# hgweb/hgweb_mod.py - Web interface for a repository. # # Copyright 21 May 2005 - (c) 2005 Jake Edge # Copyright 2005-2007 Matt Mackall # # This software may be used and distributed according to the terms # of the GNU General Public License, incorporated herein by reference. import os, mimetypes, re, mimetools, cStringIO from mercurial.node import * from mercurial import mdiff, ui, hg, util, archival, patch, hook from mercurial import revlog, templater from common import ErrorResponse, get_mtime, style_map, paritygen, get_contact from request import wsgirequest import webcommands, protocol shortcuts = { 'cl': [('cmd', ['changelog']), ('rev', None)], 'sl': [('cmd', ['shortlog']), ('rev', None)], 'cs': [('cmd', ['changeset']), ('node', None)], 'f': [('cmd', ['file']), ('filenode', None)], 'fl': [('cmd', ['filelog']), ('filenode', None)], 'fd': [('cmd', ['filediff']), ('node', None)], 'fa': [('cmd', ['annotate']), ('filenode', None)], 'mf': [('cmd', ['manifest']), ('manifest', None)], 'ca': [('cmd', ['archive']), ('node', None)], 'tags': [('cmd', ['tags'])], 'tip': [('cmd', ['changeset']), ('node', ['tip'])], 'static': [('cmd', ['static']), ('file', None)] } def _up(p): if p[0] != "/": p = "/" + p if p[-1] == "/": p = p[:-1] up = os.path.dirname(p) if up == "/": return "/" return up + "/" def revnavgen(pos, pagelen, limit, nodefunc): def seq(factor, limit=None): if limit: yield limit if limit >= 20 and limit <= 40: yield 50 else: yield 1 * factor yield 3 * factor for f in seq(factor * 10): yield f def nav(**map): l = [] last = 0 for f in seq(1, pagelen): if f < pagelen or f <= last: continue if f > limit: break last = f if pos + f < limit: l.append(("+%d" % f, hex(nodefunc(pos + f).node()))) if pos - f >= 0: l.insert(0, ("-%d" % f, hex(nodefunc(pos - f).node()))) try: yield {"label": "(0)", "node": hex(nodefunc('0').node())} for label, node in l: yield {"label": label, "node": node} yield {"label": "tip", "node": "tip"} except hg.RepoError: pass return nav class hgweb(object): def __init__(self, repo, name=None): if isinstance(repo, str): parentui = ui.ui(report_untrusted=False, interactive=False) self.repo = hg.repository(parentui, repo) else: self.repo = repo hook.redirect(True) self.mtime = -1 self.reponame = name self.archives = 'zip', 'gz', 'bz2' self.stripecount = 1 # a repo owner may set web.templates in .hg/hgrc to get any file # readable by the user running the CGI script self.templatepath = self.config("web", "templates", templater.templatepath(), untrusted=False) # The CGI scripts are often run by a user different from the repo owner. # Trust the settings from the .hg/hgrc files by default. def config(self, section, name, default=None, untrusted=True): return self.repo.ui.config(section, name, default, untrusted=untrusted) def configbool(self, section, name, default=False, untrusted=True): return self.repo.ui.configbool(section, name, default, untrusted=untrusted) def configlist(self, section, name, default=None, untrusted=True): return self.repo.ui.configlist(section, name, default, untrusted=untrusted) def refresh(self): mtime = get_mtime(self.repo.root) if mtime != self.mtime: self.mtime = mtime self.repo = hg.repository(self.repo.ui, self.repo.root) self.maxchanges = int(self.config("web", "maxchanges", 10)) self.stripecount = int(self.config("web", "stripes", 1)) self.maxshortchanges = int(self.config("web", "maxshortchanges", 60)) self.maxfiles = int(self.config("web", "maxfiles", 10)) self.allowpull = self.configbool("web", "allowpull", True) self.encoding = self.config("web", "encoding", util._encoding) def run(self): if not os.environ.get('GATEWAY_INTERFACE', '').startswith("CGI/1."): raise RuntimeError("This function is only intended to be called while running as a CGI script.") import mercurial.hgweb.wsgicgi as wsgicgi wsgicgi.launch(self) def __call__(self, env, respond): req = wsgirequest(env, respond) self.run_wsgi(req) return req def run_wsgi(self, req): self.refresh() # expand form shortcuts for k in shortcuts.iterkeys(): if k in req.form: for name, value in shortcuts[k]: if value is None: value = req.form[k] req.form[name] = value del req.form[k] # work with CGI variables to create coherent structure # use SCRIPT_NAME, PATH_INFO and QUERY_STRING as well as our REPO_NAME req.url = req.env['SCRIPT_NAME'] if not req.url.endswith('/'): req.url += '/' if req.env.has_key('REPO_NAME'): req.url += req.env['REPO_NAME'] + '/' if req.env.get('PATH_INFO'): parts = req.env.get('PATH_INFO').strip('/').split('/') repo_parts = req.env.get('REPO_NAME', '').split('/') if parts[:len(repo_parts)] == repo_parts: parts = parts[len(repo_parts):] query = '/'.join(parts) else: query = req.env['QUERY_STRING'].split('&', 1)[0] query = query.split(';', 1)[0] # translate user-visible url structure to internal structure args = query.split('/', 2) if 'cmd' not in req.form and args and args[0]: cmd = args.pop(0) style = cmd.rfind('-') if style != -1: req.form['style'] = [cmd[:style]] cmd = cmd[style+1:] # avoid accepting e.g. style parameter as command if hasattr(webcommands, cmd) or hasattr(protocol, cmd): req.form['cmd'] = [cmd] if args and args[0]: node = args.pop(0) req.form['node'] = [node] if args: req.form['file'] = args if cmd == 'static': req.form['file'] = req.form['node'] elif cmd == 'archive': fn = req.form['node'][0] for type_, spec in self.archive_specs.iteritems(): ext = spec[2] if fn.endswith(ext): req.form['node'] = [fn[:-len(ext)]] req.form['type'] = [type_] # actually process the request try: cmd = req.form.get('cmd', [''])[0] if hasattr(protocol, cmd): method = getattr(protocol, cmd) method(self, req) else: tmpl = self.templater(req) if cmd == '': req.form['cmd'] = [tmpl.cache['default']] cmd = req.form['cmd'][0] method = getattr(webcommands, cmd) method(self, req, tmpl) del tmpl except revlog.LookupError, err: req.respond(404, tmpl( 'error', error='revision not found: %s' % err.name)) except (hg.RepoError, revlog.RevlogError), inst: req.respond('500 Internal Server Error', tmpl('error', error=str(inst))) except ErrorResponse, inst: req.respond(inst.code, tmpl('error', error=inst.message)) except AttributeError: req.respond(400, tmpl('error', error='No such method: ' + cmd)) def templater(self, req): # determine scheme, port and server name # this is needed to create absolute urls proto = req.env.get('wsgi.url_scheme') if proto == 'https': proto = 'https' default_port = "443" else: proto = 'http' default_port = "80" port = req.env["SERVER_PORT"] port = port != default_port and (":" + port) or "" urlbase = '%s://%s%s' % (proto, req.env['SERVER_NAME'], port) staticurl = self.config("web", "staticurl") or req.url + 'static/' if not staticurl.endswith('/'): staticurl += '/' # some functions for the templater def header(**map): header_file = cStringIO.StringIO( ''.join(tmpl("header", encoding=self.encoding, **map))) msg = mimetools.Message(header_file, 0) req.header(msg.items()) yield header_file.read() def rawfileheader(**map): req.header([('Content-type', map['mimetype']), ('Content-disposition', 'filename=%s' % map['file']), ('Content-length', str(len(map['raw'])))]) yield '' def footer(**map): yield tmpl("footer", **map) def motd(**map): yield self.config("web", "motd", "") def sessionvars(**map): fields = [] if req.form.has_key('style'): style = req.form['style'][0] if style != self.config('web', 'style', ''): fields.append(('style', style)) separator = req.url[-1] == '?' and ';' or '?' for name, value in fields: yield dict(name=name, value=value, separator=separator) separator = ';' # figure out which style to use style = self.config("web", "style", "") if req.form.has_key('style'): style = req.form['style'][0] mapfile = style_map(self.templatepath, style) if not self.reponame: self.reponame = (self.config("web", "name") or req.env.get('REPO_NAME') or req.url.strip('/') or self.repo.root) # create the templater tmpl = templater.templater(mapfile, templater.common_filters, defaults={"url": req.url, "staticurl": staticurl, "urlbase": urlbase, "repo": self.reponame, "header": header, "footer": footer, "motd": motd, "rawfileheader": rawfileheader, "sessionvars": sessionvars }) return tmpl def archivelist(self, nodeid): allowed = self.configlist("web", "allow_archive") for i, spec in self.archive_specs.iteritems(): if i in allowed or self.configbool("web", "allow" + i): yield {"type" : i, "extension" : spec[2], "node" : nodeid} def listfilediffs(self, tmpl, files, changeset): for f in files[:self.maxfiles]: yield tmpl("filedifflink", node=hex(changeset), file=f) if len(files) > self.maxfiles: yield tmpl("fileellipses") def siblings(self, siblings=[], hiderev=None, **args): siblings = [s for s in siblings if s.node() != nullid] if len(siblings) == 1 and siblings[0].rev() == hiderev: return for s in siblings: d = {'node': hex(s.node()), 'rev': s.rev()} if hasattr(s, 'path'): d['file'] = s.path() d.update(args) yield d def renamelink(self, fl, node): r = fl.renamed(node) if r: return [dict(file=r[0], node=hex(r[1]))] return [] def nodetagsdict(self, node): return [{"name": i} for i in self.repo.nodetags(node)] def nodebranchdict(self, ctx): branches = [] branch = ctx.branch() # If this is an empty repo, ctx.node() == nullid, # ctx.branch() == 'default', but branchtags() is # an empty dict. Using dict.get avoids a traceback. if self.repo.branchtags().get(branch) == ctx.node(): branches.append({"name": branch}) return branches def showtag(self, tmpl, t1, node=nullid, **args): for t in self.repo.nodetags(node): yield tmpl(t1, tag=t, **args) def diff(self, tmpl, node1, node2, files): def filterfiles(filters, files): l = [x for x in files if x in filters] for t in filters: if t and t[-1] != os.sep: t += os.sep l += [x for x in files if x.startswith(t)] return l parity = paritygen(self.stripecount) def diffblock(diff, f, fn): yield tmpl("diffblock", lines=prettyprintlines(diff), parity=parity.next(), file=f, filenode=hex(fn or nullid)) def prettyprintlines(diff): for l in diff.splitlines(1): if l.startswith('+'): yield tmpl("difflineplus", line=l) elif l.startswith('-'): yield tmpl("difflineminus", line=l) elif l.startswith('@'): yield tmpl("difflineat", line=l) else: yield tmpl("diffline", line=l) r = self.repo c1 = r.changectx(node1) c2 = r.changectx(node2) date1 = util.datestr(c1.date()) date2 = util.datestr(c2.date()) modified, added, removed, deleted, unknown = r.status(node1, node2)[:5] if files: modified, added, removed = map(lambda x: filterfiles(files, x), (modified, added, removed)) diffopts = patch.diffopts(self.repo.ui, untrusted=True) for f in modified: to = c1.filectx(f).data() tn = c2.filectx(f).data() yield diffblock(mdiff.unidiff(to, date1, tn, date2, f, f, opts=diffopts), f, tn) for f in added: to = None tn = c2.filectx(f).data() yield diffblock(mdiff.unidiff(to, date1, tn, date2, f, f, opts=diffopts), f, tn) for f in removed: to = c1.filectx(f).data() tn = None yield diffblock(mdiff.unidiff(to, date1, tn, date2, f, f, opts=diffopts), f, tn) def changelog(self, tmpl, ctx, shortlog=False): def changelist(limit=0,**map): cl = self.repo.changelog l = [] # build a list in forward order for efficiency for i in xrange(start, end): ctx = self.repo.changectx(i) n = ctx.node() l.insert(0, {"parity": parity.next(), "author": ctx.user(), "parent": self.siblings(ctx.parents(), i - 1), "child": self.siblings(ctx.children(), i + 1), "changelogtag": self.showtag("changelogtag",n), "desc": ctx.description(), "date": ctx.date(), "files": self.listfilediffs(tmpl, ctx.files(), n), "rev": i, "node": hex(n), "tags": self.nodetagsdict(n), "branches": self.nodebranchdict(ctx)}) if limit > 0: l = l[:limit] for e in l: yield e maxchanges = shortlog and self.maxshortchanges or self.maxchanges cl = self.repo.changelog count = cl.count() pos = ctx.rev() start = max(0, pos - maxchanges + 1) end = min(count, start + maxchanges) pos = end - 1 parity = paritygen(self.stripecount, offset=start-end) changenav = revnavgen(pos, maxchanges, count, self.repo.changectx) yield tmpl(shortlog and 'shortlog' or 'changelog', changenav=changenav, node=hex(cl.tip()), rev=pos, changesets=count, entries=lambda **x: changelist(limit=0,**x), latestentry=lambda **x: changelist(limit=1,**x), archives=self.archivelist("tip")) def search(self, tmpl, query): def changelist(**map): cl = self.repo.changelog count = 0 qw = query.lower().split() def revgen(): for i in xrange(cl.count() - 1, 0, -100): l = [] for j in xrange(max(0, i - 100), i): ctx = self.repo.changectx(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 ctx.user().lower() or q in ctx.description().lower() or q in " ".join(ctx.files()).lower()): miss = 1 break if miss: continue count += 1 n = ctx.node() yield tmpl('searchentry', parity=parity.next(), author=ctx.user(), parent=self.siblings(ctx.parents()), child=self.siblings(ctx.children()), changelogtag=self.showtag("changelogtag",n), desc=ctx.description(), date=ctx.date(), files=self.listfilediffs(tmpl, ctx.files(), n), rev=ctx.rev(), node=hex(n), tags=self.nodetagsdict(n), branches=self.nodebranchdict(ctx)) if count >= self.maxchanges: break cl = self.repo.changelog parity = paritygen(self.stripecount) yield tmpl('search', query=query, node=hex(cl.tip()), entries=changelist, archives=self.archivelist("tip")) def changeset(self, tmpl, ctx): n = ctx.node() parents = ctx.parents() p1 = parents[0].node() files = [] parity = paritygen(self.stripecount) for f in ctx.files(): files.append(tmpl("filenodelink", node=hex(n), file=f, parity=parity.next())) def diff(**map): yield self.diff(tmpl, p1, n, None) yield tmpl('changeset', diff=diff, rev=ctx.rev(), node=hex(n), parent=self.siblings(parents), child=self.siblings(ctx.children()), changesettag=self.showtag("changesettag",n), author=ctx.user(), desc=ctx.description(), date=ctx.date(), files=files, archives=self.archivelist(hex(n)), tags=self.nodetagsdict(n), branches=self.nodebranchdict(ctx)) def filelog(self, tmpl, fctx): f = fctx.path() fl = fctx.filelog() count = fl.count() pagelen = self.maxshortchanges pos = fctx.filerev() start = max(0, pos - pagelen + 1) end = min(count, start + pagelen) pos = end - 1 parity = paritygen(self.stripecount, offset=start-end) def entries(limit=0, **map): l = [] for i in xrange(start, end): ctx = fctx.filectx(i) n = fl.node(i) l.insert(0, {"parity": parity.next(), "filerev": i, "file": f, "node": hex(ctx.node()), "author": ctx.user(), "date": ctx.date(), "rename": self.renamelink(fl, n), "parent": self.siblings(fctx.parents()), "child": self.siblings(fctx.children()), "desc": ctx.description()}) if limit > 0: l = l[:limit] for e in l: yield e nodefunc = lambda x: fctx.filectx(fileid=x) nav = revnavgen(pos, pagelen, count, nodefunc) yield tmpl("filelog", file=f, node=hex(fctx.node()), nav=nav, entries=lambda **x: entries(limit=0, **x), latestentry=lambda **x: entries(limit=1, **x)) def filerevision(self, tmpl, fctx): f = fctx.path() text = fctx.data() fl = fctx.filelog() n = fctx.filenode() parity = paritygen(self.stripecount) mt = mimetypes.guess_type(f)[0] rawtext = text if util.binary(text): mt = mt or 'application/octet-stream' text = "(binary:%s)" % mt mt = mt or 'text/plain' def lines(): for l, t in enumerate(text.splitlines(1)): yield {"line": t, "linenumber": "% 6d" % (l + 1), "parity": parity.next()} yield tmpl("filerevision", file=f, path=_up(f), text=lines(), raw=rawtext, mimetype=mt, rev=fctx.rev(), node=hex(fctx.node()), author=fctx.user(), date=fctx.date(), desc=fctx.description(), parent=self.siblings(fctx.parents()), child=self.siblings(fctx.children()), rename=self.renamelink(fl, n), permissions=fctx.manifest().flags(f)) def fileannotate(self, tmpl, fctx): f = fctx.path() n = fctx.filenode() fl = fctx.filelog() parity = paritygen(self.stripecount) def annotate(**map): last = None for f, l in fctx.annotate(follow=True): fnode = f.filenode() name = self.repo.ui.shortuser(f.user()) if last != fnode: last = fnode yield {"parity": parity.next(), "node": hex(f.node()), "rev": f.rev(), "author": name, "file": f.path(), "line": l} yield tmpl("fileannotate", file=f, annotate=annotate, path=_up(f), rev=fctx.rev(), node=hex(fctx.node()), author=fctx.user(), date=fctx.date(), desc=fctx.description(), rename=self.renamelink(fl, n), parent=self.siblings(fctx.parents()), child=self.siblings(fctx.children()), permissions=fctx.manifest().flags(f)) def manifest(self, tmpl, ctx, path): mf = ctx.manifest() node = ctx.node() files = {} parity = paritygen(self.stripecount) if path and path[-1] != "/": path += "/" l = len(path) abspath = "/" + path for f, n in mf.items(): if f[:l] != path: continue remain = f[l:] if "/" in remain: short = remain[:remain.index("/") + 1] # bleah files[short] = (f, None) else: short = os.path.basename(remain) files[short] = (f, n) if not files: raise ErrorResponse(404, 'Path not found: ' + path) def filelist(**map): fl = files.keys() fl.sort() for f in fl: full, fnode = files[f] if not fnode: continue fctx = ctx.filectx(full) yield {"file": full, "parity": parity.next(), "basename": f, "date": fctx.changectx().date(), "size": fctx.size(), "permissions": mf.flags(full)} def dirlist(**map): fl = files.keys() fl.sort() for f in fl: full, fnode = files[f] if fnode: continue yield {"parity": parity.next(), "path": "%s%s" % (abspath, f), "basename": f[:-1]} yield tmpl("manifest", rev=ctx.rev(), node=hex(node), path=abspath, up=_up(abspath), upparity=parity.next(), fentries=filelist, dentries=dirlist, archives=self.archivelist(hex(node)), tags=self.nodetagsdict(node), branches=self.nodebranchdict(ctx)) def tags(self, tmpl): i = self.repo.tagslist() i.reverse() parity = paritygen(self.stripecount) def entries(notip=False,limit=0, **map): count = 0 for k, n in i: if notip and k == "tip": continue if limit > 0 and count >= limit: continue count = count + 1 yield {"parity": parity.next(), "tag": k, "date": self.repo.changectx(n).date(), "node": hex(n)} yield tmpl("tags", node=hex(self.repo.changelog.tip()), entries=lambda **x: entries(False,0, **x), entriesnotip=lambda **x: entries(True,0, **x), latestentry=lambda **x: entries(True,1, **x)) def summary(self, tmpl): i = self.repo.tagslist() i.reverse() def tagentries(**map): parity = paritygen(self.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=parity.next(), tag=k, node=hex(n), date=self.repo.changectx(n).date()) def branches(**map): parity = paritygen(self.stripecount) b = self.repo.branchtags() l = [(-self.repo.changelog.rev(n), n, t) for t, n in b.items()] l.sort() for r,n,t in l: ctx = self.repo.changectx(n) yield {'parity': parity.next(), 'branch': t, 'node': hex(n), 'date': ctx.date()} def changelist(**map): parity = paritygen(self.stripecount, offset=start-end) l = [] # build a list in forward order for efficiency for i in xrange(start, end): ctx = self.repo.changectx(i) n = ctx.node() hn = hex(n) l.insert(0, tmpl( 'shortlogentry', parity=parity.next(), author=ctx.user(), desc=ctx.description(), date=ctx.date(), rev=i, node=hn, tags=self.nodetagsdict(n), branches=self.nodebranchdict(ctx))) yield l cl = self.repo.changelog count = cl.count() start = max(0, count - self.maxchanges) end = min(count, start + self.maxchanges) yield tmpl("summary", desc=self.config("web", "description", "unknown"), owner=get_contact(self.config) or "unknown", lastchange=cl.read(cl.tip())[2], tags=tagentries, branches=branches, shortlog=changelist, node=hex(cl.tip()), archives=self.archivelist("tip")) def filediff(self, tmpl, fctx): n = fctx.node() path = fctx.path() parents = fctx.parents() p1 = parents and parents[0].node() or nullid def diff(**map): yield self.diff(tmpl, p1, n, [path]) yield tmpl("filediff", file=path, node=hex(n), rev=fctx.rev(), parent=self.siblings(parents), child=self.siblings(fctx.children()), diff=diff) archive_specs = { 'bz2': ('application/x-tar', 'tbz2', '.tar.bz2', None), 'gz': ('application/x-tar', 'tgz', '.tar.gz', None), 'zip': ('application/zip', 'zip', '.zip', None), } def archive(self, tmpl, req, key, type_): reponame = re.sub(r"\W+", "-", os.path.basename(self.reponame)) cnode = self.repo.lookup(key) arch_version = key if cnode == key or key == 'tip': arch_version = short(cnode) name = "%s-%s" % (reponame, arch_version) mimetype, artype, extension, encoding = self.archive_specs[type_] headers = [('Content-type', mimetype), ('Content-disposition', 'attachment; filename=%s%s' % (name, extension))] if encoding: headers.append(('Content-encoding', encoding)) req.header(headers) archival.archive(self.repo, req.out, cnode, artype, prefix=name) # add tags to things # tags -> list of changesets corresponding to tags # find tag, changeset, file def cleanpath(self, path): path = path.lstrip('/') return util.canonpath(self.repo.root, '', path) def changectx(self, req): if req.form.has_key('node'): changeid = req.form['node'][0] elif req.form.has_key('manifest'): changeid = req.form['manifest'][0] else: changeid = self.repo.changelog.count() - 1 try: ctx = self.repo.changectx(changeid) except hg.RepoError: man = self.repo.manifest mn = man.lookup(changeid) ctx = self.repo.changectx(man.linkrev(mn)) return ctx def filectx(self, req): path = self.cleanpath(req.form['file'][0]) if req.form.has_key('node'): changeid = req.form['node'][0] else: changeid = req.form['filenode'][0] try: ctx = self.repo.changectx(changeid) fctx = ctx.filectx(path) except hg.RepoError: fctx = self.repo.filectx(path, fileid=changeid) return fctx def check_perm(self, req, op, default): '''check permission for operation based on user auth. return true if op allowed, else false. default is policy to use if no config given.''' user = req.env.get('REMOTE_USER') deny = self.configlist('web', 'deny_' + op) if deny and (not user or deny == ['*'] or user in deny): return False allow = self.configlist('web', 'allow_' + op) return (allow and (allow == ['*'] or user in allow)) or default