odd bug in integrate: using x(t) instead of x get answer

>>>>> "Richard" == Richard Fateman <fateman at gmail.com> writes:

    Richard> s:  1/sqrt(log(x)+a*x+b)
    Richard> integrate(s,x)   --> not integrated.

    Richard> u:  subst(x(t),x,s);

    Richard> integrate(u,x(t))   --> integrated, but maybe nonsense.

    Richard> The answer looks ok if a=b=0.

    Richard> Also, if a=0, the integral can be done  (correctly).

Interesting.  I haven't looked at what maxima does when integrating u,
but the answer doesn't appear correct, based on a plot with some
random values for a and b.

Ray