Update first-class Span speclet with LDM decisions #8241
Conversation
Co-authored-by: Charles Stoner <10732005+cston@users.noreply.github.com>
proposals/first-class-span-types.md
Outdated
| @@ -50,6 +50,9 @@ An implicit span conversion permits `array_types`, `System.Span<T>`, `System.Rea | |||
|
|
|||
| ------ | |||
|
|
|||
| Any Span/ReadOnlySpan types are considered applicable for the conversion if they match by their fully-qualified name. | |||
| If they do not match the types that are selected by the compiler's well-known type resolution logic, a compile-time error is reported. | |||
There was a problem hiding this comment.
This is not what we decided? We won't check that, and we can't really specify it either, since "the compiler's well-known type resolution logic" isn't a thing in the spec.
There was a problem hiding this comment.
Ah, okay, thanks. I thought that was part of the option we discussed, but I might have misunderstood.
There was a problem hiding this comment.
Then the question is how will we emit - do we need to find the operators on the Span types manually? It would be much easier to just use the well-known member machinery; and hence report an error if the Span type doesn't match the well-known type. Maybe that doesn't have to be specified, it can be an implementation detail like well-known types are. I could replace this sentence with something like:
Compiler implementation can report a compile-time error if the Span type from the conversion
does not match the "well-known" Span type that would be selected in other places by implementation-specific logic.
But I guess it might not be so difficult to find the required helpers.
No description provided.