The range of search values for cdfgam is not consistent with the range of IEEE d
Reported by Michael BAUDIN
Originally assigned to Paul BIGNIER
-- Bug description --
The range of search values for cdfgam is not consistent with the range of IEEE doubles.
The help of cdfgam prints : "Search range: [1E-300,1E300]". This is not consistent with the range of a double, which goes from 0 to 1.e-324 (first denormal) to 1.e-308 (first normal) then 1.e308 (last normal). The value 1.e300 does not correspond to the IEEE standard.
Moreover, the message is wrongly formatted, since the displayed value is zero, while it should be 1.e-300.
-- Scilab error message --
-->shape1 = cdfgam("Shape",shape,p,q,x)
!--error 999
Answer appears to be lower than lowest search bound 0.000000
-- How to reproduce the bug --
shape = 0.5
rate = 1
p = 1.12837916D-150
q = 1
x = 1.e-300
shape1 = cdfgam("Shape",shape,p,q,x)