792137 - Update [TreatNonCallableAsNull] to spec

Status: RESOLVED FIXED

(Core :: DOM: Core & HTML, defect)

Description

[Boris Zbarsky [:bzbarsky]]

It's specified on the callback, not the attribute.

Comment 1

[Boris Zbarsky [:bzbarsky]]

Attachment: [TreatNonCallableAsNull] lives on the callback type now.

Comment 2

[Peter Van der Beken [:peterv]]

Comment on attachment 662266 [details] [diff] [review]
[TreatNonCallableAsNull] lives on the callback type now.

Review of attachment 662266 [details] [diff] [review]:
-----------------------------------------------------------------

::: dom/bindings/Codegen.py
@@ +2335,5 @@
> +            template = (
> +                "if (%s) {\n"
> +                "  ${declName} = &${val}.toObject();\n"
> +                "} else {\n"
> +                "  ${declName} = NULL;\n"

nullptr

@@ +2345,5 @@
> +                "} else {\n"
> +                "%s"
> +                "}" % CGIndenter(onFailureNotCallable(failureCode)).define(),
> +                isDefinitelyObject, type,
> +                "${declName} = NULL",

nullptr

@@ +3785,5 @@
>  
>      if type.isCallback():
> +        if type.nullable():
> +            return CGGeneric("JSObject*")
> +        return CGGeneric("NonNull<JSObject>")

Shouldn't this do the same thing as object (ie JSObject&)?

Comment 3

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 662266 [details] [diff] [review]
[TreatNonCallableAsNull] lives on the callback type now.

> Shouldn't this do the same thing as object (ie JSObject&)?

Ah, this is the accessor signature.  Yes, it should.  Good catch.

Comment 4

[Boris Zbarsky [:bzbarsky]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/c24f40ee97cb

Comment 5

[Ed Morley [:emorley]]

https://hg.mozilla.org/mozilla-central/rev/c24f40ee97cb