Open mgsloan opened 8 years ago
Also, (stack-mode-type t)
doesn't work for me - it does some insertion and then cond: Wrong type argument: char-or-string-p, nil
.
The piggy back issue of (stack-mode-type t)
is fixed.
Also in this vein, two issues with (stack-mode-type &optional INSERT-VALUE)
where INSERT-VALUE
is true:
(fn :: ...)
instead of inserting the type on the line above.::
and instead just results in the incomplete literal string (fn ::
.Insert type on a top-level function/definition incorrectly uses the inline type syntax: (fn :: ...) instead of inserting the type on the line above.
Are you selecting the word when you use C-u C-c C-t
? Because selecting an expression behaves as though you want to wrap it in a type signature.
Ah yeah that's working. So that solves (1) but (2) still remains. May be related to my other issues regarding the new regressions.
If I ask for the type of an expression, but only select part of it, then that's the only part that appears in the result.
For example, where
|
delineates my selection,put|StrLn "H|ello!"
, I'll getStrLn "H :: IO ()
. This is why the type infos come with spans that say what they refer to.Of course, if the buffer has been edited since the last compilation, then both the query span and results span may be off. Maybe it would be a good idea to have ide-backend directly provide the text of the code that the type info refers to.