HTML filters and unformatted HTML code

I experienced the following behavior of Signwriter, using the built-in output filters:

- WYSIWYG API + TinyMCE (latest)
- "Format Source code" turned off
- Filterpattern: /

.*?<\/h3>/
- H3 filter turned on for all formats

=> While parsing the HTML for the tag, only the first occurrence of the tag gets converted to an image. Also, the text after the occurrence is not being displayed at all: The following paragraphs just get lost.

Workaround: Turn on "Format Sourcecode" in the WYSIWYG API profile of TinyMCE.

Maybe the regex needs some tweaking if more than one occurrences of a search pattern per line exist? I suck at regex, otherwise I'd probably suggest something, but for now I'll be happy to receive a hint. :-)

Thanks,

Oliver