1114064 - Make it possible to have an MCallDOMNative for a jitinfo which is AliasNone

Status: RESOLVED FIXED

(Core :: JavaScript Engine: JIT, defect)

Description

[Boris Zbarsky [:bzbarsky]]

Right now the JIT assumes that the jitinfo for a DOM method (as opposed to getter) cannot be AliasNone.  I'd like to actually create some AliasNone methods, so this assumption needs to be rectified.

Comment 1

[Boris Zbarsky [:bzbarsky]]

Attachment: Support AliasNone DOM methods in ion compilation

Comment 2

[Eric Faust [:efaust]]

Comment on attachment 8539714 [details] [diff] [review]
Support AliasNone DOM methods in ion compilation

Review of attachment 8539714 [details] [diff] [review]:
-----------------------------------------------------------------

I was just thinking about these and noting that we would have to clean this up. r=me

Comment 3

[Jeff Walden [:Waldo]]

Out of curiosity, what methods are these?

Comment 4

[Eric Faust [:efaust]]

(In reply to Jeff Walden [:Waldo] (remove +bmo to email) from comment #3)
> Out of curiosity, what methods are these?

This should be MCallDOMNative::getAliasSet(). We have a system by which the DOM bindings can specify the alias sets of various functions via JitInfo, and we are expanding that system.

Comment 5

[Jeff Walden [:Waldo]]

No, what *DOM* methods.  :-)

Comment 6

[Eric Faust [:efaust]]

Oh. I think mostly performance.now() is one that aliases nothing that we want to allow people to hoist around.

Comment 7

[Boris Zbarsky [:bzbarsky]]

Yeah, for now performance.now().  But in general, methods that guarantee that all observable behavior is async might fall into this bucket, and people are adding more async methods to the platform.

Comment 8

[Boris Zbarsky [:bzbarsky]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/039ae5be340a

Comment 9

[Phil Ringnalda (:philor)]

The fun never stops - with bug 1112934 backed out, ggc started asserting like https://treeherder.mozilla.org/logviewer.html#?job_id=5051574&repo=mozilla-inbound so I backed this out in https://hg.mozilla.org/integration/mozilla-inbound/rev/811c950c8df6

Comment 10

[Phil Ringnalda (:philor)]

Comment on attachment 8539714 [details] [diff] [review]
Support AliasNone DOM methods in ion compilation

Relanded in https://hg.mozilla.org/integration/mozilla-inbound/rev/206763e2c234

Comment 11

[Phil Ringnalda (:philor)]

https://hg.mozilla.org/mozilla-central/rev/206763e2c234