Peter Stuge <peter_at_stuge.se> writes:
> Simon Josefsson wrote:
>> > I couldn't find much MPI documentation so it may not be correct
>> > at all. :\
>>
>> Just checking, you know about the libgcrypt manual?
>>
>> http://www.gnupg.org/documentation/manuals/gcrypt/
>
> Yes. It lists functions but fairly little else.. :\
>
> In particular, gcry_sexp_nth_mpi() doesn't describe the int mpifmt
> parameter. It seems from the libgcrypt source code that this is
> actually an enum gcry_mpi_format, but the manual also doesn't say how
> to learn what to pass for that parameter. The sexp in question comes
> from gcry_pk_sign() but that also says nothing about how MPIs are
> stored in the result.. I looked at dsa.c in libgcrypt source and
> guessed..
So maybe there is no real advantage in using the mpi interface over
parsing the output from the sexp interface directly? Or does this fix
the actual problem magicly? If so, maybe a simpler fix to the original
sexp-parsing would also work.
But anyway, if it now works reliably, I'd say push it.
/Simon
_______________________________________________
libssh2-devel http://cool.haxx.se/cgi-bin/mailman/listinfo/libssh2-devel
Received on 2009-12-17