Some Kannada and Telugu diacritics have content far beyond right side bearing

[pvuorela]

While implementing virtual keyboard layouts for Kannada and Telugu using Lohit fonts, noticed that some characters have problems fitting their designated area. Looking at the fonts, I see that some diacritics have right side bearing somewhere in the middle of their content. Is this intentional? Would expect mostly ligatures used for real text and isolated diacritic input on text field leaves cursor in the middle, looking silly. For comparison, Devanagari diacritics fit inside the bearings a lot better.

Some involved characters, Kannada: U+0CCC, U+0CBE, U+0CCD, Telugu: U+0C4D.

Disclaimer: don't speak the languages and Qt bug affects my case too: https://bugreports.qt.io/browse/QTBUG-50642