https://bugs.exim.org/show_bug.cgi?id=2430
--- Comment #8 from Andreas Bergmann <andreas.bergm...@cyren.com> --- Just for the records: Using gcc 6.3.0 with '-O2' optimizations in my test case, the loop is still the faster option. However, applying that patch (being targeted at overall performance) also solves the issue. So 'case closed', thanks a lot. while-loop: 1048576 bytes 0.000049 sec (?i)abc 1048576 bytes 0.000141 sec (a|A)(b|B)(c|C) 1048576 bytes 0.000103 sec [aA][bB][cC] 1048576 bytes 0.000033 sec abc memchr+patch: 1048576 bytes 0.000095 sec (?i)abc 1048576 bytes 0.000144 sec (a|A)(b|B)(c|C) 1048576 bytes 0.000107 sec [aA][bB][cC] 1048576 bytes 0.000033 sec abc -- You are receiving this mail because: You are on the CC list for the bug. -- ## List details at https://lists.exim.org/mailman/listinfo/pcre-dev
