Intent to Remove: SVGElement.offsetParent/offsetTop/offsetLeft/offsetWidth/offsetHeight

1026 views
Skip to first unread message

Philip Jägenstedt

unread,
Feb 22, 2016, 10:24:52 PM2/22/16
to blink-dev

Primary eng (and PM) emails

phi...@opera.com


Link to “Intent to Deprecate” thread

https://groups.google.com/a/chromium.org/d/msg/blink-dev/JlAEmQpWMWA/63acYbmdDwAJ


Summary

Remove the attributes offsetParent, offsetTop, offsetLeft, offsetWidth and offsetHeight on SVGElement. These have been deprecated in M48-49, with the deprecation message saying that it will be gone in M50.


Motivation

In Gecko and Edge, these attributes are only supported on HTMLElement. AFAIK, they were never spec'd for Element generally or SVGElement specifically.


Compatibility Risk

There was some pushback on the Intent to Deprecate, based on https://github.com/jquery/jquery/issues/2895. jQuery does not try to support SVG, and the issue was eventually closed. 


AFAICT, people have triggered the deprecation messages by using jQuery's custom :visible/:hidden selectors and position().


For :visible/:hidden, the issue has been fixed in jQuery 2.2.0, and if upgrading isn't feasible then simply overwriting with the new definitions after loading jQuery should work:

jQuery.expr.filters.hidden = function( elem ) {
  return !jQuery.expr.filters.visible( elem );
};
jQuery.expr.filters.visible = function( elem ) {
  return elem.offsetWidth > 0 || elem.offsetHeight > 0 || elem.getClientRects().length > 0;
};


The position() method uses a custom offsetParent() internally, which in turn depends on offsetParent. https://jsfiddle.net/3ktj99m4/4/ illustrates the difference:

Blink/WebKit:
offsetParent(): div
position(): {"top":20,"left":20}


Edge:
offsetParent(): html
position(): {"top":29.5,"left":29.5}


Gecko:
offsetParent(): html
position(): {"top":30,"left":30}


After the change Blink will match Gecko. The "simple" advice here is to do whatever currently works in Edge/Gecko. These jQuery APIs are essentially wrappers for the removed APIs, so trying to polyfill the old behavior does not sound like a good idea.


Nevertheless, one can get the position of an element relative to another by calling getBoundingClientRect() on both and calculating the difference. offsetParent can also be faked to some degree. https://jsfiddle.net/3ktj99m4/6/ sketches these ideas, but this kind of approach ought to be a last resort.


Usage information from UseCounter


OWP launch tracking bug

https://crbug.com/463116


Entry on the feature dashboard

https://www.chromestatus.com/features/5724912467574784


Rick Byers

unread,
Feb 23, 2016, 11:30:42 AM2/23/16
to Philip Jägenstedt, blink-dev
I'm a little worried about the compat risk here, but the fact that Edge (with a WebKit UA string) already successfully has this behavior mitigates the risk substantially.  

LGTM1 to remove.


Dimitri Glazkov

unread,
Feb 23, 2016, 11:31:08 AM2/23/16
to Rick Byers, Philip Jägenstedt, blink-dev
LGTM2.

:DG<

Chris Harrelson

unread,
Feb 23, 2016, 12:07:00 PM2/23/16
to Dimitri Glazkov, Rick Byers, Philip Jägenstedt, blink-dev
LGTM3

--
You received this message because you are subscribed to the Google Groups "blink-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+...@chromium.org.

tapio.ha...@gmail.com

unread,
Apr 24, 2016, 8:45:40 AM4/24/16
to blink-dev

Philip Jägenstedt

unread,
Apr 25, 2016, 5:46:30 AM4/25/16
to Chris Harrelson, Dimitri Glazkov, Rick Byers, blink-dev
There was some fallout from this, with a broken SVG graph on six-swiss-exchange.com. That was also broken in Edge (see screenshot in comment) and would probably also be fixed by using a newer version of dojo, so I closed the issue as WontFix. Given enough breakage of this sort we should reconsider, so if anybody sees more of it, please poke issue 463116 or this thread.

oro.ja...@gmail.com

unread,
Jun 21, 2016, 11:09:04 AM6/21/16
to blink-dev

oro.ja...@gmail.com

unread,
Jun 21, 2016, 11:21:23 AM6/21/16
to blink-dev

Implementation Status

Enabled by default in desktop Chrome 52 (launch bug)

Available in Chrome for Android release 52.

Available in Android WebView release 52.

Consensus & Standardization

  • In development
  • No public signals
  • Shipped in release 39
  • Shipped in release 39
  • No public signals
  • No signals

Owners

Last updated on 2016-05-31


On Tuesday, 23 February 2016 11:24:52 UTC+8, Philip Jägenstedt wrote:

PhistucK

unread,
Jun 21, 2016, 12:11:36 PM6/21/16
to oro.ja...@gmail.com, blink-dev
Stop spamming the group.


PhistucK

naveenbro...@gmail.com

unread,
Jul 14, 2016, 6:32:08 AM7/14/16
to blink-dev
Message has been deleted

wadood...@gmail.com

unread,
Aug 27, 2016, 3:29:27 PM8/27/16
to blink-dev, phi...@opera.com
when i open Facebook .automatically closed and aw loop option come.

wadood...@gmail.com

unread,
Aug 27, 2016, 3:30:36 PM8/27/16
to blink-dev, phi...@opera.com
when i open facebook .automatically closed .aw loop option come.


On Monday, February 22, 2016 at 7:24:52 PM UTC-8, Philip Jägenstedt wrote:

PhistucK

unread,
Aug 27, 2016, 3:51:25 PM8/27/16
to wadood...@gmail.com, blink-dev, Philip Jägenstedt
This is completely off topic. Use the product forums instead.


PhistucK

--
You received this message because you are subscribed to the Google Groups "blink-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+unsubscribe@chromium.org.

madr...@orange.fr

unread,
Mar 13, 2017, 2:02:43 PM3/13/17
to blink-dev, phi...@opera.com

vivosmar...@gmail.com

unread,
Mar 22, 2017, 4:47:34 AM3/22/17
to blink-dev, phi...@opera.com
เมื่อ วันอังคารที่ 23 กุมภาพันธ์ ค.ศ. 2016 10 นาฬิกา 24 นาที 52 วินาที UTC+7, Philip Jägenstedt เขียนว่า:

lordo...@gmail.com

unread,
Mar 23, 2020, 2:33:24 PM3/23/20
to blink-dev
Remove. I am not interested
Reply all
Reply to author
Forward
0 new messages