Dec 30, 2025

I got bored enough to add an RSS feed to this journal. Having a custom plain-text format for these records, I had to add a custom Atom generator. Overall it was a smooth sailing, with two gotchas:

1. CommaFeed (at least) doesn’t like escaped HTML inside content fields, even if type="html". So I had to switch to CDATA encoding.

2. Python’s builtin xml.etree library doesn’t support CDATA (CPython#81055 hanging from 2019).

3. So I switched to lxml that does. Turns out lxml.etree is slightly incompatible in managing namespaces. It took a bit of googling to find out, that instead of ET.register_namespace('', 'http://www.w3.org/2005/Atom'), I should do ET.Element(..., nsmap={None: 'http://www.w3.org/2005/Atom'}).

So, business as usual.