Re: libnettle/libhogweed WIP

[Ted Zlatanov]

On Wed, 26 Jul 2017 21:51:30 +0300 Eli Zaretskii <address@hidden> wrote: 

EZ> I don't know.  Since no one else reported the same problem, perhaps
EZ> it's something specific to my local configuration, although I doubt
EZ> that.

Yes, I also think it's likely to be a real problem, that's why I've
spent a lot of time on it. But it seems to only occur on that platform.
Do you have a way to test with a newer GnuTLS release?

Is there anything in the code that looks suspicious to you? I've looked
it over several times and don't see how it could be causing garbage data.
Maybe the storage math is wrong somehow?

Can you trace the exact point at which the GnuTLS functions are called
on encryption and decryption?

  ret = ((encrypting ? gnutls_aead_cipher_encrypt : gnutls_aead_cipher_decrypt)
         (acipher, vdata, vsize, aead_auth_data, aead_auth_size,
          cipher_tag_size, idata, isize, storage, &storage_length));

Do those numbers match up with the cipher parameters?

  (:cipher-id 16 :type gnutls-symmetric-cipher :cipher-aead-capable t 
:cipher-tagsize 16 :cipher-blocksize 16 :cipher-keysize 32 :cipher-ivsize 12)

EZ> I offer my help.  What would you like me to do, try, find out?  Just
EZ> ask.  Did the information I provided already give any clues, or any
EZ> leads, or any ideas for further tinkering?

Your information was thorough but didn't tell me where the error is in
the Emacs code, and I don't yet know how to reproduce it.

Ted