3 ms·
It would be just as fast if it detected the locale correctly, and used the UTF8 state machine for the UTF8 locale only. Other locales could have their own state
by leni536 2y ago
It would be just as fast if it detected the locale correctly, and used the UTF8 state machine for the UTF8 locale only. Other locales could have their own state machines or just fall back to a generic implementation.
- stkdump 2y agoI'm wondering if you could easily build the state machine quickly at startup time. Unless it is a multi byte encoding it should be trivial.