Today, helping JN with the EMRC site, we found an oddity in the UVic Events calendar which results in invalid pages.

This tag has begun to show up in the <description> field:

<br type="_moz" />

Of course this is invalid -- there's no type attribute on <br/>. Trying to find out how it happens, I went and started to enter a new event in the calendar, using Firefox (I also get the same result on Chrome). I typed some text and some returns in the editor, and then clicked on Source. This is what I saw:

This is a quick test to see where the type="_moz" invalid HTML code is coming from.<br />
<br />
There were two breaks there.<br />
There was one break there.<br type="_moz" />

It looks as though my explicit linebreaks are converted to just <br/>, but the final linebreak (which I didn't enter -- it must be supplied by the editor) gets this invalid attribute.

This will cause pages which parse the RSS feed for HTML content to be invalid, unless we add a specific regexp to strip them out in our RSS feed reader (which is used on GRS, EMRC, and at least one other site). I've reported it to DW.