Unable to get the string_encoding from the c/c++ API headers

[GordonSmith]

Feature

I am trying to implement the "load_string_from_range" ABI function in c++:

def load_string_from_range(cx, ptr, tagged_code_units):
  match cx.opts.string_encoding:
    case 'utf8':
...
    case 'utf16':
...
    case 'latin1+utf16':
...

  return (s, cx.opts.string_encoding, tagged_code_units)

But have been unable to locate anything relating to cx.opts.string_encoding in the c/c++ headers

Benefit

Avoid potential issues dealing with strings

Implementation

Alternatives

Assume string type is "utf8"?

[alexcrichton]

Thanks for the report, but the component model is not supported in the C API yet. So in that sense it's expected that the string encoding isn't availble.

It's worth pointing out though that this string encoding isn't even available in the Rust API, though. This is an internal implementation detail of components which embedders shouldn't in theory need to know about (other than optimizing perhaps)

[GordonSmith]

Thanks for the update - also if the "tagged_code_units" is meant to represent the string length (and not the buffer length) - then the python utf8 case looks wrong to my eye (or is that a question for the spec repo)?

Ref: https://github.com/WebAssembly/component-model/blob/main/design/mvp/canonical-abi/definitions.py#L448C1-L474C57

[alexcrichton]

I think that's a question/issue for the spec repo

[GordonSmith]

Thx... closing.