PolyRatFun Division by zero during normalization #336
Description
Related: #178 , possibly #270.
There are still division by zero problems when using PolyRatFun, in both form and tform (#270 seems to be a tform-only problem).
I don't have a minimal example currently, terms which cause the problem (found using `print "%t";) do not cause problems when run alone in a simple script.
I have bisected the problem to a0b635c.
The polynomials here are not large, a crash is obtained for example when the product
prf(1,1 + t)*prf( - 40 - 40*t - 136*t^2 - 40*t^3,9*t^2 + 18*t^3 + 9*t^4)
is produced, with output
prf( - 40 - 40*t - 136*t^2 - 40*t^3,0)
which causes Division by zero during normalization at the following sort.
There are no valgrind errors.
The DEBUG spam for this term is as follows:
*** [57.578] CALL: gcd(0,9*a^3)
*** [57.5781] CALL : poly_ratfun_normalize
*** [57.5781] CALL : poly_ratfun_read
*** [57.5781] CALL: gcd(1,a+1)
*** [57.5781] CALL: integer_content(a+1)
*** [57.5781] RES : integer_content(a+1) = 1
*** [57.5781] CALL: integer_content(1)
*** [57.5781] RES : integer_content(1) = 1
*** [57.5781] CALL: integer_gcd(1,1)
*** [57.5781] RES : integer_gcd(1,1) = 1
*** [57.5781] CALL: gcd(1,a+1)
*** [57.5781] CALL: integer_content(a+1)
*** [57.5781] RES : integer_content(a+1) = 1
*** [57.5781] CALL: integer_content(1)
*** [57.5781] RES : integer_content(1) = 1
*** [57.5781] CALL: integer_gcd(1,1)
*** [57.5781] RES : integer_gcd(1,1) = 1
*** [57.5781] CALL: gcd(1,1)
*** [57.5781] CALL : poly_ratfun_read
*** [57.5781] CALL: gcd(1,9*a^2+18*a^3+9*a^4)
*** [57.5781] CALL: integer_content(9*a^2+18*a^3+9*a^4)
*** [57.5781] RES : integer_content(9*a^2+18*a^3+9*a^4) = 9
*** [57.5781] CALL: integer_content(1)
*** [57.5781] RES : integer_content(1) = 1
*** [57.5781] CALL: integer_gcd(1,9)
*** [57.5782] RES : integer_gcd(1,9) = 1
*** [57.5782] CALL: gcd(-40-40*a-136*a^2-40*a^3,a+1)
*** [57.5782] CALL: integer_content(-40-40*a-136*a^2-40*a^3)
*** [57.5782] RES : integer_content(-40-40*a-136*a^2-40*a^3) = -8
*** [57.5782] CALL: integer_content(a+1)
*** [57.5782] RES : integer_content(a+1) = 1
*** [57.5782] CALL: integer_gcd(-8,1)
*** [57.5782] RES : integer_gcd(-8,1) = 1
*** [57.5782] CALL: gcd_heuristic(5+5*a+17*a^2+5*a^3,a+1,{0})
*** [57.5782] CALL: gcd_heuristic(5,37,{})
*** [57.5782] CALL: integer_content(37)
*** [57.5782] RES : integer_content(37) = 37
*** [57.5782] CALL: integer_gcd(5,37)
*** [57.5782] RES : integer_gcd(5,37) = 1
*** [57.5782] CALL: integer_content(1)
*** [57.5782] RES : integer_content(1) = 1
*** [57.5782] RES : gcd_heuristic(5+5*a+17*a^2+5*a^3,a+1,{0}) = 1
*** [57.5782] CALL: integer_content(1)
*** [57.5782] RES : integer_content(1) = 1
*** [57.5782] RES : gcd(-40-40*a-136*a^2-40*a^3,a+1) = 1
Thanks,
Josh.