950315 - Make more nsDOMEventTargetHelper subclasses use auto-binding constructors

Status: RESOLVED FIXED

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

Description

[Boris Zbarsky [:bzbarsky]]

I was actually doing this because I wanted to hunt down which DETH subclasses were not on WebIDL, but this was a convenient way to do that.

This doesn't convert everything; for one thing on workers we don't BindToOwner.

Comment 1

[Boris Zbarsky [:bzbarsky]]

Attachment: part 1. Convert a bunch of event targets to passing a Window or nsDOMEventTargetHelper directly to the constructor of their ancestor nsDOMEventTargetHelper.

Comment 2

[Boris Zbarsky [:bzbarsky]]

Attachment: part 2. Convert some idb classes to passing a Window or nsDOMEventTargetHelper directly to the constructor of their ancestor nsDOMEventTargetHelper.

Comment 3

[Boris Zbarsky [:bzbarsky]]

Attachment: part 1. Convert a bunch of event targets to passing a Window or nsDOMEventTargetHelper directly to the constructor of their ancestor nsDOMEventTargetHelper.

Comment 4

[Boris Zbarsky [:bzbarsky]]

Attachment: part 2. Convert some idb classes to passing a Window or nsDOMEventTargetHelper directly to the constructor of their ancestor nsDOMEventTargetHelper.

Comment 5

[Boris Zbarsky [:bzbarsky]]

https://tbpl.mozilla.org/?tree=Try&rev=d4db993ff810

Comment 6

[:Ms2ger (he/him; ⌚ UTC+1/+2)]

Comment on attachment 8347572 [details] [diff] [review]
part 1.  Convert a bunch of event targets to passing a Window or nsDOMEventTargetHelper directly to the constructor of their ancestor nsDOMEventTargetHelper.

Review of attachment 8347572 [details] [diff] [review]:
-----------------------------------------------------------------

::: content/base/src/nsDOMDataChannel.cpp
@@ +82,5 @@
>  
> +nsDOMDataChannel::nsDOMDataChannel(already_AddRefed<mozilla::DataChannel> aDataChannel,
> +                                   nsPIDOMWindow* aWindow)
> +  : nsDOMEventTargetHelper(aWindow && aWindow->IsOuterWindow() ?
> +                             aWindow->GetCurrentInnerWindow() : aWindow)

Ugh. Followup to decide which kind of window we get here?

::: dom/devicestorage/nsDeviceStorage.cpp
@@ +2703,5 @@
>    if (NS_FAILED(ds->Init(aWin, aType, storageName))) {
>      *aStore = nullptr;
>      return;
>    }
>    NS_ADDREF(*aStore = ds.get());

Feel free to make this ds.forget(aStore) while you're around :)

Comment 7

[Boris Zbarsky [:bzbarsky]]

Attachment: part 1. Convert a bunch of event targets to passing a Window or nsDOMEventTargetHelper directly to the constructor of their ancestor nsDOMEventTargetHelper.

Comment 8

[Boris Zbarsky [:bzbarsky]]

Attachment: part 1. Convert a bunch of event targets to passing a Window or nsDOMEventTargetHelper directly to the constructor of their ancestor nsDOMEventTargetHelper.

Comment 9

[Boris Zbarsky [:bzbarsky]]

Attachment: part 1. Convert a bunch of event targets to passing a Window or nsDOMEventTargetHelper directly to the constructor of their ancestor nsDOMEventTargetHelper.

Comment 10

[Boris Zbarsky [:bzbarsky]]

Attachment: part 1. Convert a bunch of event targets to passing a Window or nsDOMEventTargetHelper directly to the constructor of their ancestor nsDOMEventTargetHelper.

Comment 11

[Olli Pettay [:smaug][[e-mail]]]

Comment on attachment 8347702 [details] [diff] [review]
part 1.  Convert a bunch of event targets to passing a Window or nsDOMEventTargetHelper directly to the constructor of their ancestor nsDOMEventTargetHelper.

Could you add an assertion to 
nsDOMEventTargetHelper::BindToOwner(nsPIDOMWindow* aOwner) that only inner window is passed there,
well perhaps MOZ_ASSERT(!aOwner || aOwner->IsInnerWindow());


And looks like all the changes are for stuff implementing webidl interfaces so SetIsDOMBinding
in DETH ctor is ok.

Comment 12

[Boris Zbarsky [:bzbarsky]]

> And looks like all the changes are for stuff implementing webidl interfaces

Yes, I carefully avoided the others.

I added the assert; will see how try likes it.

Comment 13

[Boris Zbarsky [:bzbarsky]]

https://tbpl.mozilla.org/?tree=Try&rev=34e356956c0c fwiw, looks good so far.

Comment 14

[Boris Zbarsky [:bzbarsky]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/e9d6fd04d08e
https://hg.mozilla.org/integration/mozilla-inbound/rev/1fc3ac9b4a2c

Comment 15

[Boris Zbarsky [:bzbarsky]]

> Ugh. Followup to decide which kind of window we get here?

Filed bug 957027.

Comment 16

[Ed Morley [:emorley]]

https://hg.mozilla.org/mozilla-central/rev/e9d6fd04d08e
https://hg.mozilla.org/mozilla-central/rev/1fc3ac9b4a2c