Attachment 302170 Details for Bug 441917 – go through all the steps to undo gmail's weirdness

[patch] go through all the steps to undo gmail's weirdness

gmail-unscrewage.patch (text/plain), 5.29 KB, created by Colin Walters on 2008-04-11 20:57:38 UTC

(hide)

Description:

Filename:

MIME Type:

Creator: Colin Walters

Created: 2008-04-11 20:57:38 UTC

Size: 5.29 KB

patch

obsolete

>Index: bigboard/stocks/mail/MailStock.py
>===================================================================
>--- bigboard/stocks/mail/MailStock.py	(revision 7291)
>+++ bigboard/stocks/mail/MailStock.py	(working copy)
>@@ -1,8 +1,13 @@
>-import logging, re, htmlentitydefs, time, urllib2
>+import logging, re, time, urllib2
> 
> import gobject, gtk
> import hippo
>+import xml.dom.minidom
>+from xml.dom.minidom import Node
>+from StringIO import StringIO
> 
>+import simplejson
>+
> from bigboard.stock import Stock
> from bigboard.slideout import ThemedSlideout
> import bigboard.google as google
>@@ -16,48 +21,39 @@
> 
> _logger = logging.getLogger('bigboard.stocks.MailStock')
> 
>-def remove_strange_tags(s, markup=False):
>- if "\\u003cb\\>" in s:
>- if markup == True:
>- b = ""
>- e = ""
>+def gmail_jshtml_str_parse(s, markup=False):
>+ # Convert the current data back into an ECMAScript string
>+ jsstr = '"' + s + '"'
>+ # These escape sequences are not legal ECMAScript; strip them
>+ jsstr = jsstr.replace(r'\>', '>')
>+ jsstr = jsstr.replace(r'\<', '<')
>+ # Use an actual ECMAScript parser to turn it safely back into
>+ # a Python string.
>+ parsed_jsstr = simplejson.loads(jsstr)
>+ # At this point, we have a Python unicode string which *should* hold
>+ # an XML fragment. Convert that fragment into a document string.
>+ pystr = "<html>" + parsed_jsstr + "</html>"
>+ # Parse that document string into a DOM. 
>+ dom = xml.dom.minidom.parseString(pystr)
>+ textContent = StringIO()
>+ # Now we parse the XML, only allowing the bold tag through, and eating everything else
>+ def DomToText(node): 
>+ if node.nodeType == Node.TEXT_NODE:
>+ textContent.write(node.data)
>+ if markup and node.nodeType == Node.ELEMENT_NODE and node.nodeName == 'b':
>+ in_bold = True
>+ textContent.write('')
> else:
>- b = ""
>- e = ""
>- s = s.replace("\\u003cb\\>", b)
>- s = s.replace("\\u003c/b\\>", e)
>- return s
>+ in_bold = False 
>+ if node.hasChildNodes():
>+ for child in node.childNodes:
>+ DomToText(child)
>+ if in_bold:
>+ textContent.write('')
>+ DomToText(dom.documentElement)
>+ # Return the sanely filtered content
>+ return textContent.getvalue()
> 
>-_CONVERT_ENTITIES_RE = re.compile("&(?:(#[0-9]+)|(#x[0-9A-Fa-f]+)|([A-Za-z]+));")
>-
>-def _convert_entity(m):
>- try:
>- if m.group(1) is not None:
>- return unichr(int(m.group(1)[1:]))
>- elif m.group(2) is not None:
>- return unichr(int(m.group(2)[2:], 16))
>- else:
>- return unichr(htmlentitydefs.name2codepoint[m.group(3)])
>- except ValueError:
>- return m.group(0)
>- except KeyError:
>- return m.group(0)
>- except OverflowError:
>- return m.group(0)
>-
>-def convert_entities(s):
>- """Replace standard HTML entities and numeric character references in the string"""
>- return _CONVERT_ENTITIES_RE.sub(_convert_entity, s) 
>-
>-# assert convert_entities("&amp;") == "&"
>-# assert convert_entities("&amp;foo&lt;") == "&foo<"
>-# assert convert_entities("&#x41;") == "A"
>-# assert convert_entities("&#65;") == "A"
>-# assert convert_entities("&zzz_amp;") == "&zzz_amp;" # not something we parse as an entity
>-# assert convert_entities("&zzzamp;") == "&zzzamp;" # unknown entity
>-# assert convert_entities("&#x1000000;") == "&#x1000000;" # not a unicode character
>-# assert convert_entities("&#x1000000000;") == "&#x1000000000;" # overflow
>-
> class LabelSlideout(ThemedSlideout):
> __gsignals__ = {
> 'changed' : (gobject.SIGNAL_RUN_LAST, gobject.TYPE_NONE, (gobject.TYPE_STRING, )),
>@@ -94,7 +90,7 @@
> self.__header = Header(topborder=False)
> self.id = thread.id 
> 
>- subject = remove_strange_tags(thread.subject)
>+ subject = gmail_jshtml_str_parse(thread.subject)
> 
> subject_box = hippo.CanvasText(classes='header', text=subject)
> self.__header.append(subject_box, hippo.PACK_EXPAND)
>@@ -106,9 +102,8 @@
> if type(value) is list:
> s = ", ".join(value)
> if type(value) is str:
>- s = remove_strange_tags(value)
>- 
>- s = convert_entities(s)
>+ s = gmail_jshtml_str_parse(value)
>+
> box = hippo.CanvasText(text=s, xalign=hippo.ALIGNMENT_START)
> vbox.append(box)
> 
>@@ -189,7 +184,7 @@
> for thread in threads:
> if i >= self.__display_limit: break
> 
>- subject = remove_strange_tags(thread.subject, True)
>+ subject = gmail_jshtml_str_parse(thread.subject, True)
> 
> box = PrelightingCanvasBox()
> box.connect("button-press-event", self.create_email_slideout, thread)
>@@ -246,3 +241,10 @@
> 
> def __on_more_button(self):
> libbig.show_url("http://mail.google.com/mail")
>+
>+if __name__ == '__main__':
>+ # We want to keep bold tags
>+ assert gmail_jshtml_str_parse(r'test \u003cb\>hi\u003c/b\> moo', True) == 'test hi moo'
>+ # Strip unknown tag "A"
>+ assert gmail_jshtml_str_parse(r'test \u003ca\>hi\u003c/a\> moo', True) == 'test hi moo'
>+ 
>\ No newline at end of file

Actions: View | Diff

Attachments on bug 441917: 302170