REfactor and rewrite all regexes for syllbale break identification

[andjc]

Refactor and rewrite all regexes for syllable break identification

Myanmar (Burmese) and Sgaw Karen currently have bugs in them. But would be fest to refactor them with improved regular expressions.

To approaches: 1) A regular expression designed for full syllable identification

• identify existing syllable identifiiers 2) Assume break before consonants and remove unnecessary or incorrect breakpoints:
• insert zwsp before consonant (but suppress zwsp before sub-stacked consonant)
• remove zwsp at places that js recognises as a word boundary, ie, at start of string, after spaces or punctuation.
• remove zwsp before final consonants within syllablee
• other points?

[andjc]

Replaced issue #6

[andjc]

([^\s`~!@#\$%^&*()_-=+[{]}|;:'",<.>/\?][\u1000-\u1021\u1023-\u1027\u1029\u102A])