Skip to content
This repository has been archived by the owner on Aug 17, 2022. It is now read-only.

Q: Using records in function signatures #120

Open
mikevoronov opened this issue Jul 10, 2020 · 3 comments
Open

Q: Using records in function signatures #120

mikevoronov opened this issue Jul 10, 2020 · 3 comments

Comments

@mikevoronov
Copy link

mikevoronov commented Jul 10, 2020

My question is related to Record type. Am I right considering Record as an analog of structs in high-level languages? If so, lets consider simple example of such struct and function that use it:

pub struct S {
    pub a: String,
    pub b: String,
}

fn foo(arg: S) -> S { arg }

How could signature of generated foo look like (keeping in mind, that Wasm supports 5 basic types)? Record.lower unlike string.lower_memory only just unwraps values on a stack. It doesn't write this type in Wasm memory anyhow and so on. So, this signature could look like fn foo(arg_0: String, arg_1: String) -> (String, String) {...}, but this behaviour is hard to obtain on languages like Rust/C++.

@mikevoronov mikevoronov changed the title Q: Passing records in function types Q: Using records in function signatures Jul 10, 2020
@fgmccabe
Copy link
Contributor

not completely sure I understand this question.
there will be an Interface Types analog for record types. It has not been completely nailed down, but the IT type signature for your function will be allowed to mention record types directly:
(type $s (record
($a string)
($b string))
(func foo (param $arg (type $s)) (result (type $s)))
Hope this helps

@jgravelle-google
Copy link
Contributor

Would it be able to use the record types directly from core-wasm, or just at the IT layer?
Because I interpret this question as being about "how do core wasm and IT match up?"

My understanding of that would be that it is going to be ABI-dependent, so the toolchain that generates the function will also be responsible for passing that information up to the IT adapters. But that's just my model of how the tools will interact here.

@mikevoronov
Copy link
Author

mikevoronov commented Jul 13, 2020

@jgravelle-google yes, exactly, thank you for clarifying.

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants