[HarfBuzz] A problem about indic shape (for khmer)

[Behdad Esfahbod]

So, looks like these are all split matras (according to
IndicMatraCategory.txt) that do not have a decomposition:

U+0AC9
U+0B57
U+0F77
U+0F79
U+17BE
U+17BF
U+17C0
U+17C4
U+17C5
U+1925
U+1926
U+1B3C
U+1C29
U+A9C0
U+1112E
U+1112F
U+111BF

Can people familiar with these characters please clarify what the desired
behavior of these characters is?

Thanks,
behdad

On 04/19/2012 02:36 PM, Behdad Esfahbod wrote:
> On 04/19/2012 02:24 PM, datao zhang wrote:
> 
>> The following Matra can’t be decomposed:
>> 0x17C0
>> 0x17C4
>> 0x17C5
>> 0x17BE
>> 0x17BF
> 
> None of these are decomposable in Unicode, so I have to to adjust the shaper
> for that.  Looks like for Khmer all split matras add U+17C1 to the left.  I
> have a couple ideas how to implement this now.  Will take a look.
> 
> Thanks,
> behdad