Compilation error in translated code in rosetta fft about pythran

[serge-sans-paille]

The problem is related to recursive typing, as the return type of fft depends on the return type of fft and the return type of the list comprehension.
As the list comprehension is turned into a function call, we reach the limit of current (clumsy) type inference system...

I keep the bug open to remind me that there is a huge amount of work to do there.

The easy fix is to add a complex number in the test case input :-)

[pbrunet]

It was not a duplicate as recent modification in typing solved this issue but Issue #89 is not solved

[serge-sans-paille]

Bug no longer reproducible, hurray, consider it fixed :-)