Sorry, somehow I missed Volker's post and your response to it. Its confusing to have things discussed on sage-release and sage-devel simultaneously (one reason I'm opposed to so many sage-* discussion groups...).
If there is already an upstream patch than we should keep supporting linux ppc. I was mistakenly thinking that it might be a long time, if ever, for the numpy folks to correct the problem. -Marshall On Sep 29, 5:17 pm, David Kirkby <david.kir...@onetel.net> wrote: > On 29 September 2010 22:29, Marshall Hampton <hampto...@gmail.com> wrote: > > > Both people wrote back - one is no longer using Sage on linux ppc, the > > other one is. > > > My guess would be that there are fewer than 10 people using Sage on > > linux ppc. I think there are more people who would be glad to have > > updated numpy, scipy, and python versions. Overall my personal > > feeling is that its worthwhile to include the update from 9808 now, > > and if the problems in numpy with ppc get fixed we should update as > > soon as possible after that to try to restore linuc ppc > > functionality. > > > Otherwise maybe we could have this update be an optional spkg. > > You may be right about 10 people - I've no idea of the number. But > like Leif, I think that other platforms provide a useful test when > problems are found. The more unusual operating systems ard platforms > tend to highlight bugs, which lie in wait to hit anyone, but are true > bugs. > > As someone who took a fairly major role in porting Sage to a different > opreating systems (Solaris( and a different processor (SPARC), I'm > somewhat puzzled why PPC should be presenting a particular problem. > I've even built quite large parts of Sage under AIX on PPC. But AIX is > a very different system to Linux, so the Linux issues should not be > hard to solve. > > As there are bug report on the numpy/scipy site I can look at? > > I''ve little intension myself of running Linux on PPC, but I would not > like to see the platform dropped, since 20+ years of software > development have taught me that these rarer platforms often show real > bugs, that can hit anyone at a later date. > > Perhaps even this PPC "bug" is just a bug that could hit any > processor, but has to date only shown up on PPC. > > 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
