On 2025-01-04 09:37, Bruno Haible wrote:
No, that unfortunate property ... is a problem with the BIG5-HKSCS encoding, not with the GB18030 encoding. (Recall that GB18030 is more-or-less a Unicode transformation format like UTF-8, UTF-16, UTF-32.) See https://www.gnu.org/software/gnulib/manual/html_node/mbrtoc32.html
Thanks for correcting my fallible memory. Since that URL says this problem is only theoretical, I suppose we can assume we are not running on these purely-theoretical platforms.
Maybe the two new modules should depend on mbrtoc32-regular, to make this assumption clear?
