On Mon, 2003-01-06 at 20:42, Jeffrey Stedfast wrote:
On Mon, 2003-01-06 at 14:26, Adrian 'Dagurashibanipal' von Bidder wrote:On Mon, 2003-01-06 at 19:49, Steven P. Auerbach wrote:I recently stated that GPG/PGP messages from myself were correctly decrypted in Evolution 1.2.1, but messages from others are not. A reply from vbi clarified the problem, pointing out that that stated that earlier versions had support for inline PGP, but that this has been dropped, because the related problem of reliably verifying inline signatures is not solvable. Most GPG-encrypted messages sent to me are inline, so dropping support for them makes Evolution basically unusable for reading GPG-encrypted messages. KMail handles such messages just fine, so I don't believe there is any fundamental issue.Hi again! There are no fundamental issues with inline encrypted messages - this one is only the implementation problem of the MUA being required to scan the message body.actually, you are slightly wrong here. what's the content-type of the data that is in the encrypted blob?
Point taken. If the encrypted blob is not in an attachment, I'd say it's safe to assume that it's text in the encoding specified in the Content-Type header of the mail. If it's not, the user will have to deal with the consequences (we agree that inline pgp support can only be a works-mostly, after all).
[snip]I strongly suggest that support for inline GPG be restored, in some form.All this said, I agree with you: inline pgp should be supported. Thanks to evolutions bad performance in interoperability with other mail agents, I have actually much more PGP/MIME signed messages failing than I have inline signed messages fail. For all the theoretical benefits of MIME and the pitfalls of inline pgp usage, I find this really sad.apparently you aren't using a new enough evolution.
Excuse me? Latest Debian 1.2.1 package plus the CRLF patch you recently sent me, custom built (painfully). I'll be very happy if there was a next point release that fixes the gpg issues. (Oh yes, is there an easy fix to the 'pilot-link needs utf-8' error during configure that I keep getting. I've hacked evo to ignore it, but ... Attached is another message I could not verify, but others apparently could. Is there a way I can check that the CRLF patch is really in the code I've built? I didn't notice a difference. cheers -- vbi -- featured link: http://fortytwo.ch/gpg/subkeys
Attachment:
badsig.gz
Description: GNU Zip compressed data
Attachment:
signature.asc
Description: This is a digitally signed message part