1316619 - Add a [NeedsCallerType] IDL annotation

Status: RESOLVED FIXED

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

Description

[Boris Zbarsky [:bzbarsky]]

See discussion in bug 1316480.

I'm going to build it on top of IsCallerChrome/ThreadsafeIsCallerChrome for now, I guess.  Once bug 1316616 is sorted out we can change to that.

Comment 1

[Boris Zbarsky [:bzbarsky]]

Per IRC discussion, will do a mozilla::dom::CallerType enum.

Comment 2

[Boris Zbarsky [:bzbarsky]]

Attachment: Add a [NeedsCallerType] WebIDL annotation

Comment 3

[Bobby Holley (:bholley)]

Comment on attachment 8809484 [details] [diff] [review]
Add a [NeedsCallerType] WebIDL annotation

Review of attachment 8809484 [details] [diff] [review]:
-----------------------------------------------------------------

::: dom/bindings/Codegen.py
@@ +14085,5 @@
>          # And the nsIPrincipal
>          if self.member.getExtendedAttribute('NeedsSubjectPrincipal'):
> +            # Cheat and assume self.descriptorProvider is a descriptor
> +            if self.descriptorProvider.interface.isExposedInAnyWorker():
> +                args.append(Argument("Maybe<nsIPrincipal*>", "aSubjectPrincipal"))

I don't see the code changes that cause us to set this Maybe<> up when being called from JS (instead of the nsIPrincipal& we had before). And isn't that whole thing kind of orthogonal to your patch here?

::: dom/bindings/test/TestExampleGen.webidl
@@ +806,5 @@
> +interface TestExampleWorkerInterface {
> +  [NeedsSubjectPrincipal] void needsSubjectPrincipalMethod();
> +  [NeedsSubjectPrincipal] attribute boolean needsSubjectPrincipalAttr;
> +  [NeedsCallerType] void needsCallerTypeMethod();
> +  [NeedsCallerType] attribute boolean needsCallerTypeAttr;  

Nit: whitespace.

Comment 4

[Boris Zbarsky [:bzbarsky]]

> And isn't that whole thing kind of orthogonal to your patch here?

It's orthogonal in the sense that the code to pass Maybe<nsIPrincipal*> already exists.  I just encountered this bug in the example codegen when I added a worker-exposed interface that exercised some of this stuff, including my new thing...  I suppose I could have spun this part off to a separate patch, sorry.

Comment 5

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/mozilla-inbound/rev/a0d9aedec307
Add a [NeedsCallerType] WebIDL annotation.  r=bholley

Comment 6

[Boris Zbarsky [:bzbarsky]]

Added docs at https://developer.mozilla.org/en-US/docs/Mozilla/WebIDL_bindings#NeedsCallerType

Comment 7

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/a0d9aedec307