On Fri, 2004-07-30 at 03:40, Mahesh T. Pai wrote: > Owen Taylor said on Thu, Jul 29, 2004 at 02:44:46PM -0400,: > > > ka + h + ya + u > > I am not sure that I have understood you fully here, but ... > > In respect of the first line on the png, the sequence, > > (ka + h + u + ya) > > Should show k, explicit virama, u and ya separately. That is, in > akruti fonts, this should display 4 independent glyphs. But the first line is "ka + h + ya + u" which in my understanding is the correct spelling for the syllable "kyu". I was trying to debug why it was rendering incorrectly. > I think the correct unicode sequence for the image shown should be:- > > ka + u + h + ya (U+0D15 + U+0D41 + U+0D4D + U+0D2F) > > The halant is necessary, since situations where the ya has to render > as a separate character exist. > > Kindly have a look at http://www.unicode.org/review/pr-37.pdf, which I > believe may be relevant to the rendering shown here. The _proposal_ is > to require use of a zwnj in situations where the subsequent character > has to render separately. As I read Peter Constable's proposal, I think you would need: ka + u + <space> + zwj + h + ya That is, the syllable 'ku' followed by an isolated post-base form of ya. ka + u + h + ya Is currently rendered by Pango as 'ku' followed by a lonely chandrakkala, followed by the isolated form of ya, which I believe is correct. It isn't really a well formed sequence. > The second sequence shown is correct. > > Also, it will be helpful if you list the bugs you are looking at. This problem was actually unrelated to the bugs I was looking at (Mostly http://bugzilla.gnome.org/show_bug.cgi?id=121671), it just happened to show up in one of my test cases. Regards, Owen
Attachment:
signature.asc
Description: This is a digitally signed message part