Neil Mitchell comments

Results 1198 comments of


                                            Neil Mitchell

String literal generic parameter to `Type[]` not allowed

The problem with rechecking it as an annotation is that many that require Usage bindings which we only generate for things we have a suspicion are types. This one will...

Allow global before variable is defined

I found another internal example with this problem. I've sent you the details. Happy to patch if we consider it rare and bad practice.

How to derive UnpackValue for a struct

You can write your own UnpackValue by hand, which shouldn't be too difficult. I do agree that a reasonable UnpackValue implementation for a struct would be to UnpackValue all the...

Go to definition should be able to handle `ctx.attr(s).foo`

I don't have any great ideas here - it seems a really hard problem. The definition is miles away and in a very different format. As such it seems like...

Go to definition should be able to handle `ctx.attr(s).foo`

The LSP was written after the type checker, so there isn't any type information used in the LSP. We should definitely fix that one day.

How to update a mutable attribute?

You will need to use a RefCell for all operations. Assuming you also want to use this for frozen values (e.g. non-mutable), you'll have to have both a frozen version...

Feature request: `typing.Literal`

We don't currently have any plans to refine the type system to follow the latest Python type specification.

Make Control.Arrow.first and second easier to find

Using the new Hoogle at http://hoogle.haskell.org/?hoogle=%3A%3A%20(a-%3Eb)%20-%3E%20(c%2Ca)%20-%3E%20(c%2Cb) this shows up as "No query", which is definitely wrong.