Hi David,

David J. Weller-Fahy wrote:
> I have not tried to modify the default algorithm yet (no time today
> other than the initial patch and install), however I have used the patch
> to sign emails to myself and others and compared the claimed algorithm
> used.  Without the patch today's tip show's SHA-1 as the algorithm, with
> the patch the algorithm today's tip show's SHA-256 as the algorithm.

So far, I've defaulted $smime_sign_digest_alg to "sha256".  I'll be
glad to change that, but was just picking what I hoped was a reasonable
value.

> Tomorrow (or Sunday) I'll take a look at the signatures in Thunderbird,
> and look at the extracted algorithms.  I'll also try setting the digest
> alrogithm manually tomorrow, and report back.
> 
> Let me know if there are any other specific things I need to look for,
> other than mismatches.

Thank you for taking a look!  I mostly wanted to make sure this patch
fixed the problem and didn't cause any other issues.

I think the original reporter must have had a smime cert that specified
sha256 as the default message digest algorithm.  smime was then *generating*
the signature using sha256, but the micalg parameter in the email header
was hardcoded to say "sha1".  Thunderbird apparently didn't like that
and so was rejecting the signature.

This patch hopefully aligns the header with the actual digest algorithm,
but requires updating $smime_sign_command to have a placeholder
specifying the algorithm to use: "-md %d".

In any case, if you were able to just test the signatures with the patch
to make sure Thunderbird and mutt were happy with them, that would be
very helpful.  If you were somehow able to replicate the original
problem and see if the patch "fixed" it, that would be even better (but
not required).

Thank you in any case!

-- 
Kevin J. McCarthy
GPG Fingerprint: 8975 A9B3 3AA3 7910 385C  5308 ADEF 7684 8031 6BDA
http://www.8t8.us/configs/gpg-key-transition-statement.txt

Attachment: signature.asc
Description: PGP signature

Reply via email to