On 8 February 2012 12:57, Jeroen Demeyer <jdeme...@cage.ugent.be> wrote:
> Does anyone here have a system on which > * building without setting SAGE64 works > * building with SAGE64=yes works > * the result in both cases is actually different > > Yes > The whole SAGE64 thing seems like a huge hack. I tend to agree - the sort of thing that autoconf/automake would have handled better! But it does work. > I'm also afraid that it > isn't tested much. > It adds compiler flags which 1) Are needed for some older versiouns of OS X. 2) Are needed to build Sage on 64-bit Solaris (both x64 and SPARC) So on Solaris, the system I know best, setting SAGE64 =yes results in 64-bit binaries, and not setting it results in 32-bit binaries. So it works, and gives different results - a completely different build. Unfortunately, due to what appears to be a bug with Pynac, importing a lot of python modules, 64-bit binaries on Solaris do not run well. The x86 vesion crashes at startup and the SPARC version is very unstable. It appears Pynac is trying to use something that's no loaded. I don't have time to look for the ticket number as I'm at work. But basically the 64-bit builds on Solaris are stuck until someone sort out that mess. So unless someone comes up with something better, I think SAGE64 should not be altered. 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
