Re: mbcel module for Gnulib?, incomplete multibyte sequences

[Bruno Haible]

Paul Eggert wrote:
> >> in UTF-8 the byte sequence E0 80 is not an incomplete character
> >> (in the sense that additional bytes may lead to a complete character),
> >> because every byte you append to E0 80 causes glibc mbrtoc32 to return
> >> (size_t) -1. Yet glibc mbrtoc32 returns (size_t) -2 for E0 80.
> > 
> > And gnulib/lib/unistr/u8-mbtouc-aux.c does it wrong as well!
> > The return value for E0 {80..9F} should be (size_t) -1, because
> > U+0800 is E0 A0 80.
> > 
> > I'll fix the gnulib part soon. Very good point. It looks like few people
> > understood the implications of
> > https://www.unicode.org/versions/Unicode15.0.0/ch03.pdf page 125, table 3-7.
> 
> I hope we don't need to replace mbrtoc32 merely because of this obscure 
> issue.

We don't need to, because
  - it's not blatant POSIX failure,
  - this bug has been in glibc for 20 years and in GNU libunistring for
    more than 10 years, and no one noticed,
  - the difference is only whether mbrtowc() returns (size_t)-1 vs.
    (size_t)-2.
 
Bruno