Avoid unnecessary copies of AccessibilityObject::AccessibilityChildrenVector

• Component: New Bugs
• Status: RESOLVED FIXED

[Zan Dobersek]

Avoid unnecessary copies of AccessibilityObject::AccessibilityChildrenVector

[Zan Dobersek]

Avoiding the copies is a huge win on the GTK port. Testing the load of the singe-page HTML Living Standard[1] with WebKitTestRunner, the number of copies is reduced from hundreds of thousands to 75-300.

[1] http://www.whatwg.org/specs/web-apps/current-work/

[Zan Dobersek]

Created attachment 221016 [details]
Patch

[Zan Dobersek]

Committed r161873: <http://trac.webkit.org/changeset/161873>

[chris fleizach]

Comment on attachment 221016 [details]
Patch

I think we probably should use auto in all these cases

[Zan Dobersek]

(In reply to comment #4)
> (From update of attachment 221016 [details])
> I think we probably should use auto in all these cases

Do these cases align with any of the guildelines on using auto that were recently discussed on webkit-dev? AccessibilityObject::children() doesn't specifically note that a reference to the vector of object's children is being returned.

I'd very much like to use auto here, as long as it is used in line with the guidelines.

[chris fleizach]

(In reply to comment #4)
> (From update of attachment 221016 [details])
> I think we probably should use auto in all these cases

https://bugs.webkit.org/show_bug.cgi?id=126915

[chris fleizach]

(In reply to comment #5)
> (In reply to comment #4)
> > (From update of attachment 221016 [details] [details])
> > I think we probably should use auto in all these cases
> 
> Do these cases align with any of the guildelines on using auto that were recently discussed on webkit-dev? AccessibilityObject::children() doesn't specifically note that a reference to the vector of object's children is being returned.
> 
> I'd very much like to use auto here, as long as it is used in line with the guidelines.

After reading that discussion, I feel like this would probably fit the criteria. It removes a lot of disgusting code to write

[Zan Dobersek]

(In reply to comment #7)
> (In reply to comment #5)
> > (In reply to comment #4)
> > > (From update of attachment 221016 [details] [details] [details])
> > > I think we probably should use auto in all these cases
> > 
> > Do these cases align with any of the guildelines on using auto that were recently discussed on webkit-dev? AccessibilityObject::children() doesn't specifically note that a reference to the vector of object's children is being returned.
> > 
> > I'd very much like to use auto here, as long as it is used in line with the guidelines.
> 
> After reading that discussion, I feel like this would probably fit the criteria. It removes a lot of disgusting code to write

Ah, changing to using the range-based for loops in bug #126876 makes things much better-looking, and auto also makes sense in that context.

[chris fleizach]

(In reply to comment #8)
> (In reply to comment #7)
> > (In reply to comment #5)
> > > (In reply to comment #4)
> > > > (From update of attachment 221016 [details] [details] [details] [details])
> > > > I think we probably should use auto in all these cases
> > > 
> > > Do these cases align with any of the guildelines on using auto that were recently discussed on webkit-dev? AccessibilityObject::children() doesn't specifically note that a reference to the vector of object's children is being returned.
> > > 
> > > I'd very much like to use auto here, as long as it is used in line with the guidelines.
> > 
> > After reading that discussion, I feel like this would probably fit the criteria. It removes a lot of disgusting code to write
> 
> Ah, changing to using the range-based for loops in bug #126876 makes things much better-looking, and auto also makes sense in that context.

The question I had is whether a range based loop will ensure that only a reference is used.