1221292 - add nondeterministicGetWeakMapKeys to ThreadSafeChromeUtils

Status: RESOLVED FIXED

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

Description

[Tom Tromey :tromey]

On irc :fitzgen suggested adding nondeterministicGetWeakMapKeys to
ThreadSafeChromeUtils as a precursor to bug 1165807.
That way the new functionality could be available in workers as well.

Comment 1

[Tom Tromey :tromey]

Attachment: move nondeterministicGetWeakMapKeys to ThreadSafeChromeUtils

Comment 2

[Tom Tromey :tromey]

Comment on attachment 8686176 [details] [diff] [review]
move nondeterministicGetWeakMapKeys to ThreadSafeChromeUtils

Nick suggested this in bug 1165807, the idea being, IIUC, that we're going
to want access to this information in workers as well.

I did not move the tests out of xpconnect.  It was unclear to me if I should.
(Some of them purely test this function; I suppose moving those might make
sense.  Others use this just for testing the results though.)

I don't know how to test the shumway change but I presume try will do the right thing.  I believe I got all the spots calling this method on Cu.

I'm curious about the need for the Bindings.conf change.  The webidl bindings
doc on MDN says that use of "any" should add a JSContext* argument; but this
seems not to work for static methods.  Should I file a bug for this?  Or is
it intentional?

I removed the "only for tests" comment, since it's clear that hasn't been the
case for a while.

Comment 3

[Nick Fitzgerald [:fitzgen] [⏰PST; UTC-8]]

Comment on attachment 8686176 [details] [diff] [review]
move nondeterministicGetWeakMapKeys to ThreadSafeChromeUtils

Review of attachment 8686176 [details] [diff] [review]:
-----------------------------------------------------------------

\o/ Thanks!

Comment 4

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 8686176 [details] [diff] [review]
move nondeterministicGetWeakMapKeys to ThreadSafeChromeUtils

> The webidl bindings doc on MDN says that use of "any" should add a JSContext*
> argument; but this seems not to work for static methods.

Looks like the documentation didn't get updated after bug 1025973.  I've done that now, but long story short you don't need the implicitJSContext thing and can just use aGlobal.Context().

r=me with that change and HandleValue replaced by Handle<JS::Value> and MutableHandleValue replaced by MutableHandle<JS::Value> and RootedObject replaced by Rooted<JSObject*>.

Comment 5

[Tom Tromey :tromey]

Attachment: move nondeterministicGetWeakMapKeys to ThreadSafeChromeUtils

Updated per review.

Comment 6

[Tom Tromey :tromey]

https://treeherder.mozilla.org/#/jobs?repo=try&revision=244b70ec20d5

Comment 7

[Tom Tromey :tromey]

https://treeherder.mozilla.org/#/jobs?repo=try&revision=a6fa818aedb2

Comment 8

[Pulsebot]

https://hg.mozilla.org/integration/mozilla-inbound/rev/7c22b95c0e29

Comment 9

[Phil Ringnalda (:philor)]

https://hg.mozilla.org/mozilla-central/rev/7c22b95c0e29