Code actions have bad offsets with Unicode in file

[RblSb]

Code actions have bad offsets with Unicode in file

[kLabz]

vshaxe/haxe-language-server#76 could maybe be applied there

[RblSb]

Seems like problem with most actions is:
token.getPos() returns offsets in utf8 (1 byte for abc chars, 2 bytes for абв chars)
and then document.rangeAt(pos.min, pos.max) iterates with charAt like arg offsets are single-byte
So there should be some way to convert utf8 byte offsets to characters first, before making range from it.

But Haxe4DisplayOffsetConverter only offsets surrogates currently

[kLabz]

Linked PR deals exactly with this on various commands (but not all that need it, apparently)

[RblSb]

Lets rename byteOffsetToCharacterOffset to utf16OffsetToCharacterOffset and make separate utf8OffsetToCharacterOffset?
It would be better to make int offsets as encoding-based abstracts, but this change can be too big.