[Indic] Recategorize some Kannada right matras

Kannada failures down from 3.5% to 2.93%.

src/hb-ot-shape-complex-indic.cc

@@ -142,7 +142,7 @@ consonant_position (hb_codepoint_t u, hb_ot_map_t *map, hb_font_t *font)
 				  IS_ORYA(u) ? POS_AFTER_POST  : \
 				  IS_TAML(u) ? POS_AFTER_POST  : \
 				  IS_TELU(u) ? (u <= 0x0C42 ? POS_BEFORE_SUB : POS_AFTER_SUB) : \
-				  IS_KNDA(u) ? (u != 0x0CD5 ? POS_BEFORE_SUB : POS_AFTER_SUB) : \
+				  IS_KNDA(u) ? (u < 0x0CC3 || u > 0xCD6 ? POS_BEFORE_SUB : POS_AFTER_SUB) : \
 				  IS_MLYM(u) ? POS_AFTER_POST  : \
 				  IS_SINH(u) ? POS_AFTER_SUB   : \
 				  /*default*/  POS_AFTER_SUB     \

test/shaping/texts/in-tree/shaper-indic/indic/script-kannada/misc/MANIFEST

 misc.txt
+right-matras.txt

test/shaping/texts/in-tree/shaper-indic/indic/script-kannada/misc/right-matras.txt

+ಸ್ಕು
+ಸ್ಕೂ
+ಸ್ಕೃ
+ಸ್ಕೄ
+ಸ್ಕಾ
+ಸ್ಕೕ
+ಸ್ಕೕ