Re: Bug 130397

[David Kastrup <dak@gnu.org>]

Geoff Kuenning <geoff@cs.hmc.edu> writes:

> Ken writes:
>
>> Geoff has a much better understanding of the underlying spell search
>> engine.  Perhaps he can shed additional light on this topic.
>
> I just looked at the code to be sure my memory is correct.  Here's
> the short rundown: in the '-a' interface, ispell interfaces with the
> outside world purely in a byte-indexed mode.  It is perfectly
> capable of handling UTF-8 and similar multi-byte encodings, but when
> it reports the offsets of incorrect words, it does so as a byte
> offset, not a character offset.
>
> Does emacs provide an underlying byte-indexed interface to the
> buffer?  If so, life should be easy: just have ispell.el use that
> interface.

You are wrongly assuming that the buffer is maintained in UTF-8.  It
isn't.  Byte indexing is not going to be fun with regard to
efficiency, unless we get some interface that will, while writing out
a file in UTF-8, store an array of byte/character correspondences for
the UTF-8 (or whatever other) character conversion somewhere.

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum