[HarfBuzz] indic differences

[Jonathan Kew]

On 7 Sep 2011, at 10:57, Martin Hosken wrote:

> Dear Jonathan,
> 
>> Could you confirm (a) which rendering is which in your listing, and (b) what font was being used - thanks!
> 
> The first [] contains the hbng glyph sequence and the second [] contains the icu. The font is Annapurna SIL available from http://scripts.sil.org/AnnapurnaSIL

Thanks. Picking an item at random from your list...

> Failed at line 26: अंग्रेजी ['uni0905', 'uni0902', 'uni0917', 'uni094D', 'uni0947', 'uni0930', 'uni091C', 'uni0940'], ['uni0905', 'uni0902', 'uni0917094D0930', 'uni0947', 'uni091C', 'uni0940']

it looks to me like HBNG is incorrectly reordering the e-matra (0947) to precede the ra (0930). This then prevents it finding the conjunct for "gre" (0917, 094D, 0930), and the visual result would be quite wrong (with the e-matra appearing on the ga-halant instead of the ra).

And a bunch of other items show a similar pattern. Taking another example:

> Failed at line 13: अंकों ['uni0905', 'uni0902', 'uni0915', 'uni0902', 'uni094B'], ['uni0905', 'uni0902', 'uni0915', 'uni094B', 'uni0902']

the final anusvara (0902) has been incorrectly moved before the o-matra (although the anusvara on the initial a-vowel is ok).

JK