nls: add compose seq's for Cyrillic letters from KOI8-C into the main table
KOI8-C is a pre-Unicode codepage for Caucasus and Central Asia Cyrillic-based alphabets (Tatar, Buryat, Kalmyk, etc.) It was rarely used even before Unicode. Yet, it has its own compose table for letters like Җ, Қ, Ү, that are outside of basic Russian alphabet (which was fitting, since it is challenge to find a physical keyboard engraved with them). The table sits in nls/koi8-c/Compose.pre
.
Unfortunately, it was never converted and included into any of modern Unicode compose tables. As a result, to be able to type a few letters in Tatar, for example, the user has to install the whole input language into the system.
My proposal is to add those sequences either to en_us.UTF-8, or to the affected Cyrillic layouts - the latter option is justified by the fact that all proposed sequences are based on Cyrillic letters (those which are common across all Cyrillic layouts).
I have converted the sequences to Unicode, fixed and extended the table to match the scope of Latin and Greek tables. Current version: dotXCompose4libX11 - I'll keep updating it here, as well as in the repo.
I was initially working on an exhaustive table for as many Cyrillic letters as possible to be typable from any Cyrillic layout, but to make it acceptable for merging I'm planning to consult with fellow linguists & translators and make an "essentials" table, that is concise enough for either en_US, or affected Cyrillic layouts. Meanwhile, I would like to discuss the patterns in use, the scope to aim for, and everything that is needed to be accounted for to make this table (or a part of it) acceptable for libX11. Thanks.