Abel Braaksma

> Backporting is too much pain, and I personally would avoid it unless it's something hugely breaking. "backporting" was the wrong term, I actually meant: we should not shy away...

> F# draws a line in the sand and says "no" to "protected" to say "this is not the way you share things in F#, this is not the way...

> I personally think we should just remove it, and have it as analyzer (when the sdk is ready). I'm not entirely against that. However, currently the missing `new` is...

Yes, I think adding `#region` would be beneficial. It will need a [language suggestion](https://github.com/fsharp/fslang-suggestions), though. See this comment as well: https://github.com/fsharp/fslang-suggestions/issues/278#issuecomment-386617042

@dsyme, you say “the new type directed conversions”. But that’s my point here. That doesn’t happen. What did not compile before, now compiles _without warning_. So I’d still say it’s...

> Also example 3 is a spectacularly good example of how bad the Task --> Task subtyping is Indeed it is, @gusty will be delighted to hear that! (We’ve both...

@dsyme is right, `Unit` is always compiled as `null`. While a type exists, this type cannot explicitly be instantiated. For instance, calling any of the methods you dumped above will...

> In what way would it break compatibility? @charlesroddie, at a minimum, every single public function, method, interface that exposes `Unit` will be binary incompatible. [Here's an example](https://sharplab.io/#v2:DYLgZgzgNAJiDUAfA9gBwKYDsAEBlAnhAC7oC2AsAFBH4bYCqmAlkQCq3oAUAlNgLxVsQ7MHRFsYZMmwBDbD2wAjebwDG/QcK2jxAD36zsAKiXZ42VZq1DdVKyLFKZAJxVu+EqdgAsAJhX2OtgAXsFuCh6S0n5u3gDMQA===) where you...

I think this is generally a good suggestion. Another use case may be where people want to create a bunch of tasks and execute them in parallel (this has been...

I think the actual problem here is in the "S" of SRTP. The use of `inline` is just there to statically resolve the parameters during compile time (static). To make...