Re: GnuTLS soname bump in rawhide

2013-02-07 Thread Daniel P. Berrange
On Thu, Feb 07, 2013 at 08:34:18AM -0500, Dan Winship wrote: > On 02/07/2013 07:27 AM, Daniel P. Berrange wrote: > > The other potential impact is that gnutls no longer uses libgcrypt, > > but instead uses nettle. > > > > This can impact apps that were using gnutls in a threaded environment > > be

Re: GnuTLS soname bump in rawhide

2013-02-07 Thread Dan Winship
On 02/07/2013 07:27 AM, Daniel P. Berrange wrote: > The other potential impact is that gnutls no longer uses libgcrypt, > but instead uses nettle. > > This can impact apps that were using gnutls in a threaded environment > because they probably have used 'gcry_control' to register a thread > impl

Re: GnuTLS soname bump in rawhide

2013-02-07 Thread Daniel P. Berrange
On Wed, Feb 06, 2013 at 09:59:36PM +0100, Tomas Mraz wrote: > I'm rebasing gnutls in rawhide to gnutls-3.1.7. This is much awaited > upgrade by many. However note that potentially patented ECC sources are > removed. > > The rebase bumps soname to libgnutls.so.28. The other potentially > disrupting

Re: GnuTLS soname bump in rawhide

2013-02-06 Thread David Woodhouse
On Wed, 2013-02-06 at 21:59 +0100, Tomas Mraz wrote: > I'm rebasing gnutls in rawhide to gnutls-3.1.7. This is much awaited > upgrade by many. However note that potentially patented ECC sources are > removed. Yay! at last! Thanks. Although RFC6090 *really* ought to have solved the nonsense about