Re: [Python-markdown-discuss] HTML Encoding Issues within Output of Custom Extension

[Waylan L. <wa...@gm...>]

On Wed, Jan 18, 2012 at 11:43 AM, andrew thornton
<ath...@gm...> wrote:
> Hello,
>
> I am kind of new to python, but I needed the ability to insert a
> gallery into my django project. I have been working on creating a
> python markdown extension that will insert an image gallery within my
> django project when a custom tag is used. The actual extension is
> working, but the HTML that the extension returns is all encoded. Here
> is the extension that I am using:
>
[snip]
>    class Galleria(markdown.inlinepatterns.Pattern):
>        def handleMatch(self, m):
>            try:
>                images =
> SuperPage.objects.get(id=m.group('superpage_id')).superpageimage_set.all()
>            except:
>                images = None
>            if images:
>                rendered = render_to_string('galleria.html', {
> 'images': images })
>            else:
>                rendered = '<b>There are no images for the given ID</b>'
>            return rendered

Your problem is in the last few lines above. Inlinepatterns must
return ElementTree instances, not strings. Therefore, you need to
build a ET Element (or tree of elements) and return that. For example
consider this abbreviated rewrite of your code:

    if images:
        # build and return gallery here
    else:
        b = markdown.util.etree.Element('b')
        b.text = 'There are no images for the given ID'
        return b

Of course, this means no Django templates for your gallery without a
few workarounds.

A few approaches you might want to look at:

1) Follow the method used by the CodeHilite Extension and save the
html output in the htmlStash (marked as 'safe' so is doesn't get eat
by safe_mode) and return the placeholder. Then the placeholder will be
automatically replaced with the html after the Tree is searialized
into a string. See the code here [1].

2) Parse the html into a ET Tree and return it. But be careful here,
most parsers (including ElementTree's) fall flat on there face with
the simplest of html syntax errors. I would suggest lxml or html5lib,
both of which adds an extra dependency (lxml is the best IMO, but
requires a C lib).  See this excellent comparison [2] of the options.
Just make sure the one you use will build an ET Tree.

[1]: https://github.com/waylan/Python-Markdown/blob/master/markdown/extensions/codehilite.py#L184
[2]: http://blog.ianbicking.org/2008/03/30/python-html-parser-performance/

-- 
----
\X/ /-\ `/ |_ /-\ |\|
Waylan Limberg