Re: Scan of regexps in Emacs (March 17)

[Paul Eggert]

On 4/2/19 9:52 PM, Eli Zaretskii wrote:
> Emacs sometimes does use regexps when dealing with unibyte
> buffers and strings, where these ranges could be significant. 

Quite possibly we'd want to support the case where a range in a unibyte
pattern is used against a unibyte buffer or string, as the semantics are
straightforward there and the code should work (or at least I think it
should work; the implementation is not clear). The troublesome cases
with ranges and raw 8-bit bytes occur when either the pattern or the
buffer/string is multibyte - there it's not clear what the
implementation *should* do, much less what it does.