Re: Possible redisplay performance enhancements (revisited)

[YAMAMOTO Mitsuharu]

>>>>> On Mon, 24 Apr 2006 16:35:28 +0900, Kenichi Handa <address@hidden> said:

> I don't know why Gerd put that condition in the code.  But, as the
> code in fontset always uses a font specified for ASCII for ASCII
> code (see below), I think the proposed patch is safe.

Thanks.  It seems that the patch would not make things worse, at least.

I suspect get_glyph_face_and_encoding is not needed if `struct glyph'
contains `char2b' info and it is filled in append_glyph.  That would
reduce the number of calls to rif->encode_char and enhance the
performance when CCL is thoroughly used as in Mac.  But such a change
might be too drastic to do at this stage.

                                     YAMAMOTO Mitsuharu
                                address@hidden