It doesn't matter for html and xhtml, but I'm updating some code that used python-markdown 2.0.x to generate RSS to current versions, and I'm getting output with <pubDate>...</pubdate> - that is, the tags don't match (rss is not case-insensitive, or at least most of the XML-based tools for chewing on it aren't.) _serialize_html at https://github.com/waylan/Python-Markdown/blob/master/markdown/serializers.py#L179 downcases the tag to compare it against some constants (script, style, HTML_EMPTY) which seems sensible, but then it writes out the downcased version and not the original.

Should I instead be writing a custom serializer (2.0 didn't even have them so it didn't come up...) or are there other approaches that make sense now?