ExpressionEngine’s Textile plugin seems to be a bit borked, at least when comparing its output to not only the official implementation but a few others (e.g., TextMate).
Here’s an example: I was writing a comment on my blog, and typed in the following Textile snippet:
bq. Hopefully you’ll be there.
Hey, "Todd":http://unstoppablerobotninja.com/entry/an-event-apart-2009/#comment-606—at the moment, I’m actually speaking at Seattle and San Francisco. So basically, I’m currently on the lineup for the _not_-Boston AEA(An Event Apart)s. But hey, do keep an eye on the "MS&S(Markup and Style Society)":http://markupandstyle.org/, and hopefully we can catch up soon.
Here’s what the official Textile page returns:
<blockquote>
<p>Hopefully you’ll be there.</p>
</blockquote>
<p>Hey, <a href="http://unstoppablerobotninja.com/entry/an-event-apart-2009/#comment-606—at">Todd</a> the moment, I’m actually speaking at Seattle and San Francisco. So basically, I’m currently on the lineup for the <em>not</em>-Boston <acronym title="An Event Apart">AEA</acronym>s. But hey, do keep an eye on the <a href="http://markupandstyle.org/" title="Markup and Style Society">MS&S</a>, and hopefully we can catch up soon.</p>
This is what ExpressionEngine returns, however:
<p>
bq. Hopefully you’ll be there.
</p>
<p>
Hey, “Todd”:http://unstoppablerobotninja.com/entry/an-event-apart-2009/#comment-606—at the moment, I’m actually speaking at Seattle and San Francisco. So basically, I’m currently on the lineup for the <em>not</em>-Boston <acronym title="An Event Apart">AEA</acronym>s. But hey, do keep an eye on the <a href="http://markupandstyle.org/" title="Markup and Style Society">MS&S</a>, and hopefully we can catch up soon.
</p>
A few things seem to be going on here:
# Block-level elements such as blockquote, h1, h2, etc. are ignored by the Textile plugin if they appear on the first line of the textarea.
# Links aren’t processed if any non-URL characters are appended to the <code>“link”:URL</code> string.
There is, I believe, some wacky processing with lists (both ordered and unordered) as well, but I don’t have a test case handy. I’ve just experienced a lot of output inconsistencies with the plugin, and thought I’d raise the issue.
If anyone has any insight, it’d be much appreciated.
