GNOME Bugzilla – Bug 569161
Kannada Character U+0CB1 does not render its below-base form.
Last modified: 2012-08-25 20:47:41 UTC
+++ This bug was initially created as a clone of Bug #358224 +++ Bug 358224 was for allowing U+0C31 to render a below-base form. On similar lines some Kannada fonts have a below-base form for U+0CB1, e.g. Red Hat's Lohit Kannada, Sampige distributed by various linux distros and Sampige's successors Kedage/Mallige from http://brahmi.sourceforge.net/downloads.html. Should Pango allow U+0CB1 also to render a below-base form? Sampige, Kedage & Mallige also have a glyph looking like a below-base form for U+0CDE but they do not have ligature rules to convert U+0CDE U+0CCD to the glyph. Should Pango allow U+0CDE also to render a below-base form?
Created attachment 127242 [details] [review] Changes to allow U+0CB1 & U+0CDE render below-base forms Thie diff is taken against the current version of the file available at http://svn.gnome.org/svn/pango/trunk/modules/indic/indic-ot-class-tables.c. In order to disallow below-base form of U+0CDE, revert second change in the diff. Note that fonts without a below-base form of U+0CDE do not display an acceptable alternative with U+0CDE's character class set to _bb: C (any consonant) U+0CCD U+0CDE -> C U+0CDE (blwf if possible) U+0CCD (blwf if possible) -> C zha halanth (no ligation) which doesn't make it obvious to the reader that the halanth is supposed to be for C, not on zha.
Created attachment 127245 [details] Rendering of "0ca8 0ccd 0cb1 0ca8 0ccd 0cde" with & without my patch The bottom text entry is from gnome-character-map as rendered by the vanilla Pango version from Mandriva Linux's pango-1.10.0-3mdk. The top text entry is from the font selection dialog as rendered using Pango compiled from pango-1.10.0-3mdk.src.rpm with my patch. The font in both cases is Lohit Kannada from lohit-fonts-2.3.1-1.fc10.
I too noticed that above said combination has an issue with rendering of below base characters(I used the latest version of pango (pango-1.26.0)). I also noticed it is happening only with the combination having 0cb1 and 0cde. I would also like to add that as Unicode itself says 0cde is an obsolete historic character, you may prefer to ignore the issue with the 0cde. Shankar
Thanks Shankar for your comment. I will include this bug fix in patch to 594101 bug.
HarfBuzz has been merged, Indic shaper removed. Marking OBSOLETE.