The following challenge posted by Vladimir yields EXPR COMPLEX FRAC INT as result type. I think it should be EXPR COMPLEX INT. Declaring the variable helps.

(1) -> ex1 := cot(1/2 * (3/8 * %pi * 2^(1/2)-1/2 * 2^(1/2) * atan(3/5)-1/4 * 2^(1/2) * atan(27+73/4 * 2^(1/2))-1/4 * 2^(1/2) * atan(42/41+12/41 * 2^(1/2))+1/4 * 2^(1/2) * atan(144/103+148/103 * 2^(1/2))-1/4 * %i * 2^(1/2) * atanh(-51945/813577+28/2440731 * ((714295812 * (833932+86508 * 93^(1/2))^(1/3)-4881462 * (833932+86508 * 93^(1/2))^(2/3)+3963747144)/(833932+86508 * 93^(1/2))^(1/3))^(1/2)+28/2440731 * 6^(1/2) * ((238098604 * (833932+86508 * 93^(1/2))^(1/3) * ((714295812 * (833932+86508 * 93^(1/2))^(1/3)-4881462 * (833932+86508 * 93^(1/2))^(2/3)+3963747144)/(833932+86508 * 93^(1/2))^(1/3))^(1/2)+813577 * ((714295812 * (833932+86508 * 93^(1/2))^(1/3)-4881462 * (833932+86508 * 93^(1/2))^(2/3)+3963747144)/(833932+86508 * 93^(1/2))^(1/3))^(1/2) * (833932+86508 * 93^(1/2))^(2/3)-660624524 * ((714295812 * (833932+86508 * 93^(1/2))^(1/3)-4881462 * (833932+86508 * 93^(1/2))^(2/3)+3963747144)/(833932+86508 * 93^(1/2))^(1/3))^(1/2)+4852325397096 * (833932+86508 * 93^(1/2))^(1/3))/(833932+86508 * 93^(1/2))^(1/3)/((714295812 * (833932+86508 * 93^(1/2))^(1/3)-4881462 * (833932+86508 * 93^(1/2))^(2/3)+3963747144)/(833932+86508 * 93^(1/2))^(1/3))^(1/2))^(1/2))+1/8 * %i * 2^(1/2) * log(18/7)+1/8 * log(81 * (69/97+2/97 * 2^(1/2))^(%i * 2^(1/2)) * ((26+7 * %i-(10+4 * %i) * 2^(1/2)) * (1/3+1/18 * 6^(1/2) * ((62 * (1052+108 * 93^(1/2))^(1/3)+(1052+108 * 93^(1/2))^(2/3)+28)/(1052+108 * 93^(1/2))^(1/3))^(1/2)+1/18 * ((744 * (1052+108 * 93^(1/2))^(1/3) * ((62 * (1052+108 * 93^(1/2))^(1/3)+(1052+108 * 93^(1/2))^(2/3)+28)/(1052+108 * 93^(1/2))^(1/3))^(1/2)-6 * ((62 * (1052+108 * 93^(1/2))^(1/3)+(1052+108 * 93^(1/2))^(2/3)+28)/(1052+108 * 93^(1/2))^(1/3))^(1/2) * (1052+108 * 93^(1/2))^(2/3)-168 * ((62 * (1052+108 * 93^(1/2))^(1/3)+(1052+108 * 93^(1/2))^(2/3)+28)/(1052+108 * 93^(1/2))^(1/3))^(1/2)+2376 * 6^(1/2) * (1052+108 * 93^(1/2))^(1/3))/(1052+108 * 93^(1/2))^(1/3)/((62 * (1052+108 * 93^(1/2))^(1/3)+(1052+108 * 93^(1/2))^(2/3)+28)/(1052+108 * 93^(1/2))^(1/3))^(1/2))^(1/2))/(3+2^(1/2)))^(2 * %i * 2^(1/2))/((2+3 * %i-2 * %i * 2^(1/2))^(4 * %i * 2^(1/2))))-acoth(2)) * 2^(1/2));

Why do you say EXPR COMPLEX FRAC INT is "forbidden"?

Hm, actually, I'm not quite sure. Usually, domain constructions containing "redundancy", like FRAC FRAC INT or EXPR POLY INT should be avoided, especially in the interpreter. (In SPAD code, they are sometimes unavoidable) Curiously, the interpreter does not reject EXPR FRAC INT, nor EXPR FRAC COMPLEX INT. On the other hand, EXPR COMPLEX FRAC INT is actually different from EXPR COMPLEX INT.

In any case, it seems quite difficult to work with objects of type EXPR COMPLEX FRAC INT. For example, you cannot convert them to EXPR COMPLEX FLOAT...

Severity: normal => wishlist