Re: [arch-dev-public] libgcrypt 1.5.0 needs testing

2011-07-10 Thread Allan McRae
On 09/07/11 17:22, Gaetan Bisson wrote: [2011-07-08 19:08:18 +0200] Andreas Radke: I've switched gnutls now using the default nettle crypto backend instead of the not well upstream supported libgrypt backend. All test suite errors are gone and gnutls should work fine now. Give libgcrypt + gnutl

Re: [arch-dev-public] libgcrypt 1.5.0 needs testing

2011-07-09 Thread Gaetan Bisson
[2011-07-08 19:08:18 +0200] Andreas Radke: > I've switched gnutls now using the default nettle crypto backend > instead of the not well upstream supported libgrypt backend. All test > suite errors are gone and gnutls should work fine now. > > Give libgcrypt + gnutls/nettle good testing. I can giv

Re: [arch-dev-public] libgcrypt 1.5.0 needs testing

2011-07-08 Thread Andreas Radke
> Please give it a good test. I'm going to look into some gnutls test > suite failures now caused by libgcrypt bump. I've switched gnutls now using the default nettle crypto backend instead of the not well upstream supported libgrypt backend. All test suite errors are gone and gnutls should work

[arch-dev-public] libgcrypt 1.5.0 needs testing

2011-07-08 Thread Andreas Radke
I've pushed new libgcrypt to testing. There's no .so bump but other major changes. http://lists.gnupg.org/pipermail/gnupg-announce/2011q2/000307.html I had to disable aes-ni cpu extension support for i686 architecture because it gives segmentation faults in the test suite. Not sure if this is cau