent, the default is ``tip``.
 
     The ``revcount`` query string argument can define the number of changesets
     to show information for.
 
     The ``graphtop`` query string argument can specify the starting changeset
     for producing ``jsdata`` variable that is used for rendering graph in
     JavaScript. By default it has the same value as ``revision``.
 
     This handler will render the ``graph`` template.
     """
 
     if 'node' in req.req.qsparams:
         ctx = webutil.changectx(web.repo, req)
         symrev = webutil.symrevorshortnode(req, ctx)
     else:
         ctx = web.repo['tip']
         symrev = 'tip'
     rev = ctx.rev()
 
     bg_height = 39
     revcount = web.maxshortchanges
     if 'revcount' in req.req.qsparams:
         try:
             revcount = int(req.req.qsparams.get('revcount', revcount))
             revcount = max(revcount, 1)
             tmpl.defaults['sessionvars']['revcount'] = revcount
         except ValueError:
             pass
 
     lessvars = copy.copy(tmpl.defaults['sessionvars'])
     lessvars['revcount'] = max(revcount // 2, 1)
     morevars = copy.copy(tmpl.defaults['sessionvars'])
     morevars['revcount'] = revcount * 2
 
     graphtop = req.req.qsparams.get('graphtop', ctx.hex())
     graphvars = copy.copy(tmpl.defaults['sessionvars'])
     graphvars['graphtop'] = graphtop
 
     count = len(web.repo)
     pos = rev
 
     uprev = min(max(0, count - 1), rev + revcount)
     downrev = max(0, rev - revcount)
     changenav = webutil.revnav(web.repo).gen(pos, revcount, count)
 
     tree = []
     nextentry = []
     lastrev = 0
     if pos != -1:
         allrevs = web.repo.changelog.revs(pos, 0)
         revs = []
         for i in allrevs:
             revs.append(i)
             if len(revs) >= revcount + 1:
                 break
 
         if len(revs) > revcount:
             nextentry = [webutil.commonentry(web.repo, web.repo[revs[-1]])]
             revs = revs[:-1]
 
         lastrev = revs[-1]
 
         # We have to feed a baseset to dagwalker as it is expecting smartset
         # object. This does not have a big impact on hgweb performance itself
         # since hgweb graphing code is not itself lazy yet.
         dag = graphmod.dagwalker(web.repo, smartset.baseset(revs))
         # As we said one line above... not lazy.
         tree = list(item for item in graphmod.colored(dag, web.repo)
                     if item[1] == graphmod.CHANGESET)
 
     def nodecurrent(ctx):
         wpnodes = web.repo.dirstate.parents()
         if wpnodes[1] == nullid:
             wpnodes = wpnodes[:1]
         if ctx.node() in wpnodes:
             return '@'
         return ''
 
     def nodesymbol(ctx):
         if ctx.obsolete():
             return 'x'
         elif ctx.isunstable():
             return '*'
         elif ctx.closesbranch():
             return '_'
         else:
             return 'o'
 
     def fulltree():
         pos = web.repo[graphtop].rev()
         tree = []
         if pos != -1:
             revs = web.repo.changelog.revs(pos, lastrev)
             dag = graphmod.dagwalker(web.repo, smartset.baseset(revs))
             tree = list(item for item in graphmod.colored(dag, web.repo)
                         if item[1] == graphmod.CHANGESET)
         return tree
 
     def jsdata():
         return [{'node': pycompat.bytestr(ctx),
                  'graphnode': nodecurrent(ctx) + nodesymbol(ctx),
                  'vertex': vtx,
                  'edges': edges}
                 for (id, type, ctx, vtx, edges) in fulltree()]
 
     def nodes():
         parity = paritygen(web.stripecount)
         for row, (id, type, ctx, vtx, edges) in enumerate(tree):
             entry = webutil.commonentry(web.repo, ctx)
             edgedata = [{'col': edge[0],
                          'nextcol': edge[1],
                          'color': (edge[2] - 1) % 6 + 1,
                          'width': edge[3],
                          'bcolor': edge[4]}
                         for edge in edges]
 
             entry.update({'col': vtx[0],
                           'color': (vtx[1] - 1) % 6 + 1,
                           'parity': next(parity),
                           'edges': edgedata,
                           'row': row,
                           'nextrow': row + 1})
 
             yield entry
 
     rows = len(tree)
 
     web.res.setbodygen(tmpl(
         'graph',
         rev=rev,
         symrev=symrev,
         revcount=revcount,
         uprev=uprev,
         lessvars=lessvars,
         morevars=morevars,
         downrev=downrev,
         graphvars=graphvars,
         rows=rows,
         bg_height=bg_height,
         changesets=count,
         nextentry=nextentry,
         jsdata=lambda **x: jsdata(),
         nodes=lambda **x: nodes(),
         node=ctx.hex(),
         changenav=changenav))
 
     return web.res
 
 def _getdoc(e):
     doc = e[0].__doc__
     if doc:
         doc = _(doc).partition('\n')[0]
     else:
         doc = _('(no help text available)')
     return doc
 
 @webcommand('help')
 def help(web, req, tmpl):
     """
     /help[/{topic}]
     ---------------
 
     Render help documentation.
 
     This web command is roughly equivalent to :hg:`help`. If a ``topic``
     is defined, that help topic will be rendered. If not, an index of
     available help topics will be rendered.
 
     The ``help`` template will be rendered when requesting help for a topic.
     ``helptopics`` will be rendered for the index of help topics.
     """
     from .. import commands, help as helpmod  # avoid cycle
 
     topicname = req.req.qsparams.get('node')
     if not topicname:
         def topics(**map):
             for entries, summary, _doc in helpmod.helptable:
                 yield {'topic': entries[0], 'summary': summary}
 
         early, other = [], []
         primary = lambda s: s.partition('|')[0]
         for c, e in commands.table.iteritems():
             doc = _getdoc(e)
             if 'DEPRECATED' in doc or c.startswith('debug'):
                 continue
             cmd = primary(c)
             if cmd.startswith('^'):
                 early.append((cmd[1:], doc))
             else:
                 other.append((cmd, doc))
 
         early.sort()
         other.sort()
 
         def earlycommands(**map):
             for c, doc in early:
                 yield {'topic': c, 'summary': doc}
 
         def othercommands(**map):
             for c, doc in other:
                 yield {'topic': c, 'summary': doc}
 
         web.res.setbodygen(tmpl(
             'helptopics',
             topics=topics,
             earlycommands=earlycommands,
             othercommands=othercommands,
             title='Index'))
         return web.res
 
     # Render an index of sub-topics.
     if topicname in helpmod.subtopics:
         topics = []
         for entries, summary, _doc in helpmod.subtopics[topicname]:
             topics.append({
                 'topic': '%s.%s' % (topicname, entries[0]),
                 'basename': entries[0],
                 'summary': summary,
             })
 
         web.res.setbodygen(tmpl(
             'helptopics',
             topics=topics,
             title=topicname,
             subindex=True))
         return web.res
 
     u = webutil.wsgiui.load()
     u.verbose = True
 
     # Render a page from a sub-topic.
     if '.' in topicname:
         # TODO implement support for rendering sections, like
         # `hg help` works.
         topic, subtopic = topicname.split('.', 1)
         if topic not in helpmod.subtopics:
             raise ErrorResponse(HTTP_NOT_FOUND)
     else:
         topic = topicname
         subtopic = None
 
     try:
         doc = helpmod.help_(u, commands, topic, subtopic=subtopic)
     except error.Abort:
         raise ErrorResponse(HTTP_NOT_FOUND)
 
     web.res.setbodygen(tmpl(
         'help',
         topic=topicname,
         doc=doc))
 
     return web.res
 
 # tell hggettext to extract docstrings from these functions:
 i18nfunctions = commands.values()

tests/hgweberror.py

@@ -1,21 +1,24 b''
 # A dummy extension that installs an hgweb command that throws an Exception.
 
 from __future__ import absolute_import
 
 from mercurial.hgweb import (
     webcommands,
 )
 
 def raiseerror(web, req, tmpl):
     '''Dummy web command that raises an uncaught Exception.'''
 
     # Simulate an error after partial response.
-    if 'partialresponse' in req.req.qsparams:
-        req.respond(200, 'text/plain')
-        req.write('partial content\n')
+    if 'partialresponse' in web.req.qsparams:
+        web.res.status = b'200 Script output follows'
+        web.res.headers[b'Content-Type'] = b'text/plain'
+        web.res.setbodywillwrite()
+        list(web.res.sendresponse())
+        web.res.getbodyfile().write(b'partial content\n')
 
     raise AttributeError('I am an uncaught error!')
 
 def extsetup(ui):
     setattr(webcommands, 'raiseerror', raiseerror)
     webcommands.__all__.append('raiseerror')