Tomas Grosup

Option: Specialize type unification for list/array/sequence builders.

Diagnostics can offer: Explicit annotation or "be first" rule.

1 comment on the gist linked: While this works for a selected example, it is hard to imagine how this would work universally for any kind of DU. Since DUs...

If there is a just one field in a case, yes. As soon as there are 2+, you have to lay them out after each other and therefore need to...

Yes, that would work memory-layout-wise, in exchange for the small addition of 1 more indirection. But indeed, this way it would be workable even without introducing new support for it...

> Is this also going to be backported to 8.0? We compile all our repositories with the .NET 8 SDK due to [this issue](https://github.com/dotnet/fsharp/issues/14313). Can you elaborate please? The issue...

For this bug, it's has not been a regression of the compiler - it was a consequence of many interface layers added to numeric types in NET. The cache has...

Indeed, this is now only implemented for standalone CLI compilation. I will follow up on that PR to progress it to tooling as well. This regression was particularly difficult to...

I am marking this as probably not - the examples given here show that existing F# syntax can express the desired intent. https://github.com/fsharp/fslang-suggestions/issues/1378#issuecomment-2299136473 https://github.com/fsharp/fslang-suggestions/issues/1378#issuecomment-2291487458 Allowing arbitrary expressions after `with let...

I big downside I see is the potential applicability of the optimization. Since it does change evaluation order even in the most basic case ( `data |> List.map f |>...