While this doesn't impact UTF8 strings built up in Swift, Strings decoded from data (like through Vapor, for example) may not be marked as UTF8, which makes the lexer UTF8 conversions very expensive. In evaluation of 450kb requests, performance improved from 27min to 5sec with this adjustment.
While this doesn't impact UTF8 strings built up in Swift, Strings decoded from data (like through Vapor, for example) may not be marked as UTF8, which makes the lexer UTF8 conversions very expensive. In evaluation of 450kb requests, performance improved from 27min to 5sec with this adjustment.