Things have been quiet here recently. The UTF8 text literal packet
idea (using a 'u' instead of 't') had a bit of discussion. Here's a
formal proposal, with text:
In section 5.9 change this:
If it is a 't' (0x74), then it contains text data, and thus may
need line ends converted to local form, or other text-mode
changes.
To this:
If it is a 't' (0x74), then it contains text data, and thus may
need line ends converted to local form, or other text-mode
changes. The character set used is not specified.
And add:
If it is a 'u' (0x75), then it contains text data known to be UTF-8
encoded, and thus may need line ends converted to local form,
character set conversion, or other text-mode changes.
I think this best describes what was discussed here. Notably, it does
not say that 't' should not be UTF-8. 't' can be whatever text people
want to put in it. All 'u' means is that this text is *known* to be
UTF-8. Implementations can handle this however they like (convert
text in and out, always use 't', etc.)
David