Re: new gmime/gpg bug

[Jeffrey Stedfast <fejj stampede org>]

both this bug and the set_sender() bug are now fixed in CVS (the
set_sender() thing originally expected you to pass it a string
constructed with internet_address*_to_string() but it doesn't matter
now).

Jeff

On Sat, 2004-07-03 at 21:36 +0200, Albrecht Dreß wrote:
> Am 03.07.04 19:26 schrieb(en) Jeffrey Stedfast:
> > I don't see how GMime would change the multipart bounaries. I can only
> > presume that perhaps Balsa changed them after signing?
> > 
> > If you can show me where GMime went wrong, please do...
> 
> Please see the attached test program and the results output (combines  
> stdout and stderr). The app uses plain gmime, no gpgme. In the output, the  
> boundaries of the embedded multipart/mixed object have changed during the  
> signing process, so I assume this is done by gmime. And I don't think  
> calling g_mime_multipart_signed_sign() for an object should change it's  
> contents, shouldn't it? Furthermore, the source file responsible for  
> signing in balsa didn't change since 2 months, and it *did* work before  
> the latest 2 or 3 gmime cvs upgrades.
> 
> Cheers, Albrecht.
> 
> 
--