On some architectures (PowerPC at least -- I think
Sparc might be affected as well), the char type
defaults to unsigned -- therefore the test at line 50
of dktest.c is always false.
Attached is a minimal patch which fixes this exact bug,
but I suspect there might be other related ones....
I'm reachable at : svallet at gmail dot com
Patch against libdomainkeys-0.68
Logged In: YES
user_id=1254642
Included in cvs, thanks
Logged In: YES
user_id=268278
i just compile exim with dkim support on a new Athlon64
usign the very same options i used on other non 64b
platforms (intel and amd) and it hangs on dk tests. Looks
like it could be the same problem