Creating undefined elements in iframes uses prototypes from the calling frame |
||||||
Issue descriptionv55.0.2883.87 is OK, r440736 is broken. What steps will reproduce the problem? 1. Open https://jsfiddle.net/rahmtq3c/1/ What is the expected result? Console contains: proto from iframe? true proto from parent? false What happens instead of that? proto from iframe? false proto from parent? true This means that creating an undefined element in an iframe uses the element interface of the caller, instead of the owner.
,
Dec 27 2016
Slightly clearer repro in https://jsfiddle.net/rahmtq3c/3/
,
Dec 28 2016
Bisected this to https://chromium.googlesource.com/chromium/src/+log/b887e7c4d46340e170ab47c12b3a0d75433250af..21d496a2c607e8348e6a0bc66f655947cffd86fe yukiishino, it looks like 23d2ae42f228825c3edb45ca50c4cb5e2e1045c0 broke wrapping nodes in iframes, PTAL. Please rebaseline imported/wpt/custom-elements when you fix this, there's one test in there that detected this.
,
Dec 28 2016
Oops, un-starting this.
,
Dec 28 2016
Specifically the test results in imported/wpt/custom-elements/upgrading.html will need rebaselining. IIRC there are six failures in there; it should go down to three failures after you fix this. (The "...in a new document..." failures are expected because we do not implement the Document constructor.)
,
Dec 28 2016
,
Dec 28 2016
Looks like custom-elements/parser/parser-uses-registry-of-owner-document.html might have one failure that needs rebaselining after this too.
,
Dec 29 2016
Removing Needs-Bisect label as bisect is already provided.
,
Feb 3 2017
I think that https://crrev.com/2633303002 fixed this issue, too. At least, the repro case at https://jsfiddle.net/rahmtq3c/3/ at #2 is now showing "GOOD". dominicc@, is it okay to close this issue, or are any actions needed?
,
Feb 6 2017
Rebaseline those tests? "imported" has moved to "external" FYI.
,
May 30 2017
As far as I checked, no more work seem needed. I don't see any related entry in LayoutTests/TestExpectations. |
||||||
►
Sign in to add a comment |
||||||
Comment 1 by dominicc@chromium.org
, Dec 27 2016