753642 - Rip out the codegen for the Xray-related bits of worker bindings, since we don't have Xrays in workers

Status: RESOLVED FIXED

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

Description

[Boris Zbarsky [:bzbarsky]]

If you look at the generated code in XMLHttpRequestBinding.cpp, in the worker part, then CreateInterfaceObjects does this:

  if (sMethods_ids[0] == JSID_VOID &&
      (!InitIds(aCx, sMethods, sMethods_ids) ||
       !InitIds(aCx, sAttributes, sAttributes_ids) ||
       !InitIds(aCx, sConstants, sConstants_ids))) {

but EnumerateProperties does this:

    for (size_t i = 0; i < sizeof(sChromeAttributes_ids); ++i) {
    if ((sChromeAttributes[i].flags & JSPROP_ENUMERATE) &&
        !props.append(sChromeAttributes_ids[i])) {

and similar for ResolveProperty.  Looks pretty fishy.  This happens because CGCreateInterfaceObjectsMethod is doing:

            if props.hasChromeOnly() and not self.descriptor.workers:
                idsToInit.append(props.variableName(True))

Maybe this is all a non-issue because there are no Xrays in workers, but then we just shouldn't generate EnumerateProperties/ResolveProperty for workers, right?

Comment 1

[Boris Zbarsky [:bzbarsky]]

Attachment: Rip out the codegen for the Xray-related bits of worker bindings, since we don't have Xrays in workers.

Ben says no Xrays in workers, so let's just not generate this gunk.

Comment 2

[Boris Zbarsky [:bzbarsky]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/6223b41be383

Comment 3

[Ed Morley [:emorley]]

https://hg.mozilla.org/mozilla-central/rev/6223b41be383