Bump revision for 4.0 release
[geekigeeki.git] / geekigeeki.py
index 8bb4d366eabde261e4fdb104dcb7a3c73e1d3083..71920f370b9dd74f6c0df8b057bf308a404fa9f0 100755 (executable)
@@ -3,7 +3,7 @@
 #
 # Copyright 1999, 2000 Martin Pool <mbp@humbug.org.au>
 # Copyright 2002 Gerardo Poggiali
-# Copyright 2007, 2008 Bernardo Innocenti <bernie@codewiz.org>
+# Copyright 2007, 2008, 2009 Bernie Innocenti <bernie@codewiz.org>
 #
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU General Public License as published by
 # You should have received a copy of the GNU General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 
-__version__ = '$Id$'[4:12]
+__version__ = '4.0-' + '$Id$'[4:11]
 
 from time import clock
 start_time = clock()
+title_done = False
 
 import cgi, sys, os, re, errno, stat
-from os import path, environ
 
-# Regular expression defining a WikiWord
-# (but this definition is also assumed in other places)
-word_re = re.compile(r"^\b((([A-Z][a-z0-9]+){2,}/)*([A-Z][a-z0-9]+){2,})\b$")
+image_ext = 'png|gif|jpg|jpeg|bmp|ico'
+video_ext = "ogg|ogv|oga" # Not supported by Firefox 3.5: mkv|mpg|mpeg|mp4|avi|asf|flv|wmv|qt
+image_re = re.compile(r".*\.(" + image_ext + "|" +  video_ext + ")", re.IGNORECASE)
+video_re = re.compile(r".*\.(" + video_ext + ")", re.IGNORECASE)
 # FIXME: we accept stuff like foo/../bar and we shouldn't
-file_re = re.compile(r"^\b([A-Za-z0-9_\-][A-Za-z0-9_\.\-/]*)\b$")
-img_re = re.compile(r"^.*\.(png|gif|jpg|jpeg)$", re.IGNORECASE)
-url_re = re.compile(r"^[a-z]{3,8}://[^\s'\"]+\S$")
-
-title_done = False
-
+file_re  = re.compile(r"([A-Za-z0-9_\-][A-Za-z0-9_\.\-/]*)")
+url_re   = re.compile(r"[a-z]{3,8}://[^\s'\"]+\S")
+ext_re   = re.compile(r"\.([^\./]+)$")
 
 # CGI stuff ---------------------------------------------------------
-
 def script_name():
-    return environ.get('SCRIPT_NAME', '')
+    return os.environ.get('SCRIPT_NAME', '')
+
+def query_string():
+    path_info = os.environ.get('PATH_INFO', '')
+    if len(path_info) and path_info[0] == '/':
+        return path_info[1:] or 'FrontPage'
+    else:
+        return os.environ.get('QUERY_STRING', '') or 'FrontPage'
 
 def privileged_path():
     return privileged_url or script_name()
 
 def remote_user():
-    user = environ.get('REMOTE_USER', '')
+    user = os.environ.get('REMOTE_USER', '')
     if user is None or user == '' or user == 'anonymous':
         user = 'AnonymousCoward'
     return user
 
 def remote_host():
-    return environ.get('REMOTE_ADDR', '')
+    return os.environ.get('REMOTE_ADDR', '')
 
 def get_hostname(addr):
     try:
         from socket import gethostbyaddr
         return gethostbyaddr(addr)[0] + ' (' + addr + ')'
-    except Exception, er:
+    except Exception:
         return addr
 
+def is_external_url(pathname):
+    return (url_re.match(pathname) or pathname.startswith('/'))
+
 def relative_url(pathname, privileged=False):
-    if not (url_re.match(pathname) or pathname.startswith('/')):
+    if not is_external_url(pathname):
         if privileged:
             url = privileged_path()
         else:
             url = script_name()
         pathname = url + '/' + pathname
-    return pathname
+    return cgi.escape(pathname, quote=True)
 
 def permalink(s):
     return re.sub(' ', '-', re.sub('[^a-z0-9_ ]', '', s.lower()).strip())
 
+def humanlink(s):
+    return re.sub(r'(?:.*[/:]|)([^:/\.]+)(?:\.[^/:]+|)$', r'\1', s.replace('_', ' '))
+
+# Split arg lists like "blah| blah blah| width=100 | align = center",
+# return a list containing anonymous arguments and a map containing the named arguments
+def parse_args(s):
+    args = []
+    kwargs = {} 
+    for arg in s.strip('<[{}]>').split('|'):
+        m = re.match('\s*(\w+)\s*=\s*(.+)\s*', arg)
+        if m is not None:
+            kwargs[m.group(1)] = m.group(2)
+        else:
+            args.append(arg.strip())
+    return (args, kwargs)
+
+def url_args(kvargs):
+    argv = []
+    for k, v in kvargs.items():
+        argv.append(k + '=' + v)
+    if argv:
+        return '?' + '&amp;'.join(argv)
+    return ''
+
 # Formatting stuff --------------------------------------------------
 def emit_header(mime_type="text/html"):
-    print "Content-type: " + mime_type + "; charset=utf-8\n"
+    print("Content-type: " + mime_type + "; charset=utf-8\n")
 
 def send_guru(msg_text, msg_type):
     if not msg_text: return
-    print '<pre id="guru" onclick="this.style.display = \'none\'" class="' + msg_type + '">'
+    print('<pre id="guru" onclick="this.style.display = \'none\'" class="' + msg_type + '">')
     if msg_type == 'error':
-        print '    Software Failure.  Press left mouse button to continue.\n'
-    print msg_text
+        print('    Software Failure.  Press left mouse button to continue.\n')
+    print(msg_text)
     if msg_type == 'error':
-        print '      Guru Meditation #DEADBEEF.ABADC0DE'
-    print '</pre>'
-    # FIXME: This little JS snippet is harder to pass than ACID 3.0 
-    print """
-    <script language="JavaScript" type="text/javascript">
-        var guru = document.getElementById('guru');
-        // Firefox 2.0 doesn't take border-color, but returns border-top-color fine
-        var color = document.defaultView.getComputedStyle(guru,null).getPropertyValue('border-top-color');
-
-        function guruOn() {
-            guru.style.setProperty('border-color', color, '');
-            setTimeout('guruOff()', 1000);
-        }
-        function guruOff() {
-            guru.style.setProperty('border-color', '#000000', '');
-            setTimeout('guruOn()', 1000);
-        }
-        // Safari 2.0 returns this rgba crap
-        // Konqueror 3.5.6 doesn't seem to support computed properties
-        if (color && color != 'rgba(0, 0, 0, 0)') {
-            //window.alert("enabled! color='" + color + "'");
-            guruOn();
-        }
-    </script>"""
+        print '\n           Guru Meditation #DEADBEEF.ABADC0DE'
+    print('</pre><script language="JavaScript" type="text/javascript" src="%s" defer="defer"></script>' \
+        % relative_url('sys/GuruMeditation.js'))
 
-def send_title(name, text="Limbo", msg_text=None, msg_type='error'):
+def send_title(name, text="Limbo", msg_text=None, msg_type='error', writable=False):
     global title_done
     if title_done: return
 
     # Head
     emit_header()
-    print '<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"'
-    print '  "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">'
-    print '<html xmlns="http://www.w3.org/1999/xhtml" lang="en" xml:lang="en">'
+    print('<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"')
+    print('  "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">')
+    print('<html xmlns="http://www.w3.org/1999/xhtml" lang="en" xml:lang="en">')
 
-    site_name = globals().get('site_name', 'Unconfigured Site')
-    print "<head><title>%s: %s</title>" % (site_name, text)
-    print ' <meta http-equiv="Content-Type" content="application/xhtml+xml; charset=utf-8" />'
+    print("<head><title>%s: %s</title>" % (site_name, text))
+    print(' <meta http-equiv="Content-Type" content="application/xhtml+xml; charset=utf-8" />')
     if not name:
-        print ' <meta name="robots" content="noindex,nofollow" />'
-    for css in css_url:
-        print ' <link rel="stylesheet" type="text/css" href="%s" />' % relative_url(css)
-    print '</head>'
+        print(' <meta name="robots" content="noindex,nofollow" />')
+
+    for meta in meta_urls:
+        http_equiv, content = meta
+        print(' <meta http-equiv="%s" content="%s" />' % (http_equiv, relative_url(content)))
+
+    for link in link_urls:
+        rel, href = link
+        print(' <link rel="%s" href="%s" />' % (rel, relative_url(href)))
+
+    if name and writable and privileged_url is not None:
+        print(' <link rel="alternate" type="application/x-wiki" title="Edit this page" href="%s" />' \
+            % (privileged_path() + '?a=edit&q=' + name))
+
+    if history_url is not None:
+        print(' <link rel="alternate" type="application/rss+xml" title="RSS" href="%s" />' \
+            % relative_url(history_url + '?a=rss'))
+
+    print('</head>')
 
     # Body
-    if name and privileged_url is not None:
-        print '<body ondblclick="location.href=\'' + privileged_path() + '?edit=' + name + '\'">'
+    if name and writable and privileged_url is not None:
+        print('<body ondblclick="location.href=\'' + privileged_path() + '?a=edit&q=' + name + '\'">')
     else:
-        print '<body>'
+        print('<body>')
 
     title_done = True
     send_guru(msg_text, msg_type)
 
     # Navbar
-    print '<div class="nav">'
+    print('<div class="nav">')
+    print link_tag('FrontPage', site_icon or 'Home', cssclass='navlink')
     if name:
-        print '  <b>' + link_tag('?fullsearch=' + name, text, 'navlink') + '</b> '
+        print('  <b>' + link_tag('?fullsearch=' + name, text, cssclass='navlink') + '</b> ')
     else:
-        print '  <b>' + text + '</b> '
-    print ' | ' + link_tag('FrontPage', 'Front Page', 'navlink')
-    print ' | ' + link_tag('FindPage', 'Find Page', 'navlink')
+        print('  <b>' + text + '</b> ')
+    print(' | ' + link_tag('FindPage', 'Find Page', cssclass='navlink'))
     if 'history_url' in globals():
-        print ' | <a href="' + history_url + '" class="navlink">Recent Changes</a>'
+        print(' | <a href="' + relative_url(history_url) + '" class="navlink">Recent Changes</a>')
         if name:
-            print ' | <a href="' + history_url + '?a=history;f=' + name + '" class="navlink">Page History</a>'
+            print(' | <a href="' + relative_url(history_url + '?a=history;f=' + name) + '" class="navlink">Page History</a>')
 
     if name:
-        print ' | ' + link_tag('?raw=' + name, 'Raw Text', 'navlink')
+        print(' | ' + link_tag(name + '?a=raw', 'Raw Text', cssclass='navlink'))
         if privileged_url is not None:
-            print ' | ' + link_tag('?edit=' + name, 'Edit Page', 'navlink', privileged=True)
+            if writable:
+                print(' | ' + link_tag('?a=edit&q=' + name, 'Edit', cssclass='navlink', privileged=True))
+            else:
+                print(' | ' + link_tag(name, 'Login', cssclass='navlink', privileged=True))
+
     else:
-        print ' | <i>Immutable Page</i>'
+        print(' | <i>Immutable Page</i>')
 
     user = remote_user()
     if user != 'AnonymousCoward':
-        print ' | <span class="login"><i>logged in as <b>' + cgi.escape(user) + '</b></i></span>'
+        print(' | <span class="login"><i><b>' + link_tag('User/' + user, user) + '</b></i></span>')
 
-    print '<hr /></div>'
+    print('<hr /></div>')
 
-def link_tag(params, text=None, ss_class=None, privileged=False):
+def send_httperror(status="403 Not Found", query=""):
+    print("Status: %s" % status)
+    send_title(None, msg_text=("%s: on query '%s'" % (status, query)))
+    send_footer()
+
+def link_tag(dest, text=None, privileged=False, **kvargs):
     if text is None:
-        text = params # default
-    classattr = ''
-    if ss_class:
-        classattr += 'class="%s" ' % ss_class
-        # Prevent crawlers from following links potentially added by spammers or to generated pages
-        if ss_class == 'external' or ss_class == 'navlink':
-            classattr += 'rel="nofollow" '
-    elif url_re.match(params):
-        classattr += 'rel="nofollow" '
-    return '<a %shref="%s">%s</a>' % (classattr, relative_url(params, privileged=privileged), text)
+        text = humanlink(dest)
+    elif image_re.match(text):
+        text = '<img border="0" src="' + relative_url(text) + '" alt="' + text + '" />'
+
+    link_class = kvargs.get('class', kvargs.get('cssclass', None))
+    if not link_class:
+        if is_external_url(dest):
+            link_class = 'external'
+        elif file_re.match(dest) and Page(dest).exists():
+            link_class = 'wikilink'
+        else:
+            text = nonexist_pfx + text
+            link_class = 'nonexistent'
+
+    # Prevent crawlers from following links potentially added by spammers or to generated pages
+    nofollow = ''
+    if link_class == 'external' or link_class == 'navlink':
+        nofollow = 'rel="nofollow" '
+
+    return '<a class="%s" %shref="%s">%s</a>' % (link_class, nofollow, relative_url(dest, privileged=privileged), text)
+
+def link_inline(name, descr=None, kvargs={}):
+    if not descr: descr = humanlink(name)
+    url = relative_url(name)
+    if video_re.match(name):
+        return '<video controls="1" src="%s">Your browser does not support the HTML5 video tag</video>' % url
+    elif image_re.match(name):
+        return '<a href="%s"><img border="0" src="%s" alt="%s" /></a>' % (url, url + url_args(kvargs), descr)
+    elif file_re.match(name) and not ext_re.search(name): # FIXME: this guesses a wiki page
+        return Page(name).send_naked()
+    else:
+        return '<iframe width="100%%" scrolling="auto" frameborder="0" src="%s"><a href="%s">%s</a></iframe>' \
+            % (url, url, name)
 
 # Search ---------------------------------------------------
 
-def handle_fullsearch(needle):
-    send_title(None, 'Full text search for "%s"' % (needle))
+def print_search_stats(hits, searched):
+    print("<p>%d hits out of %d pages searched.</p>" % (hits, searched))
+
+def handle_fullsearch(query, form):
+    needle = form['q'].value
+    send_title(None, 'Full text search for "' + needle + '"')
 
     needle_re = re.compile(needle, re.IGNORECASE)
     hits = []
@@ -195,137 +255,100 @@ def handle_fullsearch(needle):
     hits.sort()
     hits.reverse()
 
-    print "<ul>"
+    print("<ul>")
     for (count, page_name) in hits:
-        print '<li><p>' + Page(page_name).link_to()
-        print ' . . . . ' + `count`
-        print ['match', 'matches'][count != 1]
-        print '</p></li>'
-    print "</ul>"
+        print('<li><p>' + link_tag(page_name))
+        print(' . . . . ' + `count`)
+        print(['match', 'matches'][count != 1])
+        print('</p></li>')
+    print("</ul>")
 
     print_search_stats(len(hits), len(all_pages))
 
-def handle_titlesearch(needle):
-    # TODO: check needle is legal -- but probably we can just accept any RE
-    send_title(None, "Title search for \"" + needle + '"')
+def handle_titlesearch(query, form):
+    needle = form['q'].value
+    send_title(None, 'Title search for "' + needle + '"')
 
     needle_re = re.compile(needle, re.IGNORECASE)
     all_pages = page_list()
-    hits = filter(needle_re.search, all_pages)
+    hits = list(filter(needle_re.search, all_pages))
 
-    print "<ul>"
+    print("<ul>")
     for filename in hits:
-        print '<li><p>' + Page(filename).link_to() + "</p></li>"
-    print "</ul>"
+        print('<li><p>' + link_tag(filename) + "</p></li>")
+    print("</ul>")
 
     print_search_stats(len(hits), len(all_pages))
 
-def print_search_stats(hits, searched):
-    print "<p>%d hits out of %d pages searched.</p>" % (hits, searched)
+def handle_raw(pagename, form):
+    if not file_re.match(pagename):
+        send_httperror("403 Forbidden", pagename)
+        return
 
-def handle_raw(pagename):
     Page(pagename).send_raw()
 
-def handle_edit(pagename):
-    pg = Page(pagename)
+def handle_edit(pagename, form):
+    if not file_re.match(pagename):
+        send_httperror("403 Forbidden", pagename)
+        return
+
+    pg = Page(form['q'].value)
     if 'save' in form:
         if form['file'].value:
-            pg.save(form['file'].file.read())
+            pg.save(form['file'].file.read(), form['changelog'].value)
         else:
-            pg.save(form['savetext'].value.replace('\r\n', '\n'))
-        pg.send_page()
+            pg.save(form['savetext'].value.replace('\r\n', '\n'), form['changelog'].value)
+        pg.format()
     elif 'cancel' in form:
         pg.msg_text = 'Editing canceled'
         pg.msg_type = 'notice'
-        pg.send_page()
+        pg.format()
     else: # preview or edit
         text = None
         if 'preview' in form:
             text = form['savetext'].value
         pg.send_editor(text)
 
+def handle_get(pagename, form):
+        if file_re.match(pagename):
+            # FIMXE: this is all bullshit, MimeTypes bases its guess on the extension!
+            from mimetypes import MimeTypes
+            mimetype, encoding = MimeTypes().guess_type(pagename)
+            if mimetype:
+                Page(pagename).send_raw(mimetype=mimetype, args=form)
+            else:
+                Page(pagename).format()
+        else:
+            send_httperror("403 Forbidden", pagename)
+
+# Used by macros/WordIndex and macros/TitleIndex
 def make_index_key():
-    links = map(lambda ch: '<a href="#%s">%s</a>' % (ch, ch), 'abcdefghijklmnopqrstuvwxyz')
-    return '<p><center>'+ ' | '.join(links) + '</center></p>'
+    links = ['<a href="#%s">%s</a>' % (ch, ch) for ch in 'abcdefghijklmnopqrstuvwxyz']
+    return '<p style="text-align: center">' + ' | '.join(links) + '</p>'
 
-def page_list():
-    return filter(word_re.match, os.listdir(data_dir))
+def page_list(dirname=None, re=None):
+    if re is None:
+        # FIXME: WikiWord is too restrictive now!
+        re = re.compile(r"^\b((([A-Z][a-z0-9]+){2,}/)*([A-Z][a-z0-9]+){2,})\b$")
+    return sorted(filter(re.match, os.listdir(dirname or data_dir)))
 
-def send_footer(name, mod_string=None):
+def send_footer(mod_string=None):
     if globals().get('debug_cgi', False):
         cgi.print_arguments()
         cgi.print_form(form)
         cgi.print_environ()
-    print '<div id="footer"><hr />'
-    print ('<p class="copyright"><span class="benchmark">generated in %0.3fs</span>' +
-        ' by <a href="http://www.codewiz.org/wiki/GeekiGeeki">GeekiGeeki</a>' +
-        ' version %s</p>') % (clock() - start_time, __version__)
+    print('''
+<div id="footer"><hr />
+<p class="copyright">
+<a rel="license" href="http://creativecommons.org/licenses/by-sa/3.0/"><img class="license" alt="Creative Commons License" src="%s" /></a>
+<span class="benchmark">generated in %0.3fs</span> by <a href="http://www.codewiz.org/wiki/GeekiGeeki">GeekiGeeki</a> version %s
+</p>
+''' % (relative_url('cc-by-sa.png'), clock() - start_time, __version__))
     if mod_string:
-        print '<p class="modified">last modified %s</p>' % mod_string
-    print '</div></body></html>'
-
-# Macros ------------------------------------------------------------
-def _macro_TitleSearch(*vargs):
-    return _macro_search("titlesearch")
-
-def _macro_FullSearch(*vargs):
-    return _macro_search("fullsearch")
-
-def _macro_search(type):
-    default = ''
-    if 'value' in form:
-        default = form['value'].value
-    return """<form method="get"><input name="%s" size="30" value="%s" /><input type="submit" value="Search" /></form>""" % (type, default)
-
-def _macro_WordIndex(*vargs):
-    s = make_index_key()
-    pages = list(page_list())
-    map = {}
-    word_re = re.compile('[A-Z][a-z]+')
-    for name in pages:
-        for word in word_re.findall(name):
-            try:
-                map[word].append(name)
-            except KeyError:
-                map[word] = [name]
-
-    all_words = map.keys()
-    all_words.sort()
-    last_letter = None
-    # set title
-    for word in all_words:
-        letter = word[0].lower()
-        if letter != last_letter:
-            s = s + '; <a name="%s"><h3>%s</h3></a>' % (letter, letter)
-            last_letter = letter
-
-        s = s + '<b>%s</b><ul>' % word
-        links = map[word]
-        links.sort()
-        last_page = None
-        for name in links:
-            if name == last_page: continue
-            s = s + '<li>' + Page(name).link_to()
-        s = s + '</ul>'
-    return s
-
-
-def _macro_TitleIndex(*vargs):
-    s = make_index_key()
-    pages = list(page_list())
-    pages.sort()
-    current_letter = None
-    for name in pages:
-        letter = name[0].lower()
-        if letter != current_letter:
-            s += '<a name="%s"><h3>%s</h3></a>' % (letter, letter)
-            current_letter = letter
-        else:
-            s += '<br />'
-        s += Page(name).link_to()
-    return s
+        print('<p class="modified">last modified %s</p>' % mod_string)
+    print('</div></body></html>')
 
-class PageFormatter:
+class WikiFormatter:
     """Object that turns Wiki markup into HTML.
 
     All formatting commands can be parsed one line at a time, though
@@ -334,21 +357,21 @@ class PageFormatter:
     def __init__(self, raw):
         self.raw = raw
         self.h_level = 0
-        self.in_pre = self.in_table = self.in_li = False
+        self.in_pre = self.in_html = self.in_table = self.in_li = False
         self.in_header = True
         self.list_indents = []
         self.tr_cnt = 0
         self.styles = {
             #wiki   html   enabled?
             "//":  ["em",  False],
-            "''":  ["em",  False],
             "**":  ["b",   False],
-            "'''": ["b",   False],
             "##":  ["tt",  False],
-            "``":  ["tt",  False],
             "__":  ["u",   False],
             "^^":  ["sup", False],
-            ",,":  ["sub", False]
+            ",,":  ["sub", False],
+            "''":  ["em",  False], # LEGACY
+            "'''": ["b",   False], # LEGACY
+            "``":  ["tt",  False], # LEGACY
         }
 
     def _b_repl(self, word):
@@ -372,48 +395,69 @@ class PageFormatter:
     def _rule_repl(self, word):
         return self._undent() + '\n<hr size="%d" noshade="noshade" />\n' % (len(word) - 2)
 
-    def _word_repl(self, word):
-        return Page(word).link_to()
-
-    def _img_repl(self, word):
-        path = relative_url(word)
-        return '<a href="%s"><img border="0" src="%s" /></a>' % (path, path)
-
-    def _url_repl(self, word):
-        if img_re.match(word):
-            return '<a href="%s"><img border="0" src="%s" /></a>' % (word, word)
-        else:
-            return '<a href="%s" rel="nofollow" class="external">%s</a>' % (word, word)
+    def _macro_repl(self, word):
+        try:
+            args, kwargs = parse_args(word)
+            macro = globals().get('_macro_' + args[0])
+            if not macro:
+                exec(open("macros/" + name + ".py").read(), globals())
+                macro = globals().get('_macro_' + name)
+            return macro(*args, **kwargs)
+        except Exception:
+            msg = cgi.escape(word)
+            if not self.in_html:
+                msg = '<strong class="error">' + msg + '</strong>'
+            return msg
 
     def _hurl_repl(self, word):
-        m = re.compile("\[\[([^ \t\n\r\f\v\|]+)(?:\s*\|\s*([^\]]+)|)\]\]").match(word)
-        name = m.group(1)
-        descr = m.group(2) or name
-
-        macro = globals().get('_macro_' + name)
-        if macro:
-            return macro(name, descr)
-        elif img_re.match(name):
-            name = relative_url(name)
-            # The "extthumb" nonsense works around a limitation of the HTML block model
-            return '<div class="extthumb"><div class="thumb"><a href="%s"><img border="0" src="%s" alt="%s" /></a><div class="caption">%s</div></div></div>' % (name, name, descr, descr)
+        args, kvargs = parse_args(word)
+        return link_tag(*args, **kvargs)
+
+    def _inl_repl(self, word):
+        args, kvargs = parse_args(word)
+        name = args.pop(0)
+        if len(args):
+            descr = args.pop(0)
+            # This double div nonsense works around a limitation of the HTML block model
+            return '<div class="' + kvargs.get('class', 'thumb') + '">' \
+                + '<div class="innerthumb">' \
+                + link_inline(name, descr, kvargs) \
+                + '<div class="caption">' + descr + '</div></div></div>'
         else:
-            if img_re.match(descr):
-                descr = '<img border="0" src="' + descr + '" />'
-
-            return link_tag(name, descr, 'wikilink')
-
-    def _email_repl(self, word):
-        return '<a href="mailto:%s">%s</a>' % (word, word)
+            return link_inline(name, None, kvargs)
 
     def _html_repl(self, word):
+        if not self.in_html and word.startswith('<div'): word = '</p>' + word
+        self.in_html += 1
+        return word; # Pass through
+
+    def _htmle_repl(self, word):
+        self.in_html -= 1
+        if not self.in_html and word.startswith('</div'): word += '<p>'
         return word; # Pass through
 
     def _ent_repl(self, s):
+        if self.in_html:
+            return s; # Pass through
         return {'&': '&amp;',
                 '<': '&lt;',
                 '>': '&gt;'}[s]
 
+    def _img_repl(self, word): # LEGACY
+        return self._inl_repl('{{' + word + '}}')
+
+    def _word_repl(self, word): # LEGACY
+        if self.in_html: return word # pass through
+        return link_tag(word)
+
+    def _url_repl(self, word): # LEGACY
+        if self.in_html: return word # pass through
+        return link_tag(word)
+
+    def _email_repl(self, word): # LEGACY
+        if self.in_html: return word # pass through
+        return '<a href="mailto:%s">%s</a>' % (word, word)
+
     def _li_repl(self, match):
         if self.in_li:
             return '</li><li>'
@@ -480,79 +524,86 @@ class PageFormatter:
         return res
 
     def replace(self, match):
-        for type, hit in match.groupdict().items():
+        for rule, hit in list(match.groupdict().items()):
             if hit:
-                return getattr(self, '_' + type + '_repl')(hit)
+                return getattr(self, '_' + rule + '_repl')(hit)
         else:
-            raise "Can't handle match " + `match`
+            raise "Can't handle match " + repr(match)
 
     def print_html(self):
-        print '<div class="wiki"><p>'
-
-        # For each line, we scan through looking for magic
-        # strings, outputting verbatim any intervening text
-        # TODO: highlight search words (look at referrer)
-        scan_re = re.compile(
-            r"(?:"
-            # Formatting
-            + r"(?P<b>\*\*|'''|//|''|##|``|__|\^\^|,,)"
-            + r"|(?P<tit>\={2,6})"
-            + r"|(?P<br>\\\\)"
-            + r"|(?P<rule>^-{3,})"
-            + r"|(?P<html><(/|)(div|span|iframe)[^<>]*>)"
-            + r"|(?P<ent>[<>&])"
-            + r"|(?P<hi>\b(FIXME|TODO|DONE)\b)"
+        print('<div class="wiki"><p>')
+
+        scan_re = re.compile(r"""(?:
+            # Styles and formatting
+              (?P<b>     \*\*|'''|//|''|\#\#|``|__|\^\^|,,)
+            | (?P<tit>   \={2,6})
+            | (?P<br>    \\\\)
+            | (?P<rule>  ^-{3,})
+            | (?P<hi>    \b( FIXME | TODO | DONE )\b )
 
             # Links
-            + r"|(?P<img>\b[a-zA-Z0-9_-]+\.(png|gif|jpg|jpeg|bmp))"
-            + r"|(?P<word>\b(?:[A-Z][a-z]+){2,}\b)"
-            + r"|(?P<hurl>\[\[([^ \t\n\r\f\v\|]+)(?:\s*\|\s*([^\]]+)|)\]\])"
-            + r"|(?P<url>(http|https|ftp|mailto)\:[^\s'\"]+\S)"
-            + r"|(?P<email>[-\w._+]+\@[\w.-]+)"
+            | (?P<macro> \<\<([^\s\|\>]+)(?:\s*\|\s*([^\>]+)|)\>\>)
+            | (?P<hurl>  \[\[([^\s\|]+)(?:\s*\|\s*([^\]]+)|)\]\])
+
+            # Inline HTML
+            | (?P<html>  <(br|hr|div|span|form|iframe|input|textarea|a|img|h[1-5])\b )
+            | (?P<htmle> ( /\s*> | </(br|hr|div|span|form|iframe|input|textarea|a|img|h[1-5])> ) )
+            | (?P<ent>   [<>&] )
+
+            # Auto links (LEGACY)
+            | (?P<img>   \b[a-zA-Z0-9_/-]+\.(""" + image_ext + "|" + video_ext + r"""))
+            | (?P<word>  \b(?:[A-Z][a-z]+){2,}\b)
+            | (?P<url>   (http|https|ftp|mailto)\:[^\s'\"]+\S)
+            | (?P<email> [-\w._+]+\@[\w.-]+)
 
             # Lists, divs, spans
-            + r"|(?P<li>^\s+[\*#] +)"
-            + r"|(?P<pre>\{\{\{|\s*\}\}\})"
+            | (?P<li>    ^\s+[\*\#]\s+)
+            | (?P<pre>   \{\{\{|\s*\}\}\})
+            | (?P<inl>   \{\{([^\s\|]+)(?:\s*\|\s*([^\]]+)|)\}\})
 
             # Tables
-            + r"|(?P<tr>^\s*\|\|(=|)\s*)"
-            + r"|(?P<tre>\s*\|\|(=|)\s*$)"
-            + r"|(?P<td>\s*\|\|(=|)\s*)"
-            + r")")
-        pre_re = re.compile(
-            r"(?:"
-            + r"(?P<pre>\s*\}\}\})"
-            + r"|(?P<ent>[<>&])"
-            + r")")
+            | (?P<tr>    ^\s*\|\|(=|)\s*)
+            | (?P<tre>   \s*\|\|(=|)\s*$)
+            | (?P<td>    \s*\|\|(=|)\s*)
+
+            # TODO: highlight search words (look at referrer)
+          )""", re.VERBOSE)
+        pre_re = re.compile("""(?:
+              (?P<pre>\s*\}\}\})
+            | (?P<ent>[<>&])"
+            )""", re.VERBOSE)
         blank_re = re.compile(r"^\s*$")
         indent_re = re.compile(r"^\s*")
         tr_re = re.compile(r"^\s*\|\|")
         eol_re = re.compile(r"\r?\n")
-        for self.line in eol_re.split(self.raw.expandtabs()):
-            # Skip ACLs
+        # For each line, we scan through looking for magic strings, outputting verbatim any intervening text
+        #3.0: for self.line in eol_re.split(str(self.raw.expandtabs(), 'utf-8')):
+        for self.line in eol_re.split(str(self.raw.expandtabs())):
+            # Skip pragmas
             if self.in_header:
                 if self.line.startswith('#'):
                     continue
                 self.in_header = False
 
             if self.in_pre:
-                print re.sub(pre_re, self.replace, self.line)
+                print(re.sub(pre_re, self.replace, self.line))
             else:
                 if self.in_table and not tr_re.match(self.line):
                     self.in_table = False
-                    print '</tbody></table><p>'
+                    print('</tbody></table><p>')
 
                 if blank_re.match(self.line):
-                    print '</p><p>'
+                    print('</p><p>')
                 else:
                     indent = indent_re.match(self.line)
-                    print self._indent_to(len(indent.group(0)))
-                    print re.sub(scan_re, self.replace, self.line)
+                    #3.0: print(self._indent_to(len(indent.group(0))), end=' ')
+                    print(self._indent_to(len(indent.group(0))))
+                    print(re.sub(scan_re, self.replace, self.line))
 
-        if self.in_pre: print '</pre>'
-        if self.in_table: print '</tbody></table><p>'
-        print self._undent()
-        print '</p></div>'
+        if self.in_pre: print('</pre>')
+        if self.in_table: print('</tbody></table><p>')
+        print(self._undent())
+        print('</p></div>')
 
 class Page:
     def __init__(self, page_name):
@@ -564,70 +615,81 @@ class Page:
         # look for the end of words and the start of a new word and insert a space there
         return re.sub('([a-z])([A-Z])', r'\1 \2', self.page_name)
 
-    def _text_filename(self):
-        return path.join(data_dir, self.page_name)
+    def _filename(self):
+        return os.path.join(data_dir, self.page_name)
 
     def _tmp_filename(self):
-        return path.join(data_dir, ('#' + self.page_name.replace('/','_') + '.' + `os.getpid()` + '#'))
+        return os.path.join(data_dir, ('#' + self.page_name.replace('/','_') + '.' + str(os.getpid()) + '#'))
 
     def exists(self):
         try:
-            os.stat(self._text_filename())
+            os.stat(self._filename())
             return True
-        except OSError, er:
-            if er.errno == errno.ENOENT:
+        except OSError, err:
+            if err.errno == errno.ENOENT:
                 return False
-            raise er
+            raise err
 
-    def link_to(self):
-        word = self.page_name
-        if self.exists():
-            return link_tag(word, word, 'wikilink')
-        else:
-            return link_tag(word, nonexist_pfx + word, 'nonexistent')
-
-    def get_raw_body(self):
+    def get_raw_body(self, default=None):
         try:
-            return open(self._text_filename(), 'rb').read()
-        except IOError, er:
-            if er.errno == errno.ENOENT:
-                return '' # just doesn't exist, use default
-            if er.errno == errno.EISDIR:
-                return 'DIR'
-            raise er
-
-    def get_attrs(self):
-        if 'attrs' in self.__dict__:
-            return self.attrs
-        self.attrs = {}
-        try:
-            file = open(self._text_filename(), 'rt')
-            attr_re = re.compile(r"^#(\S*)(.*)$")
-            for line in file:
-                m = attr_re.match(line)
-                if not m:
-                    break
-                self.attrs[m.group(1)] = m.group(2).strip()
-                #print "bernie: attrs[" + m.group(1) + "] = " + m.group(2) + "<br>\n"
-        except IOError, er:
-            if er.errno != errno.ENOENT:
-                raise er
-        return self.attrs
-
-    def get_attr(self, name, default):
-        return self.get_attrs().get(name, default)
+            return open(self._filename(), 'rb').read()
+        except IOError, err:
+            if err.errno == errno.ENOENT:
+                if default is None:
+                    default = '//[[?a=edit&q=%s|Describe %s]]//' % (self.page_name, self.page_name)
+                return default
+            if err.errno == errno.EISDIR:
+                return self.format_dir()
+            raise err
+
+    def format_dir(self):
+        out = '== '
+        pathname = ''
+        for dirname in self.page_name.split('/'):
+            pathname = (pathname + '/' + dirname) if pathname else dirname
+            out += '[[' + pathname + '|' + dirname + ']]/'
+        out += ' ==\n'
+        for filename in page_list(self._filename(), file_re):
+            if image_re.match(filename):
+                if image_maxwidth:
+                    maxwidth_arg = ' | maxwidth=' + str(image_maxwidth)
+                out += '{{' + self.page_name + '/' + filename + ' | ' + humanlink(filename) + maxwidth_arg + ' | class=thumbleft}}\n'
+            else:
+                out += ' * [[' + self.page_name + '/' + filename + ']]\n'
+        return out
+
+    def pragmas(self):
+        if not '_pragmas' in self.__dict__:
+            self._pragmas = {}
+            try:
+                file = open(self._filename(), 'rt')
+                attr_re = re.compile(r"^#(\S*)(.*)$")
+                for line in file:
+                    m = attr_re.match(line)
+                    if not m:
+                        break
+                    self._pragmas[m.group(1)] = m.group(2).strip()
+                    #print "bernie: pragmas[" + m.group(1) + "] = " + m.group(2) + "<br>\n"
+            except IOError, err:
+                if err.errno != errno.ENOENT and err.errno != errno.EISDIR:
+                    raise er
+        return self._pragmas
+
+    def pragma(self, name, default):
+        return self.pragmas().get(name, default)
 
     def can(self, action, default=True):
         acl = None
         try:
             #acl SomeUser:read,write All:read
-            acl = self.get_attr("acl", None)
+            acl = self.pragma("acl", None)
             for rule in acl.split():
                 (user, perms) = rule.split(':')
                 if user == remote_user() or user == "All":
                     return action in perms.split(',')
             return False
-        except Exception, er:
+        except Exception:
             if acl:
                 self.msg_text = 'Illegal acl line: ' + acl
         return default
@@ -638,29 +700,31 @@ class Page:
     def can_read(self):
         return self.can("read", True)
 
-    def send_page(self):
-        page_name = None
-        if self.can_write():
-            page_name = self.page_name
-
-        #css foo.css bar.css
-        global css_url
-        css_url = self.get_attr("css", "").split() + css_url
-
-        send_title(page_name, self.split_title(), msg_text=self.msg_text, msg_type=self.msg_type)
+    def send_naked(self):
         if self.can_read():
-            PageFormatter(self.get_raw_body()).print_html()
+            WikiFormatter(self.get_raw_body()).print_html()
         else:
             send_guru("Read access denied by ACLs", "notice")
-        send_footer(page_name, self._last_modified())
+
+    def format(self):
+        #css foo.css
+        value = self.pragma("css", None)
+        if value:
+            global link_urls
+            link_urls += [ [ "stylesheet", value ] ]
+
+        send_title(self.page_name, self.split_title(),
+            msg_text=self.msg_text, msg_type=self.msg_type, writable=self.can_write())
+        self.send_naked()
+        send_footer(self._last_modified())
 
     def _last_modified(self):
         try:
             from time import localtime, strftime
-            modtime = localtime(os.stat(self._text_filename())[stat.ST_MTIME])
-        except OSError, er:
-            if er.errno != errno.ENOENT:
-                raise er
+            modtime = localtime(os.stat(self._filename())[stat.ST_MTIME])
+        except OSError, err:
+            if err.errno != errno.ENOENT:
+                raise err
             return None
         return strftime(datetime_fmt, modtime)
 
@@ -670,54 +734,70 @@ class Page:
             send_guru("Write access denied by ACLs", "error")
             return
 
-        file = ''
+        filename = ''
         if 'file' in form:
-            file = form['file'].value
+            filename = form['file'].value
 
-        print ('<p><b>Editing ' + self.page_name
+        print(('<p><b>Editing ' + self.page_name
             + ' for ' + cgi.escape(remote_user())
             + ' from ' + cgi.escape(get_hostname(remote_host()))
-            + '</b></p>')
-        print '<div class="editor"><form method="post" enctype="multipart/form-data" action="%s">' % relative_url(self.page_name)
-        print '<input type="hidden" name="edit" value="%s">' % (self.page_name)
-        print '<textarea wrap="off" spellcheck="true" id="editor" name="savetext" rows="17" cols="100">%s</textarea>' % (preview or self.get_raw_body())
-        print 'Or upload a file: <input type="file" name="file" value="%s" />' % file
-        print """
+            + '</b></p>'))
+        print('<div class="editor"><form name="editform" method="post" enctype="multipart/form-data" action="%s">' % relative_url(self.page_name))
+        print('<input type="hidden" name="a" value="edit" /><input type="hidden" name="q" value="' + self.page_name + '" />')
+        print('<input type="input" id="editor" name="changelog" value="Edit page %s" accesskey="c" /><br />' % (self.page_name))
+        print('<textarea wrap="off" spellcheck="true" id="editor" name="savetext" rows="17" cols="100" accesskey="e">%s</textarea>' \
+            % cgi.escape(preview or self.get_raw_body(default='')))
+        print('<label for="file" accesskey="u">Or Upload a file:</label> <input type="file" name="file" value="%s" />' % filename)
+        print("""
             <br />
-            <input type="submit" name="save" value="Save" />
-            <input type="submit" name="preview" value="Preview" />
+            <input type="submit" name="save" value="Save" accesskey="s">
+            <input type="submit" name="preview" value="Preview" accesskey="p" />
             <input type="reset" value="Reset" />
             <input type="submit" name="cancel" value="Cancel" />
             <br />
-            </form></div>"""
-        print "<p>" + Page('EditingTips').link_to() + "</p>"
+            </form></div>
+            <script language="javascript">
+            <!--
+            document.editform.savetext.focus()
+            //-->
+            </script>
+            """)
+        print("<p>" + link_tag('EditingTips') + "</p>")
         if preview:
-            print "<div class='preview'>"
-            PageFormatter(preview).print_html()
-            print "</div>"
-        send_footer(self.page_name)
+            print("<div class='preview'>")
+            WikiFormatter(preview).print_html()
+            print("</div>")
+        send_footer()
 
-    def send_raw(self, mimetype='text/plain'):
-        if self.can_read():
-            body = self.get_raw_body()
+    def send_raw(self, mimetype='text/plain', args=[]):
+        if not self.can_read():
+            send_title(None, msg_text='Read access denied by ACLs', msg_type='notice')
+            return
+
+        if 'maxwidth' in args:
+            import subprocess
             emit_header(mimetype)
-            print body
+            sys.stdout.flush()
+            subprocess.check_call(['gm', 'convert', self._filename(),
+                '-scale', args['maxwidth'].value + ' >', '-'])
         else:
-            send_title(None, msg_text='Read access denied by ACLs', msg_type='notice')
+            body = self.get_raw_body()
+            emit_header(mimetype)
+            print(body)
 
     def _write_file(self, data):
         tmp_filename = self._tmp_filename()
         open(tmp_filename, 'wb').write(data)
-        name = self._text_filename()
+        name = self._filename()
         if os.name == 'nt':
             # Bad Bill!  POSIX rename ought to replace. :-(
             try:
                 os.remove(name)
-            except OSError, er:
-                if er.errno != errno.ENOENT: raise er
+            except OSError, err:
+                if err.errno != errno.ENOENT: raise err
         os.rename(tmp_filename, name)
 
-    def save(self, newdata):
+    def save(self, newdata, changelog):
         if not self.can_write():
             self.msg_text = 'Write access denied by ACLs'
             self.msg_type = 'error'
@@ -726,53 +806,29 @@ class Page:
         self._write_file(newdata)
         rc = 0
         if post_edit_hook:
-            # FIXME: what's the std way to perform shell quoting in python?
-            cmd = ( post_edit_hook
-                + " '" + data_dir + '/' + self.page_name
-                + "' '" + remote_user()
-                + "' '" + remote_host() + "'"
-            )
-            out = os.popen(cmd)
-            output = out.read()
-            rc = out.close()
+            import subprocess
+            cmd = [ post_edit_hook, data_dir + '/' + self.page_name, remote_user(), remote_host(), changelog]
+            child = subprocess.Popen(cmd, stdout=subprocess.PIPE, close_fds=True)
+            output = child.stdout.read()
+            rc = child.wait()
         if rc:
-            self.msg_text += "Post-editing hook returned %d.\n" % rc
-            self.msg_text += 'Command was: ' + cmd + '\n'
+            self.msg_text += "Post-editing hook returned %d. Command was:\n'%s'\n" % (rc, "' '".join(cmd))
             if output:
                 self.msg_text += 'Output follows:\n' + output
         else:
             self.msg_text = 'Thank you for your contribution.  Your attention to detail is appreciated.'
             self.msg_type = 'success'
 
-# Main ---------------------------------------------------------------
 try:
-    execfile("geekigeeki.conf.py")
+    exec(open("geekigeeki.conf.py").read())
     form = cgi.FieldStorage()
-
-    for cmd in form:
-        handler = globals().get('handle_' + cmd)
-        if handler:
-            handler(form[cmd].value)
-            break
+    action = form.getvalue('a', 'get')
+    handler = globals().get('handle_' + action)
+    if handler:
+        handler(query_string(), form)
     else:
-        path_info = environ.get('PATH_INFO', '')
-        if len(path_info) and path_info[0] == '/':
-            query = path_info[1:] or 'FrontPage'
-        else:
-            query = environ.get('QUERY_STRING', '') or 'FrontPage'
+        send_httperror("403 Forbidden", query_string())
 
-        if file_re.match(query):
-            if word_re.match(query):
-                Page(query).send_page()
-            else:
-                from mimetypes import MimeTypes
-                type, encoding = MimeTypes().guess_type(query)
-                type = type or 'text/plain'
-                Page(query).send_raw(mimetype=type)
-        else:
-            print "Status: 404 Not Found"
-            send_title(None, msg_text='Can\'t work out query: ' + query)
-            send_footer(None)
 except Exception:
     import traceback
     msg_text = traceback.format_exc()
@@ -780,6 +836,6 @@ except Exception:
         send_guru(msg_text, "error")
     else:
         send_title(None, msg_text=msg_text)
-    send_footer(None)
+    send_footer()
 
 sys.stdout.flush()