HTML is sometimes interpreted after unescaping entities

[kentonv]

The current trunk code appears to have regressed from 0.6.1 when given this input:

echo '<pre><span class="cp">#inclued &lt;foo.h&gt;</span></pre>' | maruku

It seems to un-escape the HTML entities, and then go back and try to interpret the resulting text as HTML:

 ___________________________________________________________________________
| Maruku tells you:
+---------------------------------------------------------------------------
| Malformed HTML starting at "<foo.h>"
| ---------------------------------------------------------------------------
| #inclued <foo.h>EOF
| ---------|------------------------------------------------------------------
|          +--- Byte 9
| Shown bytes [0 to 16] of 16:
| >#inclued <foo.h>
| 
| Elements read in span: 
|  -
| Current string: 
|   "#inclued "
|
+---------------------------------------------------------------------------
!/var/lib/gems/1.9.1/gems/maruku-0.7.0.beta1/lib/maruku/input/parse_span.rb:424:in `read_inline_html'
!/var/lib/gems/1.9.1/gems/maruku-0.7.0.beta1/lib/maruku/input/parse_span.rb:93:in `read_span'
!/var/lib/gems/1.9.1/gems/maruku-0.7.0.beta1/lib/maruku/input/parse_span.rb:14:in `parse_span'
!/var/lib/gems/1.9.1/gems/maruku-0.7.0.beta1/lib/maruku/input/parse_doc.rb:175:in `block (3 levels) in substitute_markdown_inside_raw_html'
\___________________________________________________________________________

[distler]

The current trunk code appears to have regressed from 0.6.1

Correct.

This is one of several bugfixes in my (ill-named) pull request. It's fixed on my math-enabled branch of Maruku.

[distler]

Hmmm. It seems that I have problems, too, when the entities are in inline spans...

Grrrr.....

[distler]

Fixed, now.

[kentonv]

Thanks!

[distler]

Fix is now on trunk.