1319255 - Not exposing HTMLContentElement on Window breaks chrome experiments main page

Status: RESOLVED FIXED

Description

[Boris Zbarsky [:bzbarsky, bz on IRC]]

[Tracking Requested - why for this release]:

[Tracking Requested - why for this release]:

See bug 1316347 comment 6 and following.  

I just verified that the HTMLContentElement change is what causes the problem.  This confuses me quite a bit, because HTMLContentElement doesn't actually _do_ anything that I know of, when it's exposed.  Also, it's not exposed in Safari but the site works fine there...

Comment 1

[Boris Zbarsky [:bzbarsky, bz on IRC]]

Ah, I bet I know what the problem is.  Even when webcomponents are disabled, document.createElement("content") still hands back an element with HTMLContentElement.prototype as its proto.  I expect this is confusing Polymer, because now our internal (non-functional as far as they're concerned) getDistributedNodes() gets called instead of their polyfill.  Or something along those lines.

William, is it expected that even when webcomponents are disabled <content> elements end up with webcomponents stuff on their proto chain?

Comment 2

[Andrew Overholt [:overholt]]

For comment 1.

Comment 3

[Boris Zbarsky [:bzbarsky, bz on IRC]]

Attachment: part 1. Stop doing casts to HTMLShadowElement* simply based on tag

Comment 4

[Boris Zbarsky [:bzbarsky, bz on IRC]]

Attachment: part 2. Stop doing casts to HTMLContentElement* simply based on tag

Comment 5

[Boris Zbarsky [:bzbarsky, bz on IRC]]

Attachment: This fixes the chrome experiments site for me

Comment 6

[Marcia Knous [:marcia - needinfo? me]]

Tracking 53+ since we should render this correctly.

Comment 7

[William Chen [:wchen]]

Comment on attachment 8813400 [details] [diff] [review]
This fixes the chrome experiments site for me

Thanks for writing the patches.

Comment 8

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/mozilla-inbound/rev/1bacb2975324
part 1.  Stop doing casts to HTMLShadowElement* simply based on tag.  r=wchen
https://hg.mozilla.org/integration/mozilla-inbound/rev/93d4161c74bf
part 2.  Stop doing casts to HTMLContentElement* simply based on tag.  r=wchen
https://hg.mozilla.org/integration/mozilla-inbound/rev/03acde1081c3
part 3.  Don't create special elements for <shadow> and <content> unless webcomponents are enabled.  r=wchen

Comment 9

[Boris Zbarsky [:bzbarsky, bz on IRC]]

Comment on attachment 8813397 [details] [diff] [review]
part 1.  Stop doing casts to HTMLShadowElement* simply based on tag

Approval Request Comment
[Feature/regressing bug #]: Bug 1316347
[User impact if declined]: Chrome experiments site broken.  Possibly also other Polymer-using things.
[Describe test coverage new/current, TreeHerder]: Test coverage around this
   stuff is a bit sparse.
[Risks and why]: There are various changes to our content iterator code that might be a bit risky.  It might be safer to back out bug 1316347 on 52 instead.  On the other hand, continuing to ship those properties on Window in ESR 52 is a bit unfortunate.
[String/UUID change made/needed]: None

Comment 10

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/1bacb2975324
https://hg.mozilla.org/mozilla-central/rev/93d4161c74bf
https://hg.mozilla.org/mozilla-central/rev/03acde1081c3

Comment 11

[Julien Cristau [:jcristau]]

Tracking for 52.  Still pondering whether a backout or uplift is the way to go here, leaning towards backout for now...

Comment 13

[Julien Cristau [:jcristau]]

Comment on attachment 8813397 [details] [diff] [review]
part 1.  Stop doing casts to HTMLShadowElement* simply based on tag

fix regression in aurora52

I changed my mind a bit, since this is a long-ish aurora cycle I'm hoping we will have time to discover any fallout from this before 52 goes to beta.

Comment 14

[Julien Cristau [:jcristau]]

Comment on attachment 8813398 [details] [diff] [review]
part 2.  Stop doing casts to HTMLContentElement* simply based on tag

fix regression in aurora52

Comment 15

[Julien Cristau [:jcristau]]

Comment on attachment 8813400 [details] [diff] [review]
This fixes the chrome experiments site for me

fix regression in aurora52

Comment 16

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/releases/mozilla-aurora/rev/c86c1be91253
https://hg.mozilla.org/releases/mozilla-aurora/rev/e053c5265eb2
https://hg.mozilla.org/releases/mozilla-aurora/rev/a07b4a7ba48b