Niels Lohmann 6a0cf2ba2f Recover by error kind instead of by error message ...

The lexer records an error_kind with each error_message, and
recover_token() dispatches on it instead of comparing message prefixes.
add_codepoint() now uses encode_utf8(), and the surrogate pair
formula is shared by scan_string() and recover_low_surrogate().
The stubs for AllowRecovery == false select via negation, which astyle
leaves alone.

Signed-off-by: Niels Lohmann <mail@nlohmann.me>

2026-10-21 09:06:28 +02:00

..

nlohmann

Recover by error kind instead of by error message

2026-10-21 09:06:28 +02:00