wrong results and hang in gcry_mpi_invm()

[Werner Koch]

On Mon,  7 Dec 2015 09:09, hanno at hboeck.de said:

> I want to report two issues with the function gcry_mpi_invm().

Thanks.  Actually, we know about this and I have a fix stashed for quite
some time.  However, there are backward compatibility problems which is
the reason why it has been delayed.  The MPI functions have originally
been written for internal use by Libgcrypt and thus they have some
restrictions (e.g. no ENOMEM and that limited invm)

Thanks for raising this again.  We should fix it for 1.7.


Shalom-Salam,

   Werner

-- 
Die Gedanken sind frei.  Ausnahmen regelt ein Bundesgesetz.