I don't know if this has been previously noted, but
this integral works in commercial macsyma.
The fix in maxima, by interchanging
a and b suggests to me that the problem is in the management of
the varlist, genvars, and the mapping between variables in
the gemeral simplifier and the "gensyms" used in the
rational function package. That may be what you (Wolfgang)
just said.
RJF
>>Another similar problem:
>>
>> integrate( x*%E^(a*x)*SIN(SQRT(b-a^2)*x/2) , x)
>> => quotient is not exact
>>
>>Usually, this sort of problem is solved by changing GCD algorithm
>>(setting the GCD variable), but in this case, all the GCD routines have
>>the same problem.
>>
>>
>
>
>