Hi Gregor On Thu, Sep 06, 2012 at 10:22:57PM +0200, gregor herrmann wrote: > On Thu, 06 Sep 2012 22:11:28 +0200, Salvatore Bonaccorso wrote: > > > I debugged this a bit further and the problem seems to me more that as > > you said spamassassin 'is wrong', as IO::Socket::SSL is checking > > correctly his alternatives ... > > Wow, impressive! > > > Here is an output traced running spamd with --ssl, to see first what > > happens: > > Out of curiosity: How did you do this?
It's done with Devel::Trace. If you have the module, then you can run a 'trace' with perl -d:Trace ./script.pl, giving something like the shell's "-x" option. > > It does not seem to be an error of IO::Socket::SSL to me (but I'm not > > expert on this), but rather that Spamassassin is handling $^S here in > > 'strange' way? > > Agreed. > > > Opinions? I would go for reassigning the bug to spamassassin and in > > particular lowering the serverity (spamd is started and error messages > > are missleading). > > Agreed again. > > (It would be nice to hear of Bjørn sees the same, i.e. a running > spamd after the misleading error messages.) Okay, let's wait a few days until defintively reassigning. Bjørn could you confirm that indeed nevertheless these error messages printed out the spamd starts? Regards, Salvatore
signature.asc
Description: Digital signature
