Patrick Stevens

Raised https://github.com/dotnet/fsharp/issues/14653 to cover the fact that I've had to do a lot of nonsense to try and associate a better range with the type parameters. That will almost certainly...

A pretty clean implementation, by the way, is https://github.com/Smaug123/fsharp-1/commit/76ea258b45dbea90e2ea28069d76065f1de6c354#diff-99e9d8b56b1879061a50d28265e437772930bfe5c5ba465ede161a735e19adafR185 . This won't work for us as-is, though, because it's very easy for us to end up without meaningful ranges on...

> ```fsharp > JsonSerializer.Deserialize(someString) |> someGenericFunction > ``` Yes, this is recognised as an inference of `obj`. However, it is currently subject to https://github.com/dotnet/fsharp/issues/14653, with the result that a naïve...

Hmm, yes, that sounds plausible. It might be a lot harder! I'll add it to my test cases.

The rot isn't as bad as it looks - I merged https://github.com/dotnet/fsharp/pull/14642 to main a while back, which reserves an error message, so that's most of the conflicts dealt with....

(The most recent two commits on this branch are me blatting its entire contents and replacing it with https://github.com/Smaug123/fsharp-1/commit/76ea258b45dbea90e2ea28069d76065f1de6c354 .)

This is ready for review. It's built off https://github.com/dotnet/fsharp/pull/15301 ; I don't seem to be able to stack PRs across different forks.

When running this on the compiler, it identified all usages of `let __useResumableCode

Raised https://github.com/dotnet/fsharp/issues/15303 to discuss the above.

`member val Foo: obj with get, set` is behaving a bit sadly at the moment.