[HarfBuzz] Change in HarfBuzz after version 0.90 ?

Richard Wordingham richard.wordingham at ntlworld.com
Fri Dec 27 16:35:50 PST 2013

On Fri, 27 Dec 2013 13:05:05 -0500
Ed Trager <ed.trager at gmail.com> wrote:

> Hi, Johnathan,
> In your patch:
> +    if (unlikely (unicode == 0x1A60)) return 254;
> ... could the return value be something a bit more symbolic?

I suppose you're looking for something like a new identifier
HB_UNICODE_COMBINING_CLASS_MAX, 254 being the maximum ccc value
allowed, and the aim being to ensure that U+1A60 orders after anything
with non-zero combining class (though I suppose there's probably not a
lot of call for it to behave properly with Telugu length marks).  A
functionally relevant name would be HB_UNICODE_COMBINING_CLASS_COENG,
as the intended function of U+1A60 in non-defective code sequences is
to turn the immediately following consonant into a subscript
consonant (it's the same word in Khmer and the relevant Tai languages).

My suggested change
simply gave U+1A60 the same modified combining class as U+0345, which
has the greatest assigned ccc value, namely 240.


More information about the HarfBuzz mailing list