Files
sed/tests
Diomidis Spinellis 11f235d6f0 Use regex::bytes also for Unicode matching
Keep using regex_fancy for regexes containing backreferences.
This removes a complex unneeded matcher selection RE and also
provides an overall small performance boost.

     access-log-no-del pre-use-bytes is  1.02 times faster than use-re-bytes
    access-log-all-del use-re-bytes is  1.02 times faster than pre-use-bytes
   access-log-translit pre-use-bytes     is similarly fast as   use-re-bytes
access-log-complex-sub pre-use-bytes     is similarly fast as   use-re-bytes
     access-log-append pre-use-bytes     is similarly fast as   use-re-bytes
             remove-cr use-re-bytes is  1.08 times faster than pre-use-bytes
          genome-subst use-re-bytes is  1.05 times faster than pre-use-bytes
            number-fix pre-use-bytes is  1.02 times faster than use-re-bytes
           long-script use-re-bytes is  1.01 times faster than pre-use-bytes
                 hanoi pre-use-bytes     is similarly fast as   use-re-bytes
             factorial use-re-bytes is  1.03 times faster than pre-use-bytes

Suggested by: Andrew Gallant @BurntSushi
2025-12-16 21:07:09 +01:00
..
2025-12-14 18:07:22 +01:00