On Sat, May 18, 2024 at 11:29 AM Tim Bray via Datatracker <noreply@xxxxxxxx> wrote:
Reviewer: Tim Bray
Review result: Ready with Issues
Thanks for adding the Internationalization Considerations section, which
addresses the concern in my previous review that this issue needed more work.
I have small issues:
1. There is a reference to "control characters" without any citation, and the
definition is not as obvious as one might think. A good reference is [UNICODE]
section 23.1, which may be summarized as "65 code points in the ranges
U+0000-U+001F ("C0 Controls") and U+0080-U+009F (“C1 Controls”), plus U+007F,
"DEL" 2.
RFC 20, section 4.1 (or some update) would do. But, I wonder if that part is right, because of the way JSContact is defined.
That can certainly contain CR, LF, TAB, in common cases. But if you get BEL or SUB, that may be a problem.
This section might benefit from a reference to RFC9413, which has a
thorough discussion of dealing with input data that is invalid for some reason.
3. There are 3 options when receiving a message with invalid characters: delete
them, replace them (Unicode provides U+FFFD for this purpose), or reject the
message.
The draft mentions two of these but not the replacement option. Is it
never appropriate in the JMAP context?
I think most people do this, because they add new emoji every year, but I think the operating system takes care of it most of the time.
For example, I am listed as "Rob 🎧" for one person, because they call me with audio problems (it showed up in an email, so I found this out when the iPhone contact list name drifted into email). But you can see this pattern drifting to whatever new ones they add.
thanks,
Rob
-- last-call mailing list -- last-call@xxxxxxxx To unsubscribe send an email to last-call-leave@xxxxxxxx
