Re: #'query-coding-region and invalid Unicode sequences.
David Kastrup writes:
Basically I read him as "David convinced me that we don't
need sanity"
and I was protesting that without actually checking for the presence or
absence of sanity.
No, I was saying that in your use case for TeX messages, you were just
going to throw away the truncated (and therefore invalid) prefix and
suffix sequences, and resync (which UTF-8 is designed to do reliably)
and translate the rest. AIUI in earlier versions of XEmacs that was
impossible.
As far as "let David be David," the point is that this is a very
specialized usage: you *know* that TeX is going to produce breakage of
a very specific kind, that no other application I know of does. I
don't see why it should be the responsibility of the core Emacsen
developers to maintain such code, except that writing codecs is a
specialized activity due to the unfortunate way that the codec API is
deeply embedded in the C code and/or the unnecessarily arcane CCL.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta