Open ara4n opened 6 years ago
(obviously, in an ideal world, you'd be parsing both the m.in_reply_to
field to do proper semantic replies, as well as parsing the <mx-reply>
block in the HTML in order to nicely display the reply portion and clip it from subsequent replies.)
The display of this might have been fixed by 5896ec4 as there was a conflict of constraints between the image placeholder and the text placeholder which was causing multi-line messages to not expand properly.
However, there's no parsing specifically of m.in_reply_to
yet so yes, that will require a closer look so that it looks more elegant. Right now it just looks as good as the Markdown can produce, which isn't very good at all.
A typical reply looks like:
The formatted_body uses the cheeky
<mx-reply/>
custom tag to point out where the reply text is (so a client which understands it can hide it or handle it specially). Otherwise the plain text & HTML representation both include a fallback of the message contents which should show a usable version of the reply.The fallbacks are getting mangled by Seaglass however - presumably it's trying to parse the plaintext body as MD and getting it quite badly wrong. Riot shows the message based on the HTML-with-knowledge-of-the-mx-reply-tag as:
but Seaglass shows it as:
(also, where's half-shot's avatar gone? :)