1311747 - SVGTransform API calls don't trigger mutation observer records

Status: RESOLVED INVALID

(Core :: DOM: Events, defect)

Description

[benneb123]

Attachment: svgMutation.html

User Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_11_6) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/54.0.2840.59 Safari/537.36

Steps to reproduce:

1. Use a MutationObserver to observe an SVG element
2. Use the SVGTransform API to transform the element


Actual results:

The MutationObserver's callback isn't called


Expected results:

The MutationObserver's callback should be passed a MutationRecord describing the svg transformation

Comment 1

[Robert Longson [:longsonr]]

SMIL transforms wouldn't cause MutationObservers to fire as they don't change the DOM, but there's no SMIL in your testcase, should there be?

Comment 2

[benneb123]

Sorry - I meant SVGTransform API calls. They modify the transform attribute of the svg element. So I think the last javascript line in my testcase should set off the MutationObserver because it changes the transform attribute of the rect.

Comment 3

[Adrian Florinescu [:aflorinescu]]

It does seem to me that DOM Events would be the right component for this issue.

Comment 4

[Hsin-Yi Tsai (she/her) [:hsinyi]]

(In reply to benneb123 from comment #2)
> Sorry - I meant SVGTransform API calls. They modify the transform attribute
> of the svg element. So I think the last javascript line in my testcase
> should set off the MutationObserver because it changes the transform
> attribute of the rect.
If I understand the spec correctly, the current behaviour on Firefox 49 looks right to me. That is, MutationObserver callback is called once (not twice as attachment 8802994 [details] expects). Also, when the callback is triggered, it is supplied a list of the changes to the DOM; in this case, the length of the list is two. Anne, did I misunderstand anything? Thanks!

Comment 5

[Anne (:annevk)]

Yeah, there should only be a single callback here and I think a list of two is correct as well, although I'm not sure how well-defined the SVG specification is for its APIs and what those require exactly.

It does seem Chrome has a bug in that it changes the DOM but doesn't queue a mutation record for the transform change.

Comment 6

[Andrew Overholt [:overholt]]

Sounds like this isn't a bug in Gecko.