1140435 - Allow JS code to provide an async stack when calling a function

Status: RESOLVED FIXED

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

Description

[:Paolo Amadini]

See bug 1083359 for all the details.

Comment 1

[Boris Zbarsky [:bzbarsky]]

Hmm, I guess this is one of those followup bugs I asked about.  ;)

What are the DOM bits that need doing here?

Comment 2

[:Paolo Amadini]

Attachment: Part 1 - Add the asyncCause and asyncCaller properties to nsIStackFrame

Carrying over r+ from bug 1083359 comment 74.

Comment 3

[:Paolo Amadini]

Attachment: Part 2 - Allow JS code to provide an async stack when calling a function

Carrying over r+ from bug 1083359 comment 75.

Comment 4

[:Paolo Amadini]

These patches are rebased on top of bug 1122238.

Boris, I think copying the new structure of the accessors was straightforward, but if you want to take a look feel free to.

Comment 5

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 8573989 [details] [diff] [review]
Part 1 - Add the asyncCause and asyncCaller properties to nsIStackFrame

You need to add mAsyncCaller to the NS_IMPL_CYCLE_COLLECTION for StackFrame, right?

Comment 6

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 8573989 [details] [diff] [review]
Part 1 - Add the asyncCause and asyncCaller properties to nsIStackFrame

Apart from the CC bit, this looks good.

Comment 7

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 8573991 [details] [diff] [review]
Part 2 - Allow JS code to provide an async stack when calling a function

>+        JS_ReportError(cx, "Invalid argument");

No real need for that, since XPConnect will do the right thing with NS_ERROR_INVALID_ARG.

This looks great.

Comment 8

[:Paolo Amadini]

Thanks, patches updated!

If I leave the following JS_ReportError, would the more specific message be propagated?

  JS_ReportError(cx, "Must use a native JavaScript stack frame");
  return NS_ERROR_INVALID_ARG;

Comment 9

[:Paolo Amadini]

Tryserver build with updated patches:

https://treeherder.mozilla.org/#/jobs?repo=try&revision=18a89cb5764c

Comment 10

[Boris Zbarsky [:bzbarsky]]

> If I leave the following JS_ReportError, would the more specific message be propagated?

Yes, it should be.

Comment 11

[:Paolo Amadini]

https://hg.mozilla.org/integration/fx-team/rev/750120b03c83
https://hg.mozilla.org/integration/fx-team/rev/bdacb3e22b54

Comment 12

[Guilherme Lima]

Phil, something happened on the last fx-team -> mc merge because nothing got marked as fixed (from https://hg.mozilla.org/mozilla-central/rev/65d1053d7959 to https://hg.mozilla.org/mozilla-central/rev/bdacb3e22b54)

Comment 13

[Phil Ringnalda (:philor)]

https://hg.mozilla.org/mozilla-central/rev/750120b03c83
https://hg.mozilla.org/mozilla-central/rev/bdacb3e22b54

Comment 14

[Phil Ringnalda (:philor)]

I'd guess what happened was "I completely forgot to run mcMerge on it," thanks for noticing.

Comment 15

[dattaz]

Attachment: adb logcat on boot -zte open c

Hello,
It's seem (with git bisect) your second patch completely break boot on ZTE OPEN C :(

Comment 16

[:Paolo Amadini]

(In reply to dattaz from comment #15)
> It's seem (with git bisect) your second patch completely break boot on ZTE
> OPEN C :(

Can you file a separate bug on this issue? It should be investigated by someone familiar with Firefox OS.

Comment 17

[:gerard-majax]

See comment 16.

Comment 18

[dattaz]

done : https://bugzilla.mozilla.org/show_bug.cgi?id=1144622