>
> Is there a ticket for this?
>
I just created one --- trac was down when I posted originally:
http://trac.sagemath.org/ticket/20779#ticket
Nathan
--
You received this message because you are subscribed to the Google Groups
"sage-devel" group.
To unsubscribe from this group and stop recei
On Monday, June 6, 2016 at 4:08:06 PM UTC-4, Nathan Dunfield wrote:
>
> I have successfully compiled Sage 7.2, which has an earlier version of
>> NTL, from source on both 10.8 and 10.9.
>>
>
> Checking the logs for my Sage 7.2 builds, the old NTL version 9.6.2 spkg
> did *not* specify the "-ma
Yes that's new. Although it is disabled for fat binaries, but there is
no easy way to disable it from outside spkg-install.
Francois
On 06/07/16 08:08, Nathan Dunfield wrote:
I have successfully compiled Sage 7.2, which has an earlier version
of NTL, from source on both 10.8 and 10.9.
>
> I have successfully compiled Sage 7.2, which has an earlier version of
> NTL, from source on both 10.8 and 10.9.
>
Checking the logs for my Sage 7.2 builds, the old NTL version 9.6.2 spkg
did *not* specify the "-march=native" flag on 10.8 and 10.9, unlike the
current spkg for 9.8.1.
Nath
>
> Did you compile sage yourself on the same machine on which you are trying
> to update the NTL package or are using a binary distribution built on >10.9
> ?
>
I was doing a clean Sage build from scratch against the git commit
(715566f4bf9) corresponding to 7.3.beta3. I have successfully co
Did you compile sage yourself on the same machine on which you are trying
to update the NTL package or are using a binary distribution built on >10.9
?
In the latter case, I am pretty sure it is a different manifestation of the
problem from http://trac.sagemath.org/ticket/20563
On Sunday, June
I can confirm Sage 7.3.beta3 builds on OS X 10.11 (El Capitan)
On Sunday, June 5, 2016 at 6:47:29 PM UTC-7, Nathan Dunfield wrote:
>
> On OS X versions 10.8 (Mountain Lion) and 10.9 (Mavricks), the recently
> updated NTL package fails to build with the following error (complete log
> also attach
On Sunday, June 5, 2016 at 11:49:55 PM UTC-5, leif wrote:
>
> Well, GCC accepting '-march=native' doesn't imply the dumb Apple
> assembler accepts all the instructions (here AVX) GCC then emits...
>
If you pass gcc the flag "-Wa,-q", then it will use clang's assembler
rather than the ancient ve
Francois Bissey wrote:
> What’s more annoying, is that it should be ok to fail here as ntl is trying to
> establish if -march=native is a valid flag.
Well, GCC accepting '-march=native' doesn't imply the dumb Apple
assembler accepts all the instructions (here AVX) GCC then emits...
We've seen suc
