type for the optional arguments of 'in-fx/fl/extflvector' #953
Conversation
|
What would be a good test for this? Or how how to test this? I see in the unit-test/typecheck-tests.rkt file. Best to add an extra test over there? |
|
Since you are already adding a new test file, putting it there also makes sense. But really, I don't care where it's tested (though other people might, idk). |
| (check-equal? (for/list : (Listof Fixnum) ([i (in-fxvector (fxvector 1 2 3 4) 3 -1 -2)]) i) | ||
| (list 4.0 2.0)) | ||
|
|
||
| (check-equal? (for/list : (Listof extflonum) ([i (in-extflvector (extflvector 1.0t0 2.0t0 3.0t0 4.0t0) 3 -1 -2)]) i) |
There was a problem hiding this comment.
This test should be behind a (when (extflonum-available?) ...) check.
|
I see this fails for the cs branch, and looking into it, I don't understand why it succeeds on the bc branch. I don't know if typed-racket does something extra but if it is just using the racket implementation then I think the types for the indexes should be: I can change the test, to only count upwards in a for loop, |
|
@samth, I went ahead and changed the type for normalise-inputs to allow going backward through the vector. |
|
Fantastic! I'll take a look and hopefully we can merge soon. |
| (check-equal? (sequence->list (in-fxvector (fxvector 1 2 3 4) 0 #f 2)) (list 1 3)) | ||
| (check-equal? (for/list : (Listof Fixnum) ([i : Fixnum (ann (in-fxvector (fxvector 1 2 3 4) 3 -1 -2) (Sequenceof Fixnum))]) i) | ||
| (list 4 2)) | ||
| ;; I don't understand why this fails... |
There was a problem hiding this comment.
Can you say more about this failure?
There was a problem hiding this comment.
I was hoping to be able to do:
(for/list : (Listof Fixnum) ([i (in-fxvector (fxvector 1 2 3 4) 3 -1 -2)]) i)
but without the annotation of (Sequenceof Fixnum) and Fixnum for i (both are necessary) The resulting list is seen as (Listof Integer)
There was a problem hiding this comment.
I took a quick look at this failure. Seems it has something to do with unsafe-normalise-inputs and normalise-inputs. I might get back to some point later next week
|
I'm going to merge this now; we can try to make that example work with less annotation in the future. |
4 participants
add type for the optional arguments of 'in-fx/fl/extflvector' as in 'in-vector'