On Mon, May 4, 2009 at 7:52 AM, mabshoff <mabsh...@googlemail.com> wrote: > > > > On May 4, 7:40 am, William Stein <wst...@gmail.com> wrote: >> On Mon, May 4, 2009 at 7:36 AM, mabshoff <mabsh...@googlemail.com> wrote: >> >> > On May 4, 7:30 am, Jaap Spies <j.sp...@hccnet.nl> wrote: >> >> mabshoff wrote: >> >> > Hello folks, >> >> >> > the final release for 3.4.2 is done and sources, the upgrade bits and >> >> > a sage.math binary are in the usual place at >> >> >> > http://sage.math.washington.edu/home/mabshoff/release-cycles-3.4.2/ >> >> >> > I was cocky and labeled the release 3.4.2 instead of 3.4.2.final since >> >> > I am pretty confident we won't need another 3.4.2 release. >> >> >> Except setting testing in FLINT off! >> >> > Well, I knew FLINT still ran its test suite and given we updated MPIR >> > I do prefer for it to run. There was also no 3.4.2 ticket to turn it >> > off :p. >> >> It takes like 2-3 hours to run on some of my build machines, nearly >> doubling the time of building Sage from source.
> > I seriously doubt that. On sage.math the tests run in roughly 10 > minutes. Since building FLINT is more or less instant you should show You're right, it's just over an hour on ATOM: real 62m23.286s user 57m57.825s sys 3m24.089s Successfully installed flint-1.2.4.p1 Now cleaning up tmp files. Making Sage/Python scripts reloc > me the build log, but then the total build time is not going to be in > the 4 to 6 hour range. Running the FLINT test suite flushes out bugs > and I am running it on Solaris/Sparc where this is quite painful and > from experience I would claim that the build on Sparc takes even > longer than on Atom for example. > >> > In Sage 4.0 I will hopefully have a more flexible system for testing >> > so that people who don't want to run the "mandated" testing can turn >> > it off more easily. >> >> What's wrong with the current SAGE_CHECK (or whatever) system where >> spkg-check is run only if a certain environment variable is set? > > That system is all or nothing and since for example R's spkg-check > fails on every platform I do not use it. Something that white- or > blacklists individual spkgs via a control file seems a much saner > mechanism to me. You're right. Just like with "make test", we should change the spkg-check system to simply report the results at the end, instead of stopping if anything isn't perfect. > >> > Originally FLINT 1.2.5 was supposed to go into 3.4.2, but while >> > testing it I saw a doctest failure in some cohomology code and I did >> > not feel like tracking this down since the main change in FLINT 1.2.5 >> > was the update to zn_poly 0.9 (which contained a fix that was supposed >> > to resolve the problem completely). The issue that popped up might >> > also be a padics problem since 2/3 of the doctest failure has been >> > resolved by the new zn_poly, I guess we might find out in Sage 4.0 or >> > 4.0.x. ;) >> >> Good. > > I did list the failure in the FLINT 1.2.5 update ticket if anyone > wants to take a look. > > Cheers, > > Michael > > > -- William Stein Associate Professor of Mathematics University of Washington http://wstein.org --~--~---------~--~----~------------~-------~--~----~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to sage-devel-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://www.sagemath.org -~----------~----~----~----~------~----~------~--~---
