Markdown Discuss - Mar 2009 - Paragraph tags around <div> blocks

For those of you who are expert you may chuckle over this.
Perhaps if it's in the record, however, it may help someone else.


I was  getting cases of

<div id="content">

standing on a line by itself, getting converted to this:

<p><div id="content"></p>


Yes, the line above and below are blank.
Yes the div tag is starting at the first column

It displays as intended.  Firefox's browser is clever, but it won't
validate, and given all the whinging
I've done about non-validating web pages, I better get mine right.


I'm processing with template toolkit.  If I turn off the multimarkdown
filter, this is what's being passed:
     32         <div class="sidepix">
     33                 <img  src="/Images/sidepix/rotate.php"
     34                  alt = "Random scenic picture" />
     35         </div>
     36 </div>
     37 <!-- End of SideNav -->
     38
     39
     40
     41 <div id="content">
     42
     43 # Contact Info
     44    If you want to write, phone, email, drive here this is the place to
     45 find out how.
     46

     32    <div class="sidepix">
     33                 <img  src="/Images/sidepix/rotate.php"
     34                  alt = "Random scenic picture" />
     35         </div>
     36 </div>
     37 <!-- End of SideNav -->
     38
     39 <p><div id="content"></p>
     40
     41 <h1 id="contactinfo">Contact Info</h1>
     42
     43 <p>If you want to write, phone, email, drive here this is the
place to
     44 find out how.  </p>
     45

As a sanity check I reverted to Markdown instead of multimarkdown

No change.

Now it wasn't happening in all files

An hour of pruning copies of files that worked versus ones that didn't
found the following:

This file would work:

[%  USE MARKDOWN %]
[%  FILTER MARKDOWN %]

<div id=foo>
</div>

[% END  #filter markdown %]


This file would not work:
[%  USE MARKDOWN %]
[%  FILTER MARKDOWN %]

<div id=foo>


[% END  #filter markdown %]
</div>

The [% boxes are template toolkit instructions.
So TT was NOT handling markdown the closing </div> in the non-working
cases.
Markdown couldn't find a closing tag to match the opening tag,
therefore it must be text,
and gets wrapped in <p> tags.

Go figure.

-- 
Sherwood Botsford
Sherwood's Forests
Warburg, Alberta T0C 2T0
http://www.sherwoods-forests.com
780-848-2548

I can now reproduce my problem very simply.

[% USE MultiMarkdown %]
[% FILTER multimarkdown %]

# Headline outside div

<div id="content">

# Headline inside div

</div>

[% END %]

Produces:

<h1 id="headlineoutsidediv">Headline outside div</h1>

<div id="content">

# Headline inside div


</div>


-- 
Sherwood Botsford
Sherwood's Forests
Warburg, Alberta T0C 2T0
http://www.sherwoods-forests.com
780-848-2548

It's a monolog week for me.  Answering my own post. Again.

Turns out it's not that simple.  Consider the following input fragment:
(For clarity, I'm compressing the double spacing.  All <div> tags have
a blank line before and after.

# Headline outside div
<div>
# Headline inside div
</div>

Run through MMD I get:
<h1 id="headlineoutsidediv">Headline outside div</h1>
<div>
# Headline inside div
</div>

IF I use DIV instead of div
<h1 id="headlineoutsidediv">Headline outside div</h1>

<p><DIV></p>

<h1 id="headlineinsidediv">Headline inside div</h1>

<p></DIV></p>

Which causes the w3c validator the whine.
I get this form of result if I have <div id=foobar>
but the first form of result if I have <div id="foobar">


This behaviour is identical for both MMD 1.0.24 (CPAN) and for MMD
2.0.b5 (FP's website)

I think that the regex that's scaning for a div tag is looking only
for a well formed one.
Being either wrong case or with a malformed id it's treated as text,
and surrounded by <p> tags.

If the divs match the expression it doesn't process the content.

-- 
Sherwood Botsford
Sherwood's Forests
Warburg, Alberta T0C 2T0
http://www.sherwoods-forests.com
780-848-2548