Bug 113427 – Web Inspector: properly track main object for proto objects

• Status: RESOLVED
• Resolution: INVALID
• Component: Web Inspector (Deprecated)

[Peter Rybin]

When showing object property, prototype chain can be expanded. If proto object has accessor property, initial object is still needed to calculate property value: getter can be described in prototype object, but initial object must be passed as 'this'.

[Peter Rybin]

Created attachment 195359 [details]
For trybots

[Pavel Feldman]

Comment on attachment 195359 [details]
For trybots

View in context: https://bugs.webkit.org/attachment.cgi?id=195359&action=review

> Source/WebCore/ChangeLog:8
> +        Parameter 'mainObjectId' is passed along from object to object when

What is it going to be used for? In most cases, all TreeElements belong to a single TreeOutline. You could store root object id there.

[Peter Rybin]

(In reply to comment #2)
> (From update of attachment 195359 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=195359&action=review
> 
> > Source/WebCore/ChangeLog:8
> > +        Parameter 'mainObjectId' is passed along from object to object when
> 
> What is it going to be used for? In most cases, all TreeElements belong to a single TreeOutline. You could store root object id there.

All according to bug description.
In your tree you could expand a path:
var1 -> __proto__ -> field1 -> __proto__ -> __proto__ -> *foo

"*foo" represents an accessor property (i.e. with getter and setter). In order to correctly calculate value of this property, we need to call getter and pass the "main object" (not the prototype object) as 'this'. "Main object" is a value of field1. We need to track it everywhere then.

[Yury Semikhatsky]

Comment on attachment 195359 [details]
For trybots

View in context: https://bugs.webkit.org/attachment.cgi?id=195359&action=review

> Source/WebCore/inspector/front-end/ObjectPropertiesSection.js:184
> +    this.mainObjectId = mainObjectId;

main sounds too generic to me but I don't have a better suggestion.

[Pavel Feldman]

Comment on attachment 195359 [details]
For trybots

View in context: https://bugs.webkit.org/attachment.cgi?id=195359&action=review

> Source/WebCore/inspector/front-end/WatchExpressionsSidebarPane.js:343
> +WebInspector.WatchExpressionTreeElement = function(property, mainObjectId)

Sorry, I don't see how Peter's feature is allowed to affect watch expressions and such.

[Yury Semikhatsky]

Comment on attachment 195359 [details]
For trybots

View in context: https://bugs.webkit.org/attachment.cgi?id=195359&action=review

>> Source/WebCore/inspector/front-end/WatchExpressionsSidebarPane.js:343
>> +WebInspector.WatchExpressionTreeElement = function(property, mainObjectId)
> 
> Sorry, I don't see how Peter's feature is allowed to affect watch expressions and such.

That's because the constructor is passed by pointer to ObjectPropertiesSection.prototype.updateProperties

[Peter Rybin]

Created attachment 195679 [details]
Patch

[Peter Rybin]

Comment on attachment 195359 [details]
For trybots

View in context: https://bugs.webkit.org/attachment.cgi?id=195359&action=review

>> Source/WebCore/inspector/front-end/ObjectPropertiesSection.js:184
>> +    this.mainObjectId = mainObjectId;
> 
> main sounds too generic to me but I don't have a better suggestion.

I renamed it to prototypeOwner and added a comment.

[Yury Semikhatsky]

Comment on attachment 195679 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=195679&action=review

> Source/WebCore/inspector/front-end/ObjectPropertiesSection.js:187
> +    // all encountered accessor properties (with getters) are conceptually are still properties

are conceptually are -> are conceptually

[Vsevolod Vlasov]

Comment on attachment 195679 [details]
Patch

I am not sure we really need this according to my comment in https://bugs.webkit.org/show_bug.cgi?id=113357 . Clearing r? for now.