On 9 February 2012 09:10, Jeroen Demeyer <jdeme...@cage.ugent.be> wrote:
> On 2012-02-09 00:46, Dr. David Kirkby wrote: > > But on OS X, this should be easy to test - assuming you can find the > > hardware with a sufficiently old version of OS X > > Apparently this *only* applies to OS X 10.5, so the assumption is not so > easy. > I guess it would apply to older versions too. I would not waste too much of your time over this for a while at least. If t2.math is switched on, it should be possible to build Sage completely as 64-bit. If the build fails, then that's a new bug. But until the Pynac mess is sorted out, you wont be able to run many doc tests. Another issue for 64-bit Solaris x86 (but not SPARC) is that R will not build wtih gcc. The R developers know this, and don't care about it, as R builds fine with the Sun compiler. It seems that the only way to make Sage build on Solaris 64-bit on Intel/AMD hardware is to use both gcc and the Sun compiler. In the past I've just ignored R on 64-bit x86, as its only called as an external program, and so Sage *should* work (except for the R tests). But as I say, there are problems. If you want to test the Sage *build* using SAGE64 you can on "hawk", but you will need to touch the spkg/installed/r-x.y.z file, to fool the build into thinking R has built. If William puts t2.math on, then there's no need to spoof R, and Sage should build. But unless the bug where Pynac is using things that are not initialised, the 64-bit port is stalled. To be honest, despite the fact I put a lot of work into the Solaris build, I'm not actually using Sage. I have my concerns about what I personally consider the rather haphazard way some code gets into Sage. The fact many packages don't have spkg-check files, and few people seem to care, makes me less than enthustiasic about Sage. However, I still feel the Solaris build should be maintained, as it will tend to find bugs that are not seen on the more common platforms. Big-endian problems are an obvious example on SPARC, but Solaris has shown up several issues in the past, which are bad for portability. I think it's true to say an ARM port will be easier as a result of some of the fixes that were needed for Solaris. Dave -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
