Use language aliases.

[scossu]

This PR introduces some framework changes and some language-specific ones.

1. It uses alias_of instead of conf in the index file to indicate that a language is an alias of another table, and it omits the .yml suffix in the reference.
2. Internally it removes the duplication of tables previously caused by identical (alias) tables being populated. This has no effect for the end user except perhaps a faster transliteration.
3. Using "base" scripts for transliteration, earlier discouraged, is now disallowed, even in aliases. Base tables are not directly populated in the DB, so pointing an alias to a base table will result in an error when populating the database.
4. As a result of 3., some tables had to be renamed: _bengali_base to bengali, and _gurmukhi_base to gurmukhi_generic.

[scossu]

@RandyBarry In moving Gurrmukhi, I noticed that the nasalization rules only apply to Panjabi. Should I move them to the Gurmukhi generic table so that other aliases can use it, or is it a Panjabi-specific feature?

[RandyBarry]

I will have to think about the impact of the framework changes you made. You know best in terms of what will allow SS to function better. I happened to look at "https://scriptshifter-test.knowledgetx.com/" just before I saw your message about these changes. I saw that the test version has been updated. The conversions I then tested (Cherokee, Coptic, Glagolitic, and Inuktitut) looked fine. The one thing I noticed that was unexpected was the order of languages. Tibetan appears after Dzhongkha (Tibetan). It used to be alphabetically with other "T" languages. "Cyrillic (Generic)" is also out of order and appears after one of the Cyrillic-based languages. Did the framework change affect language order?

[RandyBarry]

Relating to Gurmukhi: yes, the preprocessing of the nasalization characters needs to be done with any language that uses the Gurmukhi script. Nasalization applies to more than just the Panjabi language.