SA Bugzilla – Bug 2739
Base64 issue during nightly corpora run ...
Last modified: 2004-01-23 12:28:58 UTC
I'm still trying to figure out which message(s) this comes from. Premature padding of base64 data at /home/felicity/SA/spamassassin-corpora/masses/../lib/Mail/SpamAssassin/Util.pm line 383. Premature padding of base64 data at /home/felicity/SA/spamassassin-corpora/masses/../lib/Mail/SpamAssassin/Util.pm line 383. Premature padding of base64 data at /home/felicity/SA/spamassassin-corpora/masses/../lib/Mail/SpamAssassin/Util.pm line 383. Premature padding of base64 data at /home/felicity/SA/spamassassin-corpora/masses/../lib/Mail/SpamAssassin/Util.pm line 383. Premature padding of base64 data at /home/felicity/SA/spamassassin-corpora/masses/../lib/Mail/SpamAssassin/Util.pm line 383. Premature padding of base64 data at /home/felicity/SA/spamassassin-corpora/masses/../lib/Mail/SpamAssassin/Util.pm line 383. Premature padding of base64 data at /home/felicity/SA/spamassassin-corpora/masses/../lib/Mail/SpamAssassin/Util.pm line 383.
Created attachment 1545 [details] message that triggers warning took me a bit to find this message, but ... it's from bugtraq, I've stripped out some headers and changed identification information, but it won't change the behavior.
a little more debugging ... the line that makes b64 barf is: ================================================ which is not a b64 encoded section, so I'm not sure why we're trying to decode it.
do we still have the code in the parser that looks for base64 data in MIME parts *without* a Content-Transfer-Encoding? (does that bug still exist in the field?) that strnig would probably trigger an attempted decode.
Subject: Re: Base64 issue during nightly corpora run ... On Tue, Dec 16, 2003 at 01:23:17PM -0800, bugzilla-daemon@bugzilla.spamassassin.org wrote: > do we still have the code in the parser that looks for base64 data in MIME parts > *without* a Content-Transfer-Encoding? (does that bug still exist in the field?) > > that strnig would probably trigger an attempted decode. yeah, that's what I figure it is. as I recall, we have the new parser in the code, but we're not actually doing anything with it yet, so the old code remains. it's going to be a "fun" bunch of work to convert over.
yep, the new mime parser takes care of this. ;)