236491 - Make svg use new attribute code

Status: RESOLVED FIXED

(Core :: SVG, defect)

Description

[Jonas Sicking (:sicking) No longer reading bugmail consistently]

The new attributecode is now ready enough to convert SVG to start using it.

The following should be done:
* Make nsSVGElement::GetAttributes just forward to nsGenericElement.
  This means that svg won't have to bother with creating attributenodes or
  implement an nsIDOMNamedNodeMap for attributes at all.
* Store all nsISVGValues in mAttrsAndChildren (the member already exists but is
  only used for children)
* Remove all code in nsSVGAttribute and nsSVGAttributes that implements 
  nsIDOMAttr resp. nsIDOMNamedNodeMap

Once this is done there is probably little reason for the
nsSVGAttribute/nsSVGAttributes classes to still exist. The remaining code and
members in nsSVGAttributes could probably be moved to nsSVGElement, and
nsSVGAttribute will probably more or less just disappear.

Comment 1

[Jonas Sicking (:sicking) No longer reading bugmail consistently]

Attachment: Patch to fix

The patch does the following:

* Store attributes in mAttrsAndChildren.
* Only attributes that are also available throuh dom-properties (such as .r)
are
  stored as nsISVGValues. All other attributes are just stored as strings.
* Use the code from nsGenericElement for a bunch of attribute-code, including:
    .attributes
    attribute-nodes
    HasAttr/GetAttr/UnsetAttr/GetAttrNameAt/GetAttrCount
* Send off nsIDocumentObserver and mutation-events when attributes are modified

  through dom-properties. This way a stylerule like
  circle[r="150"] { fill: red }
  will work when someone does myCircle.r.baseVal.value=150
* Kill nsISVGAttribute.h/nsSVGAttribute.h/nsSVGAttribute.cpp
* Remove code in element dtors that unregistered itself as observer of
  dom-properties. AFAICT in the old code the elements didn't actually observe
  the dom-properties. In the new code add the element as observer in
  nsSVGElement, so i'd like to remove it as observer there too.

Comment 2

[Jonas Sicking (:sicking) No longer reading bugmail consistently]

Comment on attachment 143332 [details] [diff] [review]
Patch to fix

jst: could you r/sr the nsAttrValue.cpp|h part of this patch?

Comment 3

[Johnny Stenback (:jst)]

Comment on attachment 143332 [details] [diff] [review]
Patch to fix

sr=jst for the nsAttrValue part.

Comment 4

[Alex Fritze]

Comment on attachment 143332 [details] [diff] [review]
Patch to fix

great stuff. r=afri

Comment 5

[Jonas Sicking (:sicking) No longer reading bugmail consistently]

checked in. Thanks for the quick reviews