https://bz.apache.org/SpamAssassin/show_bug.cgi?id=5590
--- Comment #57 from Mark Martinec <[email protected]> --- Please see Bug 7091 and further tickets referenced there. Text being processed by rules and plugins should not have utf8 flag enabled since 3.4.1, i.e. it is either transcoded to UTF-8 octets (with normalize-charset enabled), or left as original bytes. There was a bug in HTML decoding which could result in Unicode text (utf8 flag on) which would slow down processing. I hope to get rid of most of 'use bytes' in modules eventually, as this pragma makes no difference when we can ensure that both text and rules are in octets (utf8 flag off). -- You are receiving this mail because: You are the assignee for the bug.
