Re: [Python-markdown-discuss] Not liking Mako-like syntax

[Yuri T. <yu...@cs...>]

On Dec 31, 2007 10:49 AM, Waylan Limberg <wa...@gm...> wrote:
> Well,  the proper way would be to write your own pre-processor that
> strips out all the mako markup and adds it to the htmlstash. That is,
> after all, how html is currently ignored (in HtmlBlockPreprocessor -
> line 486).
>
> With only a quick look, my guess is that the problem stems from the
> fact that python-markdown has a very limited understanding of xml.
> Therefore, the `%`s are throwing it off. In fact, if you try your
> given test case without the `%` it works fine:

First, for the record, <%foo /> doesn't count as XML.  Second, yes,
the _right_ way to do this is to write a custom preprocessor.

However, my intension was to support this, so I suppose this is a bug.
 Let me explain how it is _supposed_ to work:

When the script encounters an HTML tag in the beginning of the block,
it classifies the tag into three types:

1. Tags that open an HTML block, such as <table> or <pre>.  In this
case, there will be no markdown processing until the block is closed
with the same tag.  There will also be no "<p>..</p>" around this
chunk of html.
2. Tags like <div> or <?php>.  In this case we don't look for the
closing tag, but we also avoid putting <p>...</p> around the block.
3. Inline tags.  We don't look for where the tag closes and we put
<p>...</p> around the block.

#2 is supposed to include "div", as well as any tags starting with
"%", "?".  The rationale is that for something like <?php> or <%foo>
we don't really know what that tag means and we don't want to make any
assumption as to whether it needs to be closed _or_ whether it needs
<p> around it.  If you want "<p>" around your <%foo/>, you can put it
manually, after all, but you won't be able to remove it easily if we
add it.

So, "%foo" and "foo" are supposed to be treated differently.
"<foo>**bar**" should become "<p><foo><b>bar</b><p>" while
"<%foo>**bar**" is supposed to become as "<%foo><b>bar</b>"

It obviously doesn't work that way.  I think that the bug is either
_get_right_tag() or _equal_tag().  The quick fix is to change line 499
to

        if left_tag in ['div'] or left_tag[0] in ["?", "@", "%"]: #
handle PHP, etc.

Though, at some point, the whole HTML_Block_Preprocessor logic needs
to be rethought.  It seems more convoluted than it needs to be.

I hope this helps.

- yuri

-- 
http://sputnik.freewisdom.org/