[manton][jacky] Late reply here… When I moved from Jekyll to Hugo, I originally had a “blog engine” field internally which I imagined could allow someone to choose from various blog types. But I decided to go all-in on Hugo and going to stick with it because it makes sharing themes easier if users can depend on one format.
[fluffy]per #meta, even ephemeral data needs to be more durable than what a cache allows for. My rule of thumb for anything that lives in redis/memcached/etc. is that if it can’t be reconstructed by other context, it doesn’t belong in a cache.
[fluffy]so many things think of redis as being their own personal stomping ground, and that’s sort of how redis is designed in the first place. so you have to decide what your tolerance is for things evicting items they don’t own.
[jacky]so for Lighthouse, I'm thinking about modifying the inbound HTML a bit (like cleaning up the HTML, removing XSS-esque things, etc) but I wonder if it's "taboo" to run it through a pull-down parser (like Commonmark) to do things like automatically linkifying URLs that are found. Could that be considered overstepping how a Webmention service could provide that to a consuming site?
[jacky]Like what's to say that the consuming site _already_ does this? Granted, I can always make this configurable but I guess I wonder if I should make it opt-out or opt-in.