https://bz.apache.org/SpamAssassin/show_bug.cgi?id=7758
--- Comment #5 from Bill Cole <billc...@apache.org> --- (In reply to Dave Hansen from comment #3) > Ahh, my reproducer is actually testing wrap() directly. The changeset you > linked to is just hacking around the result. While this does fix the > symptom, and patches the only place where wrap() is used on headers headers > are generated, isn't it wrap() which needs fixing? Not really. The empty line is not intrinsically bad, it is only wrong in a header-wrapping context, where it is *always* wrong. This can be fixed by the caller, which 'knows' that the wrap is being done for a mail header. There is one call to wrap() (lib/Mail/SpamAssassin/PerMsgStatus.pm:939) that isn't wrapping text for a mail header and that could in principle be harmed by removing empty lines. The other 2 calls (also in M::SA::PMS) fix up the result rather that relying on wrap() to not return a blank line. A more elegant wrapping solution is probably possible, so as Henrik said: patches welcome. -- You are receiving this mail because: You are the assignee for the bug.
