[x] Changed the implementation to use a temp buffer
Also a -21~24% improvements on escaped sequences. Overall better... For us the parsing is less important (and not even x86/arm assembly anyways). But no one ever complained about things being faster. This indeed saves the vector allocation due to the String conversion.
Couldn't help it ;)
Thanks to your feedback here, I decided to:
Also a -21~24% improvements on escaped sequences. Overall better... For us the parsing is less important (and not even x86/arm assembly anyways). But no one ever complained about things being faster. This indeed saves the vector allocation due to the
String
conversion.