https://bugzilla.redhat.com/show_bug.cgi?id=825115

Bug ID: 825115 QA Contact: extras-qa@fedoraproject.org Severity: unspecified Version: 16 Priority: unspecified CC: fonts-bugs@lists.fedoraproject.org, i18n-bugs@lists.fedoraproject.org, psatpute@redhat.com Assignee: psatpute@redhat.com Summary: [kn_IN] Lohit Kannada glyphs of consonants with vowel signs -II, -EE, -OO and -AI should be optimized Regression: --- Story Points: --- Classification: Fedora OS: Unspecified Reporter: samjnaa@gmail.com Type: Bug Documentation: --- Hardware: Unspecified Mount Type: --- Status: NEW Component: lohit-kannada-fonts Product: Fedora

Description of problem:

Currently the Lohit Kannada fonts unnecessarily contain separate glyphs for consonants with vowel signs -II, -EE, -OO and -AI. These are merely sequences of other glyphs.

In the case of -II, -EE, -OO the glyphs are the same as those of the corresponding short vowels plus a separate glyph 0CD5 Kannada Length Mark ೕ (unlike in Telugu where the length mark ligates with the syllable). Therefore instead of doing substitutions of:

CONSONANT + VOWEL_SIGN_II/EE/OO --> CONSONANT_VOWELSIGN_II/EE/OO

the optimized way would be to do:

CONSONANT + VOWEL_SIGN_II/EE/OO --> CONSONANT_VOWELSIGN_I/E/O LENGTH_MARK

If this is done, any changes that are reflected on the CONSONANT_VOWELSIGN_I/E/O glyphs would automatically reflect for the long vowels as well without additional work being needed. For instance, see my recent report of bug 825104. If that bug is fixed for short vowels I, E and O, automatically it would reflect for II, EE and OO also.

As for -AI, it is merely sequence of glyph for -E plus 0CD6 Kannada AI Length Mark ೖ. So instead of doing substitutions of:

CONSONANT + VOWEL_SIGN_AI --> CONSONANT_VOWELSIGN_AI

the optimized way would be to do:

CONSONANT + VOWEL_SIGN_AI --> CONSONANT_VOWELSIGN_E AI_LENGTH_MARK

with same benefits as above.

Further, when handling combinations of:

CONSONANT1 + VIRAMA + CONSONANT2 + VOWEL_SIGN_II/EE/OO/AU

by separating the length mark as recommended here, the sub-base form of CONSONANT2 can be placed closer to the base CONSONANT1 which is also typographically a desirable factor in a good font. The rules would be:

CONSONANT1 + VIRAMA + CONSONANT2 + VOWEL_SIGN_II/EE/OO/AU --> CONSONANT1_VOWEL_SIGN_I/E/O + SUB_BASE_CONSONANT_2 + (AI_)LENGTH_MARK

Version-Release number of selected component (if applicable):

2.5.1

How reproducible:

Examine the internals of Lohit Kannada font.

Actual results:

Currently there are separate glyphs for consonants with vowel signs -II, -EE, -OO and -AI unnecessarily. These are merely sequences of other glyphs. Therefore any change effected on the component glyphs has to be re-done here.

Expected results:

The superfluous glyphs should be removed and the appropriate effect should be achieved by appropriate smartfont rules as indicated above.

Additional info:

This would also reduce the size of the font. Not an issue on laptops/desktops but nowadays Lohit fonts are finding their way into smaller screens such as smartphones, and it would be good to have a trim font with small footprint.