Re: [sage-support] series Bug

[D. S. McNeil]

> Sorry for the large example, but smaller examples seem to work.

After some reductions, I think we can find a slightly simpler failure,
of a somewhat suspicious size (4.6.1, OS X 10.6):

q = var('q')
for i in [1..33]:
ix = 2**i
f=(q+1)/(q^(ix) + 1)
res = (f.series(q==0,2), f.subs(q=0), f.taylor(q, 0, 2))
print i, ix, res

1 2 (1 + 1*q + Order(q^2), 1, -q^2 + q + 1)
2 4 (1 + 1*q + Order(q^2), 1, q + 1)
3 8 (1 + 1*q + Order(q^2), 1, q + 1)
4 16 (1 + 1*q + Order(q^2), 1, q + 1)
[...]
28 268435456 (1 + 1*q + Order(q^2), 1, q + 1)
29 536870912 (1 + 1*q + Order(q^2), 1, q + 1)
30 1073741824 (1 + 1*q + Order(q^2), 1, q + 1)
31 2147483648 (Order(q^(-2147483646)), 1, q + 1)
32 4294967296 (1 + 1*q + Order(q^2), 1, q + 1)
33 8589934592 (1 + 1*q + Order(q^2), 1, q + 1)

The case with q^(3^i) is even funnier:

18 387420489 (1 + 1*q + Order(q^2), 1, q + 1)
19 1162261467 (1 + 1*q + Order(q^2), 1, q + 1)
20 3486784401 (Order(q^2), 1, q + 1)
21 10460353203 (1 + 1*q + Order(q^2), 1, q + 1)
22 31381059609 (1 + 1*q + Order(q^2), 1, q + 1)
23 94143178827 (Order(q^2), 1, q + 1)
24 282429536481 (Order(q^2), 1, q + 1)
25 847288609443 (1 + 1*q + Order(q^2), 1, q + 1)
26 2541865828329 (Order(q^2), 1, q + 1)
27 7625597484987 (1 + 1*q + Order(q^2), 1, q + 1)
28 22876792454961 (1 + 1*q + Order(q^2), 1, q + 1)
29 68630377364883 (1 + 1*q + Order(q^2), 1, q + 1)
30 205891132094649 (Order(q^2), 1, q + 1)
31 617673396283947 (1 + 1*q + Order(q^2), 1, q + 1)
32 1853020188851841 (Order(q^2), 1, q + 1)
33 555906056623 (Order(q^2), 1, q + 1)

Obvious suspect is overflow due to use of a C int somewhere along the
path (suddenly negative around 2^31; works for small numbers; is
better-behaved for powers of 2; can generate an actual OverflowError
at large enough exponents.)


Doug

-- 
Department of Earth Sciences
University of Hong Kong

-- 
To post to this group, send email to sage-support@googlegroups.com
To unsubscribe from this group, send email to 
sage-support+unsubscr...@googlegroups.com
For more options, visit this group at 
http://groups.google.com/group/sage-support
URL: http://www.sagemath.org

[sage-support] series Bug

[Clemens Heuberger]

A month ago, I encountered a bug while using series. I used the
"Report a problem" link in the sage notebook to report it and never
got a reply. I couldn't find anything in trac either, so in order to
be sure that the problem is not lost, I repost it here once again. My
apologies if this turns out to be a duplicate ...

So:

var('q')
f=(q^13362120470/((q - 1)*(q^5 - 1)*(q^21 - 1)*(q^85 - 1)*(q^341 -
1)*(q^1365 - 1)*(q^5461 - 1)*(q^21845 - 1)*(q^87381 - 1)*(q^349525 -
1)*(q^1398101 - 1)*(q^5592405 - 1)*(q^22369621 - 1)*(q^89478485 -
1)*(q^357913941 - 1)*(q^1431655765 - 1)*(q^5726623061 - 1)) +
1)/(q^7635497409/((q - 1)*(q^5 - 1)*(q^21 - 1)*(q^85 - 1)*(q^341 -
1)*(q^1365 - 1)*(q^5461 - 1)*(q^21845 - 1)*(q^87381 - 1)*(q^349525 -
1)*(q^1398101 - 1)*(q^5592405 - 1)*(q^22369621 - 1)*(q^89478485 -
1)*(q^357913941 - 1)*(q^1431655765 - 1)*(q^5726623061 - 1)) + 1)
print f.series(q==0,2)
print f.subs(q=0)

f.series(q==0,2) gives Order(q^2). However, f.subs(q=0) gives 1. These
two results contradict each other. In fact, the series result is
false, whereas the subs result is correct.

Sorry for the large example, but smaller examples seem to work.

System information: I am using
ftp://ftp.sun.ac.za/pub/mirrors/www.sagemath.org/linux/64bit/sage-4.6.1-linux-64bit-ubuntu_8.04.4_lts-x86_64-Linux.tar.gz
on debian lenny.

regards,

Clemens Heuberger

-- 
To post to this group, send email to sage-support@googlegroups.com
To unsubscribe from this group, send email to 
sage-support+unsubscr...@googlegroups.com
For more options, visit this group at 
http://groups.google.com/group/sage-support
URL: http://www.sagemath.org