Make hgweb.staticfile() more secure and portable.

Without this, files in directories next to the static directory starting
with 'static' could be retrieved, e.g. with '../static.private/foo'.
Additionally staticfile now generates platform specific pathnames from
the /-separated paths given in the URL.
Illegal file names (e.g. containing %00) now yield a sane error message.

mercurial/hgweb.py

@@ -1,1188 +1,1185 b''
 # hgweb.py - web interface to a mercurial repository
 #
 # Copyright 21 May 2005 - (c) 2005 Jake Edge <[e-mail]>
 # Copyright 2005 Matt Mackall <[e-mail]>
 #
 # This software may be used and distributed according to the terms
 # of the GNU General Public License, incorporated herein by reference.
 
 import os, cgi, sys, urllib
 import mimetypes
 from demandload import demandload
 demandload(globals(), "mdiff time re socket zlib errno ui hg ConfigParser")
 demandload(globals(), "zipfile tempfile StringIO tarfile BaseHTTPServer util")
 demandload(globals(), "mimetypes")
 from node import *
 from i18n import gettext as _
 
 def templatepath():
     for f in "templates", "../templates":
         p = os.path.join(os.path.dirname(__file__), f)
         if os.path.isdir(p):
             return os.path.normpath(p)
 
 def age(x):
     def plural(t, c):
         if c == 1:
             return t
         return t + "s"
     def fmt(t, c):
         return "%d %s" % (c, plural(t, c))
 
     now = time.time()
     then = x[0]
     delta = max(1, int(now - then))
 
     scales = [["second", 1],
               ["minute", 60],
               ["hour", 3600],
               ["day", 3600 * 24],
               ["week", 3600 * 24 * 7],
               ["month", 3600 * 24 * 30],
               ["year", 3600 * 24 * 365]]
 
     scales.reverse()
 
     for t, s in scales:
         n = delta / s
         if n >= 2 or s == 1:
             return fmt(t, n)
 
 def nl2br(text):
     return text.replace('\n', '<br/>\n')
 
 def obfuscate(text):
     return ''.join(['&#%d;' % ord(c) for c in text])
 
 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 get_mtime(repo_path):
     hg_path = os.path.join(repo_path, ".hg")
     cl_path = os.path.join(hg_path, "00changelog.i")
     if os.path.exists(os.path.join(cl_path)):
         return os.stat(cl_path).st_mtime
     else:
         return os.stat(hg_path).st_mtime
 
 def staticfile(directory, fname):
-    fname = os.path.realpath(os.path.join(directory, fname))
+    """return a file inside directory with guessed content-type header
+
+    fname always uses '/' as directory separator and isn't allowed to
+    contain unusual path components.
+    Content-type is guessed using the mimetypes module.
+    Return an empty string if fname is illegal or file not found.
 
+    """
+    parts = fname.split('/')
+    path = directory
+    for part in parts:
+        if (part in ('', os.curdir, os.pardir) or
+            os.sep in part or os.altsep is not None and os.altsep in part):
+            return ""
+        path = os.path.join(path, part)
     try:
-        # the static dir should be a substring in the real
-        # file path, if it is not, we have something strange
-        # going on => security breach attempt?
-        #
-        # This will either:
-        #   1) find the `static' path at index 0  =  success
-        #   2) find the `static' path at other index  =  error
-        #   3) not find the `static' path  =  ValueError generated
-        if fname.index(directory) != 0:
-            # generate ValueError manually
-            raise ValueError()
-
-        os.stat(fname)
-
-        ct = mimetypes.guess_type(fname)[0] or "text/plain"
-        return "Content-type: %s\n\n%s" % (ct, file(fname).read())
-    except ValueError:
-        # security breach attempt
+        os.stat(path)
+        ct = mimetypes.guess_type(path)[0] or "text/plain"
+        return "Content-type: %s\n\n%s" % (ct, file(path).read())
+    except (TypeError, OSError):
+        # illegal fname or unreadable file
         return ""
-    except OSError, e:
-        if e.errno == errno.ENOENT:
-            return ""
 
 class hgrequest(object):
     def __init__(self, inp=None, out=None, env=None):
         self.inp = inp or sys.stdin
         self.out = out or sys.stdout
         self.env = env or os.environ
         self.form = cgi.parse(self.inp, self.env, keep_blank_values=1)
 
     def write(self, *things):
         for thing in things:
             if hasattr(thing, "__iter__"):
                 for part in thing:
                     self.write(part)
             else:
                 try:
                     self.out.write(str(thing))
                 except socket.error, inst:
                     if inst[0] != errno.ECONNRESET:
                         raise
 
     def header(self, headers=[('Content-type','text/html')]):
         for header in headers:
             self.out.write("%s: %s\r\n" % header)
         self.out.write("\r\n")
 
     def httphdr(self, type, file="", size=0):
 
         headers = [('Content-type', type)]
         if file:
             headers.append(('Content-disposition', 'attachment; filename=%s' % file))
         if size > 0:
             headers.append(('Content-length', str(size)))
         self.header(headers)
 
 class templater(object):
     def __init__(self, mapfile, filters={}, defaults={}):
         self.cache = {}
         self.map = {}
         self.base = os.path.dirname(mapfile)
         self.filters = filters
         self.defaults = defaults
 
         for l in file(mapfile):
             m = re.match(r'(\S+)\s*=\s*"(.*)"$', l)
             if m:
                 self.cache[m.group(1)] = m.group(2)
             else:
                 m = re.match(r'(\S+)\s*=\s*(\S+)', l)
                 if m:
                     self.map[m.group(1)] = os.path.join(self.base, m.group(2))
                 else:
                     raise LookupError(_("unknown map entry '%s'") % l)
 
     def __call__(self, t, **map):
         m = self.defaults.copy()
         m.update(map)
         try:
             tmpl = self.cache[t]
         except KeyError:
             tmpl = self.cache[t] = file(self.map[t]).read()
         return self.template(tmpl, self.filters, **m)
 
     def template(self, tmpl, filters={}, **map):
         while tmpl:
             m = re.search(r"#([a-zA-Z0-9]+)((%[a-zA-Z0-9]+)*)((\|[a-zA-Z0-9]+)*)#", tmpl)
             if m:
                 yield tmpl[:m.start(0)]
                 v = map.get(m.group(1), "")
                 v = callable(v) and v(**map) or v
 
                 format = m.group(2)
                 fl = m.group(4)
 
                 if format:
                     q = v.__iter__
                     for i in q():
                         lm = map.copy()
                         lm.update(i)
                         yield self(format[1:], **lm)
 
                     v = ""
 
                 elif fl:
                     for f in fl.split("|")[1:]:
                         v = filters[f](v)
 
                 yield v
                 tmpl = tmpl[m.end(0):]
             else:
                 yield tmpl
                 return
 
 common_filters = {
     "escape": lambda x: cgi.escape(x, True),
     "urlescape": urllib.quote,
     "strip": lambda x: x.strip(),
     "age": age,
     "date": lambda x: util.datestr(x),
     "addbreaks": nl2br,
     "obfuscate": obfuscate,
     "short": (lambda x: x[:12]),
     "firstline": (lambda x: x.splitlines(1)[0]),
     "permissions": (lambda x: x and "-rwxr-xr-x" or "-rw-r--r--"),
     "rfc822date": lambda x: util.datestr(x, "%a, %d %b %Y %H:%M:%S"),
     }
 
 class hgweb(object):
     def __init__(self, repo, name=None):
         if type(repo) == type(""):
             self.repo = hg.repository(ui.ui(), repo)
         else:
             self.repo = repo
 
         self.mtime = -1
         self.reponame = name
         self.archives = 'zip', 'gz', 'bz2'
 
     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.repo.ui.config("web", "maxchanges", 10))
             self.maxfiles = int(self.repo.ui.config("web", "maxfiles", 10))
             self.allowpull = self.repo.ui.configbool("web", "allowpull", True)
 
     def archivelist(self, nodeid):
         for i in self.archives:
             if self.repo.ui.configbool("web", "allow" + i, False):
                 yield {"type" : i, "node" : nodeid}
 
     def listfiles(self, files, mf):
         for f in files[:self.maxfiles]:
             yield self.t("filenodelink", node=hex(mf[f]), file=f)
         if len(files) > self.maxfiles:
             yield self.t("fileellipses")
 
     def listfilediffs(self, files, changeset):
         for f in files[:self.maxfiles]:
             yield self.t("filedifflink", node=hex(changeset), file=f)
         if len(files) > self.maxfiles:
             yield self.t("fileellipses")
 
     def siblings(self, siblings=[], rev=None, hiderev=None, **args):
         if not rev:
             rev = lambda x: ""
         siblings = [s for s in siblings if s != nullid]
         if len(siblings) == 1 and rev(siblings[0]) == hiderev:
             return
         for s in siblings:
             yield dict(node=hex(s), rev=rev(s), **args)
 
     def renamelink(self, fl, node):
         r = fl.renamed(node)
         if r:
             return [dict(file=r[0], node=hex(r[1]))]
         return []
 
     def showtag(self, t1, node=nullid, **args):
         for t in self.repo.nodetags(node):
              yield self.t(t1, tag=t, **args)
 
     def diff(self, 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 = [0]
         def diffblock(diff, f, fn):
             yield self.t("diffblock",
                          lines=prettyprintlines(diff),
                          parity=parity[0],
                          file=f,
                          filenode=hex(fn or nullid))
             parity[0] = 1 - parity[0]
 
         def prettyprintlines(diff):
             for l in diff.splitlines(1):
                 if l.startswith('+'):
                     yield self.t("difflineplus", line=l)
                 elif l.startswith('-'):
                     yield self.t("difflineminus", line=l)
                 elif l.startswith('@'):
                     yield self.t("difflineat", line=l)
                 else:
                     yield self.t("diffline", line=l)
 
         r = self.repo
         cl = r.changelog
         mf = r.manifest
         change1 = cl.read(node1)
         change2 = cl.read(node2)
         mmap1 = mf.read(change1[0])
         mmap2 = mf.read(change2[0])
         date1 = util.datestr(change1[2])
         date2 = util.datestr(change2[2])
 
         modified, added, removed, deleted, unknown = r.changes(node1, node2)
         if files:
             modified, added, removed = map(lambda x: filterfiles(files, x),
                                            (modified, added, removed))
 
         diffopts = self.repo.ui.diffopts()
         showfunc = diffopts['showfunc']
         ignorews = diffopts['ignorews']
         for f in modified:
             to = r.file(f).read(mmap1[f])
             tn = r.file(f).read(mmap2[f])
             yield diffblock(mdiff.unidiff(to, date1, tn, date2, f,
                             showfunc=showfunc, ignorews=ignorews), f, tn)
         for f in added:
             to = None
             tn = r.file(f).read(mmap2[f])
             yield diffblock(mdiff.unidiff(to, date1, tn, date2, f,
                             showfunc=showfunc, ignorews=ignorews), f, tn)
         for f in removed:
             to = r.file(f).read(mmap1[f])
             tn = None
             yield diffblock(mdiff.unidiff(to, date1, tn, date2, f,
                             showfunc=showfunc, ignorews=ignorews), f, tn)
 
     def changelog(self, pos):
         def changenav(**map):
             def seq(factor, maxchanges=None):
                 if maxchanges:
                     yield maxchanges
                     if maxchanges >= 20 and maxchanges <= 40:
                         yield 50
                 else:
                     yield 1 * factor
                     yield 3 * factor
                 for f in seq(factor * 10):
                     yield f
 
             l = []
             last = 0
             for f in seq(1, self.maxchanges):
                 if f < self.maxchanges or f <= last:
                     continue
                 if f > count:
                     break
                 last = f
                 r = "%d" % f
                 if pos + f < count:
                     l.append(("+" + r, pos + f))
                 if pos - f >= 0:
                     l.insert(0, ("-" + r, pos - f))
 
             yield {"rev": 0, "label": "(0)"}
 
             for label, rev in l:
                 yield {"label": label, "rev": rev}
 
             yield {"label": "tip", "rev": "tip"}
 
         def changelist(**map):
             parity = (start - end) & 1
             cl = self.repo.changelog
             l = [] # build a list in forward order for efficiency
             for i in range(start, end):
                 n = cl.node(i)
                 changes = cl.read(n)
                 hn = hex(n)
 
                 l.insert(0, {"parity": parity,
                              "author": changes[1],
                              "parent": self.siblings(cl.parents(n), cl.rev,
                                                      cl.rev(n) - 1),
                              "child": self.siblings(cl.children(n), cl.rev,
                                                     cl.rev(n) + 1),
                              "changelogtag": self.showtag("changelogtag",n),
                              "manifest": hex(changes[0]),
                              "desc": changes[4],
                              "date": changes[2],
                              "files": self.listfilediffs(changes[3], n),
                              "rev": i,
                              "node": hn})
                 parity = 1 - parity
 
             for e in l:
                 yield e
 
         cl = self.repo.changelog
         mf = cl.read(cl.tip())[0]
         count = cl.count()
         start = max(0, pos - self.maxchanges + 1)
         end = min(count, start + self.maxchanges)
         pos = end - 1
 
         yield self.t('changelog',
                      changenav=changenav,
                      manifest=hex(mf),
                      rev=pos, changesets=count, entries=changelist)
 
     def search(self, query):
 
         def changelist(**map):
             cl = self.repo.changelog
             count = 0
             qw = query.lower().split()
 
             def revgen():
                 for i in range(cl.count() - 1, 0, -100):
                     l = []
                     for j in range(max(0, i - 100), i):
                         n = cl.node(j)
                         changes = cl.read(n)
                         l.append((n, j, changes))
                     l.reverse()
                     for e in l:
                         yield e
 
             for n, i, changes in revgen():
                 miss = 0
                 for q in qw:
                     if not (q in changes[1].lower() or
                             q in changes[4].lower() or
                             q in " ".join(changes[3][:20]).lower()):
                         miss = 1
                         break
                 if miss:
                     continue
 
                 count += 1
                 hn = hex(n)
 
                 yield self.t('searchentry',
                              parity=count & 1,
                              author=changes[1],
                              parent=self.siblings(cl.parents(n), cl.rev),
                              child=self.siblings(cl.children(n), cl.rev),
                              changelogtag=self.showtag("changelogtag",n),
                              manifest=hex(changes[0]),
                              desc=changes[4],
                              date=changes[2],
                              files=self.listfilediffs(changes[3], n),
                              rev=i,
                              node=hn)
 
                 if count >= self.maxchanges:
                     break
 
         cl = self.repo.changelog
         mf = cl.read(cl.tip())[0]
 
         yield self.t('search',
                      query=query,
                      manifest=hex(mf),
                      entries=changelist)
 
     def changeset(self, nodeid):
         cl = self.repo.changelog
         n = self.repo.lookup(nodeid)
         nodeid = hex(n)
         changes = cl.read(n)
         p1 = cl.parents(n)[0]
 
         files = []
         mf = self.repo.manifest.read(changes[0])
         for f in changes[3]:
             files.append(self.t("filenodelink",
                                 filenode=hex(mf.get(f, nullid)), file=f))
 
         def diff(**map):
             yield self.diff(p1, n, None)
 
         yield self.t('changeset',
                      diff=diff,
                      rev=cl.rev(n),
                      node=nodeid,
                      parent=self.siblings(cl.parents(n), cl.rev),
                      child=self.siblings(cl.children(n), cl.rev),
                      changesettag=self.showtag("changesettag",n),
                      manifest=hex(changes[0]),
                      author=changes[1],
                      desc=changes[4],
                      date=changes[2],
                      files=files,
                      archives=self.archivelist(nodeid))
 
     def filelog(self, f, filenode):
         cl = self.repo.changelog
         fl = self.repo.file(f)
         filenode = hex(fl.lookup(filenode))
         count = fl.count()
 
         def entries(**map):
             l = []
             parity = (count - 1) & 1
 
             for i in range(count):
                 n = fl.node(i)
                 lr = fl.linkrev(n)
                 cn = cl.node(lr)
                 cs = cl.read(cl.node(lr))
 
                 l.insert(0, {"parity": parity,
                              "filenode": hex(n),
                              "filerev": i,
                              "file": f,
                              "node": hex(cn),
                              "author": cs[1],
                              "date": cs[2],
                              "rename": self.renamelink(fl, n),
                              "parent": self.siblings(fl.parents(n),
                                                      fl.rev, file=f),
                              "child": self.siblings(fl.children(n),
                                                     fl.rev, file=f),
                              "desc": cs[4]})
                 parity = 1 - parity
 
             for e in l:
                 yield e
 
         yield self.t("filelog", file=f, filenode=filenode, entries=entries)
 
     def filerevision(self, f, node):
         fl = self.repo.file(f)
         n = fl.lookup(node)
         node = hex(n)
         text = fl.read(n)
         changerev = fl.linkrev(n)
         cl = self.repo.changelog
         cn = cl.node(changerev)
         cs = cl.read(cn)
         mfn = cs[0]
 
         mt = mimetypes.guess_type(f)[0]
         rawtext = text
         if util.binary(text):
             text = "(binary:%s)" % mt
 
         def lines():
             for l, t in enumerate(text.splitlines(1)):
                 yield {"line": t,
                        "linenumber": "% 6d" % (l + 1),
                        "parity": l & 1}
 
         yield self.t("filerevision",
                      file=f,
                      filenode=node,
                      path=up(f),
                      text=lines(),
                      raw=rawtext,
                      mimetype=mt,
                      rev=changerev,
                      node=hex(cn),
                      manifest=hex(mfn),
                      author=cs[1],
                      date=cs[2],
                      parent=self.siblings(fl.parents(n), fl.rev, file=f),
                      child=self.siblings(fl.children(n), fl.rev, file=f),
                      rename=self.renamelink(fl, n),
                      permissions=self.repo.manifest.readflags(mfn)[f])
 
     def fileannotate(self, f, node):
         bcache = {}
         ncache = {}
         fl = self.repo.file(f)
         n = fl.lookup(node)
         node = hex(n)
         changerev = fl.linkrev(n)
 
         cl = self.repo.changelog
         cn = cl.node(changerev)
         cs = cl.read(cn)
         mfn = cs[0]
 
         def annotate(**map):
             parity = 1
             last = None
             for r, l in fl.annotate(n):
                 try:
                     cnode = ncache[r]
                 except KeyError:
                     cnode = ncache[r] = self.repo.changelog.node(r)
 
                 try:
                     name = bcache[r]
                 except KeyError:
                     cl = self.repo.changelog.read(cnode)
                     bcache[r] = name = self.repo.ui.shortuser(cl[1])
 
                 if last != cnode:
                     parity = 1 - parity
                     last = cnode
 
                 yield {"parity": parity,
                        "node": hex(cnode),
                        "rev": r,
                        "author": name,
                        "file": f,
                        "line": l}
 
         yield self.t("fileannotate",
                      file=f,
                      filenode=node,
                      annotate=annotate,
                      path=up(f),
                      rev=changerev,
                      node=hex(cn),
                      manifest=hex(mfn),
                      author=cs[1],
                      date=cs[2],
                      rename=self.renamelink(fl, n),
                      parent=self.siblings(fl.parents(n), fl.rev, file=f),
                      child=self.siblings(fl.children(n), fl.rev, file=f),
                      permissions=self.repo.manifest.readflags(mfn)[f])
 
     def manifest(self, mnode, path):
         man = self.repo.manifest
         mn = man.lookup(mnode)
         mnode = hex(mn)
         mf = man.read(mn)
         rev = man.rev(mn)
         node = self.repo.changelog.node(rev)
         mff = man.readflags(mn)
 
         files = {}
 
         p = path[1:]
         if p and p[-1] != "/":
             p += "/"
         l = len(p)
 
         for f,n in mf.items():
             if f[:l] != p:
                 continue
             remain = f[l:]
             if "/" in remain:
                 short = remain[:remain.find("/") + 1] # bleah
                 files[short] = (f, None)
             else:
                 short = os.path.basename(remain)
                 files[short] = (f, n)
 
         def filelist(**map):
             parity = 0
             fl = files.keys()
             fl.sort()
             for f in fl:
                 full, fnode = files[f]
                 if not fnode:
                     continue
 
                 yield {"file": full,
                        "manifest": mnode,
                        "filenode": hex(fnode),
                        "parity": parity,
                        "basename": f,
                        "permissions": mff[full]}
                 parity = 1 - parity
 
         def dirlist(**map):
             parity = 0
             fl = files.keys()
             fl.sort()
             for f in fl:
                 full, fnode = files[f]
                 if fnode:
                     continue
 
                 yield {"parity": parity,
                        "path": os.path.join(path, f),
                        "manifest": mnode,
                        "basename": f[:-1]}
                 parity = 1 - parity
 
         yield self.t("manifest",
                      manifest=mnode,
                      rev=rev,
                      node=hex(node),
                      path=path,
                      up=up(path),
                      fentries=filelist,
                      dentries=dirlist,
                      archives=self.archivelist(hex(node)))
 
     def tags(self):
         cl = self.repo.changelog
         mf = cl.read(cl.tip())[0]
 
         i = self.repo.tagslist()
         i.reverse()
 
         def entries(notip=False, **map):
             parity = 0
             for k,n in i:
                 if notip and k == "tip": continue
                 yield {"parity": parity,
                        "tag": k,
                        "tagmanifest": hex(cl.read(n)[0]),
                        "date": cl.read(n)[2],
                        "node": hex(n)}
                 parity = 1 - parity
 
         yield self.t("tags",
                      manifest=hex(mf),
                      entries=lambda **x: entries(False, **x),
                      entriesnotip=lambda **x: entries(True, **x))
 
     def summary(self):
         cl = self.repo.changelog
         mf = cl.read(cl.tip())[0]
 
         i = self.repo.tagslist()
         i.reverse()
 
         def tagentries(**map):
             parity = 0
             count = 0
             for k,n in i:
                 if k == "tip": # skip tip
                     continue;
 
                 count += 1
                 if count > 10: # limit to 10 tags
                     break;
 
                 c = cl.read(n)
                 m = c[0]
                 t = c[2]
 
                 yield self.t("tagentry",
                              parity = parity,
                              tag = k,
                              node = hex(n),
                              date = t,
                              tagmanifest = hex(m))
                 parity = 1 - parity
 
         def changelist(**map):
             parity = 0
             cl = self.repo.changelog
             l = [] # build a list in forward order for efficiency
             for i in range(start, end):
                 n = cl.node(i)
                 changes = cl.read(n)
                 hn = hex(n)
                 t = changes[2]
 
                 l.insert(0, self.t(
                     'shortlogentry',
                     parity = parity,
                     author = changes[1],
                     manifest = hex(changes[0]),
                     desc = changes[4],
                     date = t,
                     rev = i,
                     node = hn))
                 parity = 1 - parity
 
             yield l
 
         cl = self.repo.changelog
         mf = cl.read(cl.tip())[0]
         count = cl.count()
         start = max(0, count - self.maxchanges)
         end = min(count, start + self.maxchanges)
         pos = end - 1
 
         yield self.t("summary",
                  desc = self.repo.ui.config("web", "description", "unknown"),
                  owner = (self.repo.ui.config("ui", "username") or # preferred
                           self.repo.ui.config("web", "contact") or # deprecated
                           self.repo.ui.config("web", "author", "unknown")), # also
                  lastchange = (0, 0), # FIXME
                  manifest = hex(mf),
                  tags = tagentries,
                  shortlog = changelist)
 
     def filediff(self, file, changeset):
         cl = self.repo.changelog
         n = self.repo.lookup(changeset)
         changeset = hex(n)
         p1 = cl.parents(n)[0]
         cs = cl.read(n)
         mf = self.repo.manifest.read(cs[0])
 
         def diff(**map):
             yield self.diff(p1, n, file)
 
         yield self.t("filediff",
                      file=file,
                      filenode=hex(mf.get(file, nullid)),
                      node=changeset,
                      rev=self.repo.changelog.rev(n),
                      parent=self.siblings(cl.parents(n), cl.rev),
                      child=self.siblings(cl.children(n), cl.rev),
                      diff=diff)
 
     def archive(self, req, cnode, type):
         cs = self.repo.changelog.read(cnode)
         mnode = cs[0]
         mf = self.repo.manifest.read(mnode)
         rev = self.repo.manifest.rev(mnode)
         reponame = re.sub(r"\W+", "-", self.reponame)
         name = "%s-%s/" % (reponame, short(cnode))
 
         files = mf.keys()
         files.sort()
 
         if type == 'zip':
             tmp = tempfile.mkstemp()[1]
             try:
                 zf = zipfile.ZipFile(tmp, "w", zipfile.ZIP_DEFLATED)
 
                 for f in files:
                     zf.writestr(name + f, self.repo.file(f).read(mf[f]))
                 zf.close()
 
                 f = open(tmp, 'r')
                 req.httphdr('application/zip', name[:-1] + '.zip',
                         os.path.getsize(tmp))
                 req.write(f.read())
                 f.close()
             finally:
                 os.unlink(tmp)
 
         else:
             tf = tarfile.TarFile.open(mode='w|' + type, fileobj=req.out)
             mff = self.repo.manifest.readflags(mnode)
             mtime = int(time.time())
 
             if type == "gz":
                 encoding = "gzip"
             else:
                 encoding = "x-bzip2"
             req.header([('Content-type', 'application/x-tar'),
                     ('Content-disposition', 'attachment; filename=%s%s%s' %
                         (name[:-1], '.tar.', type)),
                     ('Content-encoding', encoding)])
             for fname in files:
                 rcont = self.repo.file(fname).read(mf[fname])
                 finfo = tarfile.TarInfo(name + fname)
                 finfo.mtime = mtime
                 finfo.size = len(rcont)
                 finfo.mode = mff[fname] and 0755 or 0644
                 tf.addfile(finfo, StringIO.StringIO(rcont))
             tf.close()
 
     # add tags to things
     # tags -> list of changesets corresponding to tags
     # find tag, changeset, file
 
     def run(self, req=hgrequest()):
         def clean(path):
             p = os.path.normpath(path)
             if p[:2] == "..":
                 raise "suspicious path"
             return p
 
         def header(**map):
             yield self.t("header", **map)
 
         def footer(**map):
             yield self.t("footer", **map)
 
         def expand_form(form):
             shortcuts = {
                 'cl': [('cmd', ['changelog']), ('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)]
             }
 
             for k in shortcuts.iterkeys():
                 if form.has_key(k):
                     for name, value in shortcuts[k]:
                         if value is None:
                             value = form[k]
                         form[name] = value
                     del form[k]
 
         self.refresh()
 
         expand_form(req.form)
 
         t = self.repo.ui.config("web", "templates", templatepath())
         static = self.repo.ui.config("web", "static", os.path.join(t,"static"))
         m = os.path.join(t, "map")
         style = self.repo.ui.config("web", "style", "")
         if req.form.has_key('style'):
             style = req.form['style'][0]
         if style:
             b = os.path.basename("map-" + style)
             p = os.path.join(t, b)
             if os.path.isfile(p):
                 m = p
 
         port = req.env["SERVER_PORT"]
         port = port != "80" and (":" + port) or ""
         uri = req.env["REQUEST_URI"]
         if "?" in uri:
             uri = uri.split("?")[0]
         url = "http://%s%s%s" % (req.env["SERVER_NAME"], port, uri)
         if not self.reponame:
             self.reponame = (self.repo.ui.config("web", "name")
                              or uri.strip('/') or self.repo.root)
 
         self.t = templater(m, common_filters,
                            {"url": url,
                             "repo": self.reponame,
                             "header": header,
                             "footer": footer,
                            })
 
         if not req.form.has_key('cmd'):
             req.form['cmd'] = [self.t.cache['default'],]
 
         if req.form['cmd'][0] == 'changelog':
             c = self.repo.changelog.count() - 1
             hi = c
             if req.form.has_key('rev'):
                 hi = req.form['rev'][0]
                 try:
                     hi = self.repo.changelog.rev(self.repo.lookup(hi))
                 except hg.RepoError:
                     req.write(self.search(hi))
                     return
 
             req.write(self.changelog(hi))
 
         elif req.form['cmd'][0] == 'changeset':
             req.write(self.changeset(req.form['node'][0]))
 
         elif req.form['cmd'][0] == 'manifest':
             req.write(self.manifest(req.form['manifest'][0],
                                     clean(req.form['path'][0])))
 
         elif req.form['cmd'][0] == 'tags':
             req.write(self.tags())
 
         elif req.form['cmd'][0] == 'summary':
             req.write(self.summary())
 
         elif req.form['cmd'][0] == 'filediff':
             req.write(self.filediff(clean(req.form['file'][0]),
                                     req.form['node'][0]))
 
         elif req.form['cmd'][0] == 'file':
             req.write(self.filerevision(clean(req.form['file'][0]),
                                         req.form['filenode'][0]))
 
         elif req.form['cmd'][0] == 'annotate':
             req.write(self.fileannotate(clean(req.form['file'][0]),
                                         req.form['filenode'][0]))
 
         elif req.form['cmd'][0] == 'filelog':
             req.write(self.filelog(clean(req.form['file'][0]),
                                    req.form['filenode'][0]))
 
         elif req.form['cmd'][0] == 'heads':
             req.httphdr("application/mercurial-0.1")
             h = self.repo.heads()
             req.write(" ".join(map(hex, h)) + "\n")
 
         elif req.form['cmd'][0] == 'branches':
             req.httphdr("application/mercurial-0.1")
             nodes = []
             if req.form.has_key('nodes'):
                 nodes = map(bin, req.form['nodes'][0].split(" "))
             for b in self.repo.branches(nodes):
                 req.write(" ".join(map(hex, b)) + "\n")
 
         elif req.form['cmd'][0] == 'between':
             req.httphdr("application/mercurial-0.1")
             nodes = []
             if req.form.has_key('pairs'):
                 pairs = [map(bin, p.split("-"))
                          for p in req.form['pairs'][0].split(" ")]
             for b in self.repo.between(pairs):
                 req.write(" ".join(map(hex, b)) + "\n")
 
         elif req.form['cmd'][0] == 'changegroup':
             req.httphdr("application/mercurial-0.1")
             nodes = []
             if not self.allowpull:
                 return
 
             if req.form.has_key('roots'):
                 nodes = map(bin, req.form['roots'][0].split(" "))
 
             z = zlib.compressobj()
             f = self.repo.changegroup(nodes, 'serve')
             while 1:
                 chunk = f.read(4096)
                 if not chunk:
                     break
                 req.write(z.compress(chunk))
 
             req.write(z.flush())
 
         elif req.form['cmd'][0] == 'archive':
             changeset = self.repo.lookup(req.form['node'][0])
             type = req.form['type'][0]
             if (type in self.archives and
                 self.repo.ui.configbool("web", "allow" + type, False)):
                 self.archive(req, changeset, type)
                 return
 
             req.write(self.t("error"))
 
         elif req.form['cmd'][0] == 'static':
             fname = req.form['file'][0]
             req.write(staticfile(static, fname)
                       or self.t("error", error="%r not found" % fname))
 
         else:
             req.write(self.t("error"))
 
 def create_server(repo):
 
     def openlog(opt, default):
         if opt and opt != '-':
             return open(opt, 'w')
         return default
 
     address = repo.ui.config("web", "address", "")
     port = int(repo.ui.config("web", "port", 8000))
     use_ipv6 = repo.ui.configbool("web", "ipv6")
     accesslog = openlog(repo.ui.config("web", "accesslog", "-"), sys.stdout)
     errorlog = openlog(repo.ui.config("web", "errorlog", "-"), sys.stderr)
 
     class IPv6HTTPServer(BaseHTTPServer.HTTPServer):
         address_family = getattr(socket, 'AF_INET6', None)
 
         def __init__(self, *args, **kwargs):
             if self.address_family is None:
                 raise hg.RepoError(_('IPv6 not available on this system'))
             BaseHTTPServer.HTTPServer.__init__(self, *args, **kwargs)
 
     class hgwebhandler(BaseHTTPServer.BaseHTTPRequestHandler):
         def log_error(self, format, *args):
             errorlog.write("%s - - [%s] %s\n" % (self.address_string(),
                                                  self.log_date_time_string(),
                                                  format % args))
 
         def log_message(self, format, *args):
             accesslog.write("%s - - [%s] %s\n" % (self.address_string(),
                                                   self.log_date_time_string(),
                                                   format % args))
 
         def do_POST(self):
             try:
                 self.do_hgweb()
             except socket.error, inst:
                 if inst[0] != errno.EPIPE:
                     raise
 
         def do_GET(self):
             self.do_POST()
 
         def do_hgweb(self):
             query = ""
             p = self.path.find("?")
             if p:
                 query = self.path[p + 1:]
                 query = query.replace('+', ' ')
 
             env = {}
             env['GATEWAY_INTERFACE'] = 'CGI/1.1'
             env['REQUEST_METHOD'] = self.command
             env['SERVER_NAME'] = self.server.server_name
             env['SERVER_PORT'] = str(self.server.server_port)
             env['REQUEST_URI'] = "/"
             if query:
                 env['QUERY_STRING'] = query
             host = self.address_string()
             if host != self.client_address[0]:
                 env['REMOTE_HOST'] = host
                 env['REMOTE_ADDR'] = self.client_address[0]
 
             if self.headers.typeheader is None:
                 env['CONTENT_TYPE'] = self.headers.type
             else:
                 env['CONTENT_TYPE'] = self.headers.typeheader
             length = self.headers.getheader('content-length')
             if length:
                 env['CONTENT_LENGTH'] = length
             accept = []
             for line in self.headers.getallmatchingheaders('accept'):
                 if line[:1] in "\t\n\r ":
                     accept.append(line.strip())
                 else:
                     accept = accept + line[7:].split(',')
             env['HTTP_ACCEPT'] = ','.join(accept)
 
             req = hgrequest(self.rfile, self.wfile, env)
             self.send_response(200, "Script output follows")
             hg.run(req)
 
     hg = hgweb(repo)
     if use_ipv6:
         return IPv6HTTPServer((address, port), hgwebhandler)
     else:
         return BaseHTTPServer.HTTPServer((address, port), hgwebhandler)
 
 # This is a stopgap
 class hgwebdir(object):
     def __init__(self, config):
         def cleannames(items):
             return [(name.strip('/'), path) for name, path in items]
 
         if type(config) == type([]):
             self.repos = cleannames(config)
         elif type(config) == type({}):
             self.repos = cleannames(config.items())
             self.repos.sort()
         else:
             cp = ConfigParser.SafeConfigParser()
             cp.read(config)
             self.repos = cleannames(cp.items("paths"))
             self.repos.sort()
 
     def run(self, req=hgrequest()):
         def header(**map):
             yield tmpl("header", **map)
 
         def footer(**map):
             yield tmpl("footer", **map)
 
         m = os.path.join(templatepath(), "map")
         tmpl = templater(m, common_filters,
                          {"header": header, "footer": footer})
 
         def entries(**map):
             parity = 0
             for name, path in self.repos:
                 u = ui.ui()
                 try:
                     u.readconfig(os.path.join(path, '.hg', 'hgrc'))
                 except IOError:
                     pass
                 get = u.config
 
                 url = ('/'.join([req.env["REQUEST_URI"].split('?')[0], name])
                        .replace("//", "/"))
 
                 # update time with local timezone
                 try:
                     d = (get_mtime(path), util.makedate()[1])
                 except OSError:
                     continue
 
                 yield dict(contact=(get("ui", "username") or # preferred
                                     get("web", "contact") or # deprecated
                                     get("web", "author", "unknown")), # also
                            name=get("web", "name", name),
                            url=url,
                            parity=parity,
                            shortdesc=get("web", "description", "unknown"),
                            lastupdate=d)
 
                 parity = 1 - parity
 
         virtual = req.env.get("PATH_INFO", "").strip('/')
         if virtual:
             real = dict(self.repos).get(virtual)
             if real:
                 try:
                     hgweb(real).run(req)
                 except IOError, inst:
                     req.write(tmpl("error", error=inst.strerror))
                 except hg.RepoError, inst:
                     req.write(tmpl("error", error=str(inst)))
             else:
                 req.write(tmpl("notfound", repo=virtual))
         else:
             if req.form.has_key('static'):
                 static = os.path.join(templatepath(), "static")
                 fname = req.form['static'][0]
                 req.write(staticfile(static, fname)
                           or tmpl("error", error="%r not found" % fname))
             else:
                 req.write(tmpl("index", entries=entries))