Nuqta consonants and vowels about lipika-ime

[ratreya]

As I said in Issue #5, the Unicode glyph name RRA is inconsistent for Devanagari. To fix the Nukta attachment issue, can you please edit /Library/Input Methods/LipikaIME.app/Contents/Resources/Schemes/Default.ime and add the following at line 5:

{CONSONANT}{SIGN/NUKTA}{DEPENDENT} [CONSONANT][SIGN/NUKTA][DEPENDENT]

Then restart your computer and try it out. Let me know how that works. If it works well for you, I will make that change in the application binary.

[ratreya]

The only problem I found was that Dza does not work, you have to type Daz to get the desired output. Everything else seems to work. But please try it out and confirm.

[dalahast]

I've added the line (Github seems to have replaced the tab character with a space, which initially caused LipikaIME to beep and not load correctly, but I fixed it) and ran killall LipikaIME from Terminal to restart the IME without having to restart the entire system. I also previously edited the mappings to output ड़ for RRA, which means I can easily get (e.g.) ड़, ड़ि, ड़ी, without inputting the nuqta manually.

Currently, I get the following when inputting the nuqta manually:
ड् for Dza
ड़ for Daz
ड़ि for Dzi
डि़ for Diz
ड़ो for Dzo
डो़ for Doz
This confirms the behavior you observed with having to place the modifier at the end of the sequence for ड़ (likely because ड is internally seen as "unmodified" as opposed to e.g. डि Da + i).

[ratreya]

Actually, make the following changes and it should all work correctly: 5309fde

[dalahast]

That seems to have fixed it, thanks. Typing the nuqta modifier immediately after the consonant works for any combination now (so Dza now correctly outputs ड़).

With the modified Devanāgarī mapping for RRA, this also means that ऱ is still easily accessible through rza (and ड़ remains easily accessible for users with the default mapping, using Dza).

[ratreya]

Perfect - thanks for confirming. I will close the issue after the change makes it into the next release.

[ratreya]

There were some lingering issues with the fix above. It did not work for Hindi and for languages that did not have a Nukta defined. All of this is fixed in: 2ab8e41.