Open smnandre opened 3 days ago
Based on your information, text/html
seems the best option here.
But, I don't know internals well enough to see if changing the current content-type header will break something for someone.
But this is not valid html, so text/html would be incorrect, isn't it? Also, a custom content-type is currently how the CSRF protection works. This can be replaced with another header, but this should be accounted for.
But this is not valid html, so text/html would be incorrect, isn't it?
Gray area... let's say it would not not be "full - html- document - valid"... but what we send is HTML (part of document but still)
In the end, to be 100% honnest, this header works and will continue to work.
The main problem is the DX one.
And i won't fight for this if it created trouble for a single user (this is where i will obv listen to and follow you guys)
RFC: use
text/html
ortxt/vnd.ux-live-component
To improve DX and align with IANA, we can:
txt/vnd.ux-live-component
)text/html
and add another header to check front/back exchangesOr we can keep current one :)
LiveComponent currently require the
application/vnd.live-component+html
header in both communication from both browser and server when communicating.This header has the disadvantage to not be show in browser developer tools (Chrome and Firefox can only display the content after a couple of clicks.. for every request).
The other problem is that it seems a bit .. wrong. Or let's say not perfectly in line with the different RFC.
subtype
Text
Application
So i feel we are more in line with
text
here.vendor
We should register our vendor names to be perfectly in line with the RFCs, but i guess none of us has the time, energy, and need to do such thing.
suffix
+html is not a valid suffix (see full list of suffixes). This is recommanded both for official mime registration but also as general practice