TC gives incorrect type for tuple concat with multiple labels

[PaulKlint]

In commit dd9a74d it was demonstrated that the interpreter did not drop all labels when there are duplicate labels in a tuple concatenation while this is the desired behavior.
The type checker also computes duplicate fields names instead of dropping all field names.

In this example

module experiments::Compiler::Examples::Tst2
import Type;
tuple[int x, str s] a = <10, "abc">;
tuple[real x] b = <3.14>;
value main(list[value] args) = (a + b) has x; // !((a + b) has x);

the type of (a + b) is computed as

tuple([label("x",int()),label("s",str()),label("x",real())])

while it should be

tuple([int(),str(),real()])

[PaulKlint]

Great, this works fine now.