For single-line strings it's easy to forget the termination character for these languages. This PR tries to address that by adding the SCE_..._STRINGEOL state marking unterminated strings similarly to other languages.

I tried to follow the implementation of other lexers - typically the implementation looks like

if (sc.state == STRING && sc.atLineEnd) {
    sc.ChangeState(STRINGEOL);
    sc.ForwardSetState(DEFAULT);
}

However, the sc.ForwardSetState() at line end isn't quite correct - many lexers contain

if (sc.atLineEnd) {
    styler.SetLineState(sc.currentLine, lineState);
    ...
}

at the end of the main loop where they set line state and update various state variables and sc.ForwardSetState() when handling unterminated string may cause this code to be skipped.

I only ran into this problem with the TOML lexer and the provided unit test so I implemented "correctly" only this lexer; I left the "incorrect" implementation (which is used by many existing lexers too) with the remaining lexers. It probably doesn't matter much as unterminated strings are supposed to be fixed by users soon and not to remain in the code for a long time.