Function pointer type inference #4310
Conversation
|
Done with review pass (commit 2) #Closed |
|
Done with review pass (commit 3) #Closed |
* Add exact inference section. * Correct the upper and lower bound qualifications. No implicit conversions are considered when doing inference, we simply add the correct bounds for the type (upper, lower, or exact).
|
@AlekseyTs I believe I've addressed the feedback. |
|
Done with review pass (commit 5) #Closed |
|
@AlekseyTs addressed feedback. |
|
This is not blocking, but I think we need to bring the proposal to LDM for an approval. |
I think we can do that over email, I'll send one. |
…pper or lower bound inferences.
|
@AlekseyTs made the change we discussed offline. |
| The following case is added to bullet 3: | ||
|
|
||
| > * `V` is a function pointer type `delegate*<V2..Vk, V1>` and there is a function pointer type `delegate*<U2..Uk, U1>` such that `U` is identical to | ||
| `delegate*<U2..Uk, U1>`, and the calling convention of `V` is identical to `U`, and the refness of `Vi` is identical to `Ui`. |
There was a problem hiding this comment.
Why "there is a function pointer type delegate*<U2..Uk, U1> such that U is identical to
delegate*<U2..Uk, U1>", not just "U is a function pointer type delegate*<U2..Uk, U1>"?
There was a problem hiding this comment.
"the calling convention of V is identical to U" is a bit imprecise. Either "V and U have identical calling conventions" or "the calling convention of V is identical to the calling convention of U"
| The following case is added to bullet 3: | ||
|
|
||
| > * `V` is a function pointer type `delegate*<V2..Vk, V1>` and there is a function pointer type `delegate*<U2..Uk, U1>` such that `U` is identical to | ||
| `delegate*<U2..Uk, U1>`, and the calling convention of `V` is identical to `U`, and the refness of `Vi` is identical to `Ui`. |
There was a problem hiding this comment.
"the refness of Vi is identical to Ui" same comment plus make clear that it is each of the Vi, e.g.: "the refness of each Vi is identical to the refness of the corresponding Ui" or something like that.
There was a problem hiding this comment.
Also, is refness a word? I believe we have another term for it, though I would have to dig to find it.
| The following case is added to bullet 2: | ||
|
|
||
| > * `U` is a function pointer type `delegate*<U2..Uk, U1>` and `V` is a function pointer type which is identical to `delegate*<V2..Vk, V1>`, and the | ||
| calling convention of `U` is identical to `V`, and the refness of `Ui` is identical to `Vi`. |
There was a problem hiding this comment.
Same comments as above, mutatis mutandum.
I've taken a stab at creating a spec for function pointer type inference, basing it off the rules for constructed and delegate type inference. I'm unsure that I actually got the wording right that reflects the actual rules I'm implementing in dotnet/roslyn#50249, so careful review is appreciated.