🚀 go-pugleaf

Piet van Oostrum <piet@cs.uu.nl> writes:

>>>>>> newsspam5REMOVETHIS@robf.de (R) wrote:
>
>>R> Piet van Oostrum <piet@cs.uu.nl> writes:
>>>> --=-=-=
>>>> 
>>>> Occasionally, I receive a message with a traditional PGP/GnuPG signature,
>>>> i.e. not a PGP/Mime signature but the ASCII-armored thing, like the
>>>> enclosed message. In te past I used mailcrypt for it, but now I have
>>>> switched over to VM-8's code. However this fails.
>
>>R> Clear text signatures should work, I am not sure what going wrong
>>R> here ...
>
>>R> Are you sure that vm-pgg is enabled?
>
>>R> Please check the message log and
>>R> (featurep 'vm-pgg) => Should be t
>>R> (locate-library "pgg") => Should find pgg.el(c) 
>
> This is all OK.
>
>>R> For GNU Emacs pgg.el comes with Gnus.
>
>>>> When there is a Mime-version header in the message VM seems to try to
>>>> verify the signature, but without Mime-version header it seems to just
>>>> ignore it. Would it be possible to correct this?
>
>>R> Definitely.
>
> I think I got that wrong, because now it ignores the signate if the ARE
> Mime headers and does the check when they are absent.
>
>>R> The message you have attached was in flowed format, probably
>>R> paragraphs got filled before sending or on their way to my news
>>R> server.
>
>>R> After fixing the headers it is working for me.
>
> Actually after trying again (and removing the Mime headers) it checks OK.
> Including the original that I thought failed. I can't understand what
> happened. I did not upgrade VM in the meantime, I haven't even restarted
> Emacs since.
>
> I do have another message that I received from someone else, that fails
> with vm-pgg but succeeds with mailcrypt. However it contains non-ASCII
> characters without giving proper Mime headers, so that may just be a
> problem with the character encoding.

Several VM functions are advised to call vm-pgg-cleartext-automode
which checks for the ASCII armor of a clear text signed or
encrypted message.

So if it does not work, then either vm-pgg-cleartext-automode is
not called for that message/mime-part or it fails to detect the
armor. 

There is a vm-pgg-cleartext-search-limit but IMHO it should be
high enough as the armor usually comes first.

Also some decoding might happens before searching for the armor,
but in general it should not be necessary for text parts with QP
or 8bit encoding.

Could you send me the other message for testing or ask the other
guy to prepare test message?

Thanks
Robert.