Hi,
On 29-08-17 09:54, Hans de Goede wrote:
Hi,
On 28-08-17 23:34, Stephen John Smoogen wrote:
> Can anyone look at these and look for fixes? Thanks
I checked but the 2 i686 machines I've both support SSE2,
making debugging this hard. One could log at build logs and
see if the packages in question inject something into CFLAGS
which they should not ...
An alternative option would be to as the X86 SiG declare that
we are only going to support SSE2 capable hardware.
What I wanted to add here, but forgot is that many people who
do not care about i686 hardware support will like this because
it will allow for faster and/or simpler code for 32 bits
code running on 64 bit machines,be it multi-lib or a 32 bit
vm). So this might be a good compromise vs completely dropping
i686 hardware support.
Regards,
Hans