Support inline flag groups: (?i) (?m) (?s) etc.

[halotukozak]

Currently rejected at parse time (see RegexParser.scala group-header parsing, the case _ => unsupported("flag group") fallthrough).

Case-insensitivity ((?i)) is probably the most commonly needed one — could plausibly be implemented at the CharSet/Chars level by expanding each literal/range into its case-folded union at parse time, without touching the core Regex/Subset algebra. Multiline ((?m)) and dotall ((?s)) depend on anchor and .-matching support respectively.

[halotukozak]

Worth scoping explicitly: there are two tiers here. Simple case folding (single code point maps to single code point, e.g. 'a' <-> 'A') is the straightforward CharSet-expansion approach already sketched above. Full Unicode case folding per UTS#18 also includes multi-character folds (e.g. German ß folds to "SS"), which don't fit a single-code-point CharSet expansion at all — that piece would need either a documented restriction to simple folding for 1.0.0, or a dedicated multi-code-point mechanism (same family of problem as #22's grapheme clusters).