Bug 113571 – Implement DOMFutures

• Status: RESOLVED
• Resolution: WONTFIX
• Component: DOM

[Chris Dumez]

DOM 4 is introducing Futures, which offers a nicer way to build asynchronous APIs.
Futures provide a convenient way to get access to the result of an operation.

More details at: http://dom.spec.whatwg.org/#futures

It would be nice to implement them in WebKit.

[Chris Dumez]

Created attachment 195814 [details]
WIP Patch

[Peter Beverloo (cr-android ews)]

Comment on attachment 195814 [details]
WIP Patch

Attachment 195814 [details] did not pass cr-android-ews (chromium-android):
Output: http://webkit-commit-queue.appspot.com/results/17255654

[Early Warning System Bot]

Comment on attachment 195814 [details]
WIP Patch

Attachment 195814 [details] did not pass qt-ews (qt):
Output: http://webkit-commit-queue.appspot.com/results/17311746

[WebKit Review Bot]

Comment on attachment 195814 [details]
WIP Patch

Attachment 195814 [details] did not pass chromium-ews (chromium-xvfb):
Output: http://webkit-commit-queue.appspot.com/results/17257696

[Build Bot]

Comment on attachment 195814 [details]
WIP Patch

Attachment 195814 [details] did not pass mac-wk2-ews (mac-wk2):
Output: http://webkit-commit-queue.appspot.com/results/17327415

[WebKit Review Bot]

Comment on attachment 195814 [details]
WIP Patch

Attachment 195814 [details] did not pass cr-linux-debug-ews (chromium-xvfb):
Output: http://webkit-commit-queue.appspot.com/results/17356136

[Build Bot]

Comment on attachment 195814 [details]
WIP Patch

Attachment 195814 [details] did not pass mac-ews (mac):
Output: http://webkit-commit-queue.appspot.com/results/17327421

[Build Bot]

Comment on attachment 195814 [details]
WIP Patch

Attachment 195814 [details] did not pass mac-ews (mac):
Output: http://webkit-commit-queue.appspot.com/results/17330451

[Early Warning System Bot]

Comment on attachment 195814 [details]
WIP Patch

Attachment 195814 [details] did not pass qt-wk2-ews (qt):
Output: http://webkit-commit-queue.appspot.com/results/17332298

[Chris Dumez]

Created attachment 195860 [details]
WIP Patch

[Adam Barth]

I would like to review this patch before it lands.

[Chris Dumez]

(In reply to comment #11)
> I would like to review this patch before it lands.

Sure. Is it better to send an email to webkit-dev as well to announce that I'm working on this?

[Adam Barth]

> Sure. Is it better to send an email to webkit-dev as well to announce that I'm working on this?

That might be worthwhile.  I'm more interested in the technical aspects of the patch (e.g., the usage of ScriptValue).  It looks like your work is still in progress, so it didn't seem appropriate to review the current patch.

[Chris Dumez]

(In reply to comment #13)
> > Sure. Is it better to send an email to webkit-dev as well to announce that I'm working on this?
> 
> That might be worthwhile.  I'm more interested in the technical aspects of the patch (e.g., the usage of ScriptValue).  It looks like your work is still in progress, so it didn't seem appropriate to review the current patch.

This is useful feedback, thanks. Could you please explain the issue with using ScriptValue? Is there a better way to represent the WebIDL "any" type? I see that it is used quite a lot in the IndexedDB code, which I used as reference.

I tried compiling the patch for Chromium to see if I was missing something related to V8 but the generic code seems to compile fine as ScriptValue also exists for V8 (although with a different API but it does not matter much for my use).

[Adam Barth]

You almost certainly should not be using ScriptValue.  Essentially any time we use ScriptValue, it's a memory leak.  I'd have to study this feature in more detail to figure out the best way of implementing it.

[Chris Dumez]

Created attachment 195972 [details]
WIP Patch

[Chris Dumez]

Promises should now be implemented, not futures.