On Tuesday, January 29, 2013 9:01:28 AM UTC+1, Dima Pasechnik wrote: > > > > On Monday, 28 January 2013 19:05:31 UTC+8, Snark wrote: >> >> Le 28/01/2013 11:55, Dima Pasechnik a �crit : >> > On 2013-01-28, Julien Puydt<julien...@laposte.net> wrote: >> >> Hi, >> >> >> >> here is where we stand: >> >> >> >> - build: ok, even though libm4rie's conversion.c is a pain ; >> >> >> >> - ptestlong: three failing tests, related to libc's gamma function (as >> >> usual... I should find the time to fix that...) ; >> >> >> >> - bdist: available in the bdist/ directory of my home dir (~jpuydt) on >> >> sage.math.washington.edu >> >> >> >> >> >> Snark on #sagemath >> >> >> >> PS: in case someone has access to another ARM box than ubuntu-based, >> >> here is which files give problems ; please report if you don't see >> errors : >> >> sage -t --long -force_lib devel/sage/sage/functions/other.py >> >> sage -t --long -force_lib devel/sage/sage/symbolic/expression.pyx >> >> sage -t --long -force_lib devel/sage/sage/rings/arith.py >> > >> > I wonder why exactly these files also produce numerical noise on >> Cygwin, >> > as I reported: >> > >> http://boxen.math.washington.edu/home/dima/tmp/cygwin_sage-5.7.beta0.test.log >> >> > E.g. >> > File >> > "/usr/local/src/sage/sage-5.7.beta0/devel/sage/sage/rings/arith.py", >> > line 3162: >> > sage: binomial(0.5r, 5) >> > Expected: >> > 0.02734375 >> > Got: >> > 0.02734375000000001 >> > >> > How one does get numerical noise like this on an Intel, is beyond me. >> >> Eh, I was hoping someone would tell me "I have an ARM and I don't see >> the problem" so I could have a look at the libc used and see if that >> helped fixing the issue. >> >> And here you come, with another place where similar problems arise. And >> I notice that the value you "Got:" is the same I see on ARM :-P >> > > This is a eglibc/glibc problem, we reported on > https://bugs.launchpad.net/ubuntu/+source/eglibc/+bug/713985 > Not sure it is related, but please note that lib m does not implement long double (for several or all functions, don't know) on Cygwin.
> > What I cannot recall is why this problem does not show up on (32-bit) > Linux x86. > Note that the code in question (on bug/713985) gives almost the same > output as on ARM > on such a platform, too. > But somehow Sage manages to work around this. > > Dima > > >> Snark on #sagemath >> > -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel?hl=en. For more options, visit https://groups.google.com/groups/opt_out.