Align HTMLImageElement relevant mutations to spec

Align HTMLImageElement relevant mutations to spec

This CL (almost) aligns the HTMLImageElement's implementation of
relevant mutations with the spec [1] and ensures that the "update the
image data" algorithm fires when it should, and that it doesn't fire
when it shouldn't.

Specific changes:
* Avoid firing update when srcset and sizes are set to current value
* Fire an update when an image is inserted into a <picture>
* Fire a (null) update when an image is removed from a <picture>
* Avoid firing update when a source that's after the <img> was removed,
  inserted or changed

The part where this CL diverges from the spec is upon removal of an
HTMLImageElement from an HTMLPictureElement parent. Instead of running
a full "update the image data", we run it with a null "selected
source" to avoid spurious downloads. See [2] for details.

Compatibility:
* Firefox agrees with the spec on all these points, so this change
  increases compat there.
* The relevant mutations parts we change here are not yet implemented
  in WebKit.
* Latest MS Edge seems to be failing most tests that this CL fixes.

Test:
 Imported from https://github.com/w3c/web-platform-tests/blob/master/html/semantics/embedded-content/the-img-element/relevant-mutations.html

[1] https://html.spec.whatwg.org/multipage/embedded-content.html#relevant-mutations
[2] https://github.com/ResponsiveImagesCG/picture-element/issues/274

BUG=418903
Committed: https://crrev.com/cb0e14ed4b021eff0699175f0fe3ac222ae3978f
Cr-Commit-Position: refs/heads/master@{#364307}

[Yoav Weiss, 2015-12-08 09:42:27]

Description was changed from

==========
Align HTMLImageElement relevant mutations to spec

This CL (almost) aligns the HTMLImageElement's implementation of relevant
mutations
with the spec, and makes sure that the "update the image data" algorithm fires
when
it should and doesn't when it shouldn't.

Specific changes:
* Avoid firing update when srcset and sizes are set to current value.
*

BUG=418903
==========

to

==========
Align HTMLImageElement relevant mutations to spec

This CL (almost) aligns the HTMLImageElement's implementation of relevant
mutations
with the spec, and makes sure that the "update the image data" algorithm fires
when
it should and doesn't when it shouldn't.

Specific changes:
* Avoid firing update when srcset and sizes are set to current value.
* Fire an update when an image is inserted into an HTMLPictureElement
* Fire a (null) update when an image is removed from an HTMLPictureElement
* Avoid firing update when a source that's after the <img> was removed, inserted
or changed


BUG=418903
==========

[Yoav Weiss, 2015-12-08 09:42:27]

yoav@yoav.ws changed reviewers:
+ mkwst@chromium.org, noel@chromium.org

[Yoav Weiss, 2015-12-08 09:45:27]

Hey!

This is a long overdue alignment of HTMLImageElement with the relevant mutations
part of the spec.

Can you please take a look?

[Yoav Weiss, 2015-12-08 09:48:15]

On 2015/12/08 09:45:27, Yoav Weiss wrote:
> Hey!
> 
> This is a long overdue alignment of HTMLImageElement with the relevant
mutations
> part of the spec.
> 
> Can you please take a look?

Most of the files changed are just adding a next node to removedFrom()

Other relevant changes are at ContainerNode, HTMLImageElement, HTMLSourceElement
& HTMLPictureElement.

[Noel Gordon, 2015-12-08 10:42:31]

On 2015/12/08 09:48:15, Yoav Weiss wrote:
> On 2015/12/08 09:45:27, Yoav Weiss wrote:
> > Hey!
> > 
> > This is a long overdue alignment of HTMLImageElement with the relevant
> mutations
> > part of the spec.
> > 
> > Can you please take a look?
> 
> Most of the files changed are just adding a next node to removedFrom()

Yeah, that part seems mechanical, let me look at the rest specific elements you
mentioned ...

[Yoav Weiss, 2015-12-08 10:45:41]

Description was changed from

==========
Align HTMLImageElement relevant mutations to spec

This CL (almost) aligns the HTMLImageElement's implementation of relevant
mutations
with the spec, and makes sure that the "update the image data" algorithm fires
when
it should and doesn't when it shouldn't.

Specific changes:
* Avoid firing update when srcset and sizes are set to current value.
* Fire an update when an image is inserted into an HTMLPictureElement
* Fire a (null) update when an image is removed from an HTMLPictureElement
* Avoid firing update when a source that's after the <img> was removed, inserted
or changed


BUG=418903
==========

to

==========
Align HTMLImageElement relevant mutations to spec

This CL (almost) aligns the HTMLImageElement's implementation of relevant
mutations
with the spec, and makes sure that the "update the image data" algorithm fires
when
it should and doesn't when it shouldn't.

Specific changes:
* Avoid firing update when srcset and sizes are set to current value.
* Fire an update when an image is inserted into an HTMLPictureElement
* Fire a (null) update when an image is removed from an HTMLPictureElement
* Avoid firing update when a source that's after the <img> was removed, inserted
or changed

The part where this CL diverges from the spec is upon removal of
HTMLImageElement from an HTMLPictureElement parent. Instead of running a full
"update the image data", we are running it with a null "selected source", to
avoid spurious downloads. See
https://github.com/ResponsiveImagesCG/picture-element/issues/274 for details.

BUG=418903
==========

[Yoav Weiss, 2015-12-08 10:48:31]

Description was changed from

==========
Align HTMLImageElement relevant mutations to spec

This CL (almost) aligns the HTMLImageElement's implementation of relevant
mutations
with the spec, and makes sure that the "update the image data" algorithm fires
when
it should and doesn't when it shouldn't.

Specific changes:
* Avoid firing update when srcset and sizes are set to current value.
* Fire an update when an image is inserted into an HTMLPictureElement
* Fire a (null) update when an image is removed from an HTMLPictureElement
* Avoid firing update when a source that's after the <img> was removed, inserted
or changed

The part where this CL diverges from the spec is upon removal of
HTMLImageElement from an HTMLPictureElement parent. Instead of running a full
"update the image data", we are running it with a null "selected source", to
avoid spurious downloads. See
https://github.com/ResponsiveImagesCG/picture-element/issues/274 for details.

BUG=418903
==========

to

==========
Align HTMLImageElement relevant mutations to spec

This CL (almost) aligns the HTMLImageElement's implementation of relevant
mutations
with the spec, and makes sure that the "update the image data" algorithm fires
when
it should and doesn't when it shouldn't.

Specific changes:
* Avoid firing update when srcset and sizes are set to current value.
* Fire an update when an image is inserted into an HTMLPictureElement
* Fire a (null) update when an image is removed from an HTMLPictureElement
* Avoid firing update when a source that's after the <img> was removed, inserted
or changed

The part where this CL diverges from the spec is upon removal of
HTMLImageElement from an HTMLPictureElement parent. Instead of running a full
"update the image data", we are running it with a null "selected source", to
avoid spurious downloads. See
https://github.com/ResponsiveImagesCG/picture-element/issues/274 for details.

Compatibility info:
* Firefox agrees with the spec on all these points, so this change will increase
compatibility there.
* The relevant mutations parts we change here are not yet implemented in WebKit.
* Not sure regarding the latest Edge.


BUG=418903
==========

[Noel Gordon, 2015-12-08 10:52:15]

Minor nit: after staring at the classes you listed in detail ...
HTMLImageElement, HTMLSourceElement, HTMLPictureElement and ContainerNode.

https://codereview.chromium.org/1511523002/diff/1/third_party/WebKit/Source/c...
File third_party/WebKit/Source/core/dom/ContainerNode.cpp (right):

https://codereview.chromium.org/1511523002/diff/1/third_party/WebKit/Source/c...
third_party/WebKit/Source/core/dom/ContainerNode.cpp:699: removeBetween(0, next,
*child);
0 -> nullptr?

[Yoav Weiss, 2015-12-08 10:58:42]

On 2015/12/08 10:52:15, noel gordon wrote:
> Minor nit: after staring at the classes you listed in detail ...
> HTMLImageElement, HTMLSourceElement, HTMLPictureElement and ContainerNode.
> 
>
https://codereview.chromium.org/1511523002/diff/1/third_party/WebKit/Source/c...
> File third_party/WebKit/Source/core/dom/ContainerNode.cpp (right):
> 
>
https://codereview.chromium.org/1511523002/diff/1/third_party/WebKit/Source/c...
> third_party/WebKit/Source/core/dom/ContainerNode.cpp:699: removeBetween(0,
next,
> *child);
> 0 -> nullptr?

fixed

[Yoav Weiss, 2015-12-08 11:00:25]

Description was changed from

==========
Align HTMLImageElement relevant mutations to spec

This CL (almost) aligns the HTMLImageElement's implementation of relevant
mutations
with the spec, and makes sure that the "update the image data" algorithm fires
when
it should and doesn't when it shouldn't.

Specific changes:
* Avoid firing update when srcset and sizes are set to current value.
* Fire an update when an image is inserted into an HTMLPictureElement
* Fire a (null) update when an image is removed from an HTMLPictureElement
* Avoid firing update when a source that's after the <img> was removed, inserted
or changed

The part where this CL diverges from the spec is upon removal of
HTMLImageElement from an HTMLPictureElement parent. Instead of running a full
"update the image data", we are running it with a null "selected source", to
avoid spurious downloads. See
https://github.com/ResponsiveImagesCG/picture-element/issues/274 for details.

Compatibility info:
* Firefox agrees with the spec on all these points, so this change will increase
compatibility there.
* The relevant mutations parts we change here are not yet implemented in WebKit.
* Not sure regarding the latest Edge.


BUG=418903
==========

to

==========
Align HTMLImageElement relevant mutations to spec

This CL (almost) aligns the HTMLImageElement's implementation of relevant
mutations
with the spec, and makes sure that the "update the image data" algorithm fires
when
it should and doesn't when it shouldn't.

Specific changes:
* Avoid firing update when srcset and sizes are set to current value.
* Fire an update when an image is inserted into an HTMLPictureElement
* Fire a (null) update when an image is removed from an HTMLPictureElement
* Avoid firing update when a source that's after the <img> was removed, inserted
or changed

The part where this CL diverges from the spec is upon removal of
HTMLImageElement from an HTMLPictureElement parent. Instead of running a full
"update the image data", we are running it with a null "selected source", to
avoid spurious downloads. See
https://github.com/ResponsiveImagesCG/picture-element/issues/274 for details.

Compatibility info:
* Firefox agrees with the spec on all these points, so this change will increase
compatibility there.
* The relevant mutations parts we change here are not yet implemented in WebKit.
* Latest MS Edge seems to be failing most of the tests that this CL fixes.


BUG=418903
==========

[Yoav Weiss, 2015-12-08 11:05:33]

Description was changed from

==========
Align HTMLImageElement relevant mutations to spec

This CL (almost) aligns the HTMLImageElement's implementation of relevant
mutations
with the spec, and makes sure that the "update the image data" algorithm fires
when
it should and doesn't when it shouldn't.

Specific changes:
* Avoid firing update when srcset and sizes are set to current value.
* Fire an update when an image is inserted into an HTMLPictureElement
* Fire a (null) update when an image is removed from an HTMLPictureElement
* Avoid firing update when a source that's after the <img> was removed, inserted
or changed

The part where this CL diverges from the spec is upon removal of
HTMLImageElement from an HTMLPictureElement parent. Instead of running a full
"update the image data", we are running it with a null "selected source", to
avoid spurious downloads. See
https://github.com/ResponsiveImagesCG/picture-element/issues/274 for details.

Compatibility info:
* Firefox agrees with the spec on all these points, so this change will increase
compatibility there.
* The relevant mutations parts we change here are not yet implemented in WebKit.
* Latest MS Edge seems to be failing most of the tests that this CL fixes.


BUG=418903
==========

to

==========
Align HTMLImageElement relevant mutations to spec

This CL (almost) aligns the HTMLImageElement's implementation of
relevant mutations with the spec, and makes sure that the
"update the image data" algorithm fires when it should
and doesn't fire when it shouldn't.

Specific changes:
* Avoid firing update when srcset and sizes are set to current value.
* Fire an update when an image is inserted into an HTMLPictureElement
* Fire a (null) update when an image is removed from an HTMLPictureElement
* Avoid firing update when a source that's after the <img> was removed,
  inserted or changed

The part where this CL diverges from the spec is upon removal of
HTMLImageElement
from an HTMLPictureElement parent.
Instead of running a full "update the image data", we are running it with a null
"selected source", to avoid spurious downloads.
See https://github.com/ResponsiveImagesCG/picture-element/issues/274 for
details.

Compatibility info:
* Firefox agrees with the spec on all these points, so this change increases
compat there.
* The relevant mutations parts we change here are not yet implemented in WebKit.
* Latest MS Edge seems to be failing most of the tests that this CL fixes.

BUG=418903
==========

[Noel Gordon, 2015-12-08 11:20:13]

https://codereview.chromium.org/1511523002/diff/20001/third_party/WebKit/Sour...
File third_party/WebKit/Source/core/html/HTMLPictureElement.h (right):

https://codereview.chromium.org/1511523002/diff/20001/third_party/WebKit/Sour...
third_party/WebKit/Source/core/html/HTMLPictureElement.h:13: 
Maybe not needed?

[Noel Gordon, 2015-12-08 11:25:12]

Description was changed from

==========
Align HTMLImageElement relevant mutations to spec

This CL (almost) aligns the HTMLImageElement's implementation of
relevant mutations with the spec, and makes sure that the
"update the image data" algorithm fires when it should
and doesn't fire when it shouldn't.

Specific changes:
* Avoid firing update when srcset and sizes are set to current value.
* Fire an update when an image is inserted into an HTMLPictureElement
* Fire a (null) update when an image is removed from an HTMLPictureElement
* Avoid firing update when a source that's after the <img> was removed,
  inserted or changed

The part where this CL diverges from the spec is upon removal of
HTMLImageElement
from an HTMLPictureElement parent.
Instead of running a full "update the image data", we are running it with a null
"selected source", to avoid spurious downloads.
See https://github.com/ResponsiveImagesCG/picture-element/issues/274 for
details.

Compatibility info:
* Firefox agrees with the spec on all these points, so this change increases
compat there.
* The relevant mutations parts we change here are not yet implemented in WebKit.
* Latest MS Edge seems to be failing most of the tests that this CL fixes.

BUG=418903
==========

to

==========
Align HTMLImageElement relevant mutations to spec

This CL (almost) aligns the HTMLImageElement's implementation of
relevant mutations with the spec, and ensures that the "update the
image data" algorithm fires when it should, and that it doesn't
fire when it shouldn't.

Specific changes:
* Avoid firing update when srcset and sizes are set to current value.
* Fire an update when an image is inserted into an HTMLPictureElement
* Fire a (null) update when an image is removed from an HTMLPictureElement
* Avoid firing update when a source that's after the <img> was removed,
  inserted or changed

The part where this CL diverges from the spec is upon removal of an
HTMLImageElement from an HTMLPictureElement parent. Instead of running
a full "update the image data", we run it with a null "selected
source" to avoid spurious downloads. See [1] for details.

Compatibility info:
* Firefox agrees with the spec on all these points, so this change
  increases compat there.
* The relevant mutations parts we change here are not yet implemented
  in WebKit.
* Latest MS Edge seems to be failing most tests that this CL fixes.

[1] https://github.com/ResponsiveImagesCG/picture-element/issues/274

BUG=418903
==========

[Noel Gordon, 2015-12-08 11:26:53]

Description was changed from

==========
Align HTMLImageElement relevant mutations to spec

This CL (almost) aligns the HTMLImageElement's implementation of
relevant mutations with the spec, and ensures that the "update the
image data" algorithm fires when it should, and that it doesn't
fire when it shouldn't.

Specific changes:
* Avoid firing update when srcset and sizes are set to current value.
* Fire an update when an image is inserted into an HTMLPictureElement
* Fire a (null) update when an image is removed from an HTMLPictureElement
* Avoid firing update when a source that's after the <img> was removed,
  inserted or changed

The part where this CL diverges from the spec is upon removal of an
HTMLImageElement from an HTMLPictureElement parent. Instead of running
a full "update the image data", we run it with a null "selected
source" to avoid spurious downloads. See [1] for details.

Compatibility info:
* Firefox agrees with the spec on all these points, so this change
  increases compat there.
* The relevant mutations parts we change here are not yet implemented
  in WebKit.
* Latest MS Edge seems to be failing most tests that this CL fixes.

[1] https://github.com/ResponsiveImagesCG/picture-element/issues/274

BUG=418903
==========

to

==========
Align HTMLImageElement relevant mutations to spec

This CL (almost) aligns the HTMLImageElement's implementation of
relevant mutations with the spec [X] and ensures that the "update the
image data" algorithm fires when it should, and that it doesn't
fire when it shouldn't.

Specific changes:
* Avoid firing update when srcset and sizes are set to current value.
* Fire an update when an image is inserted into an HTMLPictureElement
* Fire a (null) update when an image is removed from an HTMLPictureElement
* Avoid firing update when a source that's after the <img> was removed,
  inserted or changed

The part where this CL diverges from the spec is upon removal of an
HTMLImageElement from an HTMLPictureElement parent. Instead of running
a full "update the image data", we run it with a null "selected
source" to avoid spurious downloads. See [2] for details.

Compatibility info:
* Firefox agrees with the spec on all these points, so this change
  increases compat there.
* The relevant mutations parts we change here are not yet implemented
  in WebKit.
* Latest MS Edge seems to be failing most tests that this CL fixes.

[X]
[2] https://github.com/ResponsiveImagesCG/picture-element/issues/274

BUG=418903
==========

[Noel Gordon, 2015-12-08 11:29:33]

Description was changed from

==========
Align HTMLImageElement relevant mutations to spec

This CL (almost) aligns the HTMLImageElement's implementation of
relevant mutations with the spec [X] and ensures that the "update the
image data" algorithm fires when it should, and that it doesn't
fire when it shouldn't.

Specific changes:
* Avoid firing update when srcset and sizes are set to current value.
* Fire an update when an image is inserted into an HTMLPictureElement
* Fire a (null) update when an image is removed from an HTMLPictureElement
* Avoid firing update when a source that's after the <img> was removed,
  inserted or changed

The part where this CL diverges from the spec is upon removal of an
HTMLImageElement from an HTMLPictureElement parent. Instead of running
a full "update the image data", we run it with a null "selected
source" to avoid spurious downloads. See [2] for details.

Compatibility info:
* Firefox agrees with the spec on all these points, so this change
  increases compat there.
* The relevant mutations parts we change here are not yet implemented
  in WebKit.
* Latest MS Edge seems to be failing most tests that this CL fixes.

[X]
[2] https://github.com/ResponsiveImagesCG/picture-element/issues/274

BUG=418903
==========

to

==========
Align HTMLImageElement srcset relevant mutations to spec

This CL (almost) aligns the HTMLImageElement's implementation of
relevant mutations with the spec [X] and ensures that the "update the
image data" algorithm fires when it should, and that it doesn't fire
when it shouldn't.

Specific changes:
* Avoid firing update when srcset and sizes are set to current value
* Fire an update when an image is inserted into a <picture>
* Fire a (null) update when an image is removed from a <picture>
* Avoid firing update when a source that's after the <img> was removed,
  inserted or changed

The part where this CL diverges from the spec is upon removal of an
HTMLImageElement from an HTMLPictureElement parent. Instead of running
a full "update the image data", we run it with a null "selected
source" to avoid spurious downloads. See [2] for details.

Compatibility:
* Firefox agrees with the spec on all these points, so this change
  increases compat there.
* The relevant mutations parts we change here are not yet implemented
  in WebKit.
* Latest MS Edge seems to be failing most tests that this CL fixes.

[X]
[2] https://github.com/ResponsiveImagesCG/picture-element/issues/274

BUG=418903
==========

[Noel Gordon, 2015-12-08 11:33:14]

Description was changed from

==========
Align HTMLImageElement srcset relevant mutations to spec

This CL (almost) aligns the HTMLImageElement's implementation of
relevant mutations with the spec [X] and ensures that the "update the
image data" algorithm fires when it should, and that it doesn't fire
when it shouldn't.

Specific changes:
* Avoid firing update when srcset and sizes are set to current value
* Fire an update when an image is inserted into a <picture>
* Fire a (null) update when an image is removed from a <picture>
* Avoid firing update when a source that's after the <img> was removed,
  inserted or changed

The part where this CL diverges from the spec is upon removal of an
HTMLImageElement from an HTMLPictureElement parent. Instead of running
a full "update the image data", we run it with a null "selected
source" to avoid spurious downloads. See [2] for details.

Compatibility:
* Firefox agrees with the spec on all these points, so this change
  increases compat there.
* The relevant mutations parts we change here are not yet implemented
  in WebKit.
* Latest MS Edge seems to be failing most tests that this CL fixes.

[X]
[2] https://github.com/ResponsiveImagesCG/picture-element/issues/274

BUG=418903
==========

to

==========
Align HTMLImageElement relevant mutations to spec

This CL (almost) aligns the HTMLImageElement's implementation of
relevant mutations with the spec [X] and ensures that the "update the
image data" algorithm fires when it should, and that it doesn't fire
when it shouldn't.

Specific changes:
* Avoid firing update when srcset and sizes are set to current value
* Fire an update when an image is inserted into a <picture>
* Fire a (null) update when an image is removed from a <picture>
* Avoid firing update when a source that's after the <img> was removed,
  inserted or changed

The part where this CL diverges from the spec is upon removal of an
HTMLImageElement from an HTMLPictureElement parent. Instead of running
a full "update the image data", we run it with a null "selected
source" to avoid spurious downloads. See [2] for details.

Compatibility:
* Firefox agrees with the spec on all these points, so this change
  increases compat there.
* The relevant mutations parts we change here are not yet implemented
  in WebKit.
* Latest MS Edge seems to be failing most tests that this CL fixes.

[X]
[2] https://github.com/ResponsiveImagesCG/picture-element/issues/274

BUG=418903
==========

[Noel Gordon, 2015-12-08 11:33:35]

Description was changed from

==========
Align HTMLImageElement relevant mutations to spec

This CL (almost) aligns the HTMLImageElement's implementation of
relevant mutations with the spec [X] and ensures that the "update the
image data" algorithm fires when it should, and that it doesn't fire
when it shouldn't.

Specific changes:
* Avoid firing update when srcset and sizes are set to current value
* Fire an update when an image is inserted into a <picture>
* Fire a (null) update when an image is removed from a <picture>
* Avoid firing update when a source that's after the <img> was removed,
  inserted or changed

The part where this CL diverges from the spec is upon removal of an
HTMLImageElement from an HTMLPictureElement parent. Instead of running
a full "update the image data", we run it with a null "selected
source" to avoid spurious downloads. See [2] for details.

Compatibility:
* Firefox agrees with the spec on all these points, so this change
  increases compat there.
* The relevant mutations parts we change here are not yet implemented
  in WebKit.
* Latest MS Edge seems to be failing most tests that this CL fixes.

[X]
[2] https://github.com/ResponsiveImagesCG/picture-element/issues/274

BUG=418903
==========

to

==========
Align HTMLImageElement relevant mutations to spec

This CL (almost) aligns the HTMLImageElement's implementation of
relevant mutations with the spec [1] and ensures that the "update the
image data" algorithm fires when it should, and that it doesn't fire
when it shouldn't.

Specific changes:
* Avoid firing update when srcset and sizes are set to current value
* Fire an update when an image is inserted into a <picture>
* Fire a (null) update when an image is removed from a <picture>
* Avoid firing update when a source that's after the <img> was removed,
  inserted or changed

The part where this CL diverges from the spec is upon removal of an
HTMLImageElement from an HTMLPictureElement parent. Instead of running
a full "update the image data", we run it with a null "selected
source" to avoid spurious downloads. See [2] for details.

Compatibility:
* Firefox agrees with the spec on all these points, so this change
  increases compat there.
* The relevant mutations parts we change here are not yet implemented
  in WebKit.
* Latest MS Edge seems to be failing most tests that this CL fixes.

[X]
[2] https://github.com/ResponsiveImagesCG/picture-element/issues/274

BUG=418903
==========

[Noel Gordon, 2015-12-08 11:34:01]

Description was changed from

==========
Align HTMLImageElement relevant mutations to spec

This CL (almost) aligns the HTMLImageElement's implementation of
relevant mutations with the spec [1] and ensures that the "update the
image data" algorithm fires when it should, and that it doesn't fire
when it shouldn't.

Specific changes:
* Avoid firing update when srcset and sizes are set to current value
* Fire an update when an image is inserted into a <picture>
* Fire a (null) update when an image is removed from a <picture>
* Avoid firing update when a source that's after the <img> was removed,
  inserted or changed

The part where this CL diverges from the spec is upon removal of an
HTMLImageElement from an HTMLPictureElement parent. Instead of running
a full "update the image data", we run it with a null "selected
source" to avoid spurious downloads. See [2] for details.

Compatibility:
* Firefox agrees with the spec on all these points, so this change
  increases compat there.
* The relevant mutations parts we change here are not yet implemented
  in WebKit.
* Latest MS Edge seems to be failing most tests that this CL fixes.

[X]
[2] https://github.com/ResponsiveImagesCG/picture-element/issues/274

BUG=418903
==========

to

==========
Align HTMLImageElement relevant mutations to spec

This CL (almost) aligns the HTMLImageElement's implementation of
relevant mutations with the spec [1] and ensures that the "update the
image data" algorithm fires when it should, and that it doesn't fire
when it shouldn't.

Specific changes:
* Avoid firing update when srcset and sizes are set to current value
* Fire an update when an image is inserted into a <picture>
* Fire a (null) update when an image is removed from a <picture>
* Avoid firing update when a source that's after the <img> was removed,
  inserted or changed

The part where this CL diverges from the spec is upon removal of an
HTMLImageElement from an HTMLPictureElement parent. Instead of running
a full "update the image data", we run it with a null "selected
source" to avoid spurious downloads. See [2] for details.

Compatibility:
* Firefox agrees with the spec on all these points, so this change
  increases compat there.
* The relevant mutations parts we change here are not yet implemented
  in WebKit.
* Latest MS Edge seems to be failing most tests that this CL fixes.

[1] TODO add a spec reference here.
[2] https://github.com/ResponsiveImagesCG/picture-element/issues/274

BUG=418903
==========

[Noel Gordon, 2015-12-08 11:46:07]

https://codereview.chromium.org/1511523002/diff/20001/third_party/WebKit/Sour...
File third_party/WebKit/Source/core/html/HTMLPictureElement.h (right):

https://codereview.chromium.org/1511523002/diff/20001/third_party/WebKit/Sour...
third_party/WebKit/Source/core/html/HTMLPictureElement.h:13: 
On 2015/12/08 11:20:13, noel gordon wrote:
> Maybe not needed?

The fwd reference class HTMLSourceElement @12 right, to be precise.  Looked over
the rest of the patch in more detail, seems fine.

LGTM and thanks for the compat info in ChangeLog.  Mike will need to look also.

[Yoav Weiss, 2015-12-08 11:46:41]

Description was changed from

==========
Align HTMLImageElement relevant mutations to spec

This CL (almost) aligns the HTMLImageElement's implementation of
relevant mutations with the spec [1] and ensures that the "update the
image data" algorithm fires when it should, and that it doesn't fire
when it shouldn't.

Specific changes:
* Avoid firing update when srcset and sizes are set to current value
* Fire an update when an image is inserted into a <picture>
* Fire a (null) update when an image is removed from a <picture>
* Avoid firing update when a source that's after the <img> was removed,
  inserted or changed

The part where this CL diverges from the spec is upon removal of an
HTMLImageElement from an HTMLPictureElement parent. Instead of running
a full "update the image data", we run it with a null "selected
source" to avoid spurious downloads. See [2] for details.

Compatibility:
* Firefox agrees with the spec on all these points, so this change
  increases compat there.
* The relevant mutations parts we change here are not yet implemented
  in WebKit.
* Latest MS Edge seems to be failing most tests that this CL fixes.

[1] TODO add a spec reference here.
[2] https://github.com/ResponsiveImagesCG/picture-element/issues/274

BUG=418903
==========

to

==========
Align HTMLImageElement relevant mutations to spec

This CL (almost) aligns the HTMLImageElement's implementation of
relevant mutations with the spec [1] and ensures that the "update the
image data" algorithm fires when it should, and that it doesn't fire
when it shouldn't.

Specific changes:
* Avoid firing update when srcset and sizes are set to current value
* Fire an update when an image is inserted into a <picture>
* Fire a (null) update when an image is removed from a <picture>
* Avoid firing update when a source that's after the <img> was removed,
  inserted or changed

The part where this CL diverges from the spec is upon removal of an
HTMLImageElement from an HTMLPictureElement parent. Instead of running
a full "update the image data", we run it with a null "selected
source" to avoid spurious downloads. See [2] for details.

Compatibility:
* Firefox agrees with the spec on all these points, so this change
  increases compat there.
* The relevant mutations parts we change here are not yet implemented
  in WebKit.
* Latest MS Edge seems to be failing most tests that this CL fixes.

[1]
https://html.spec.whatwg.org/multipage/embedded-content.html#relevant-mutations
[2] https://github.com/ResponsiveImagesCG/picture-element/issues/274

BUG=418903
==========

[Noel Gordon, 2015-12-08 11:47:06]

Bots seems unhappy: not sure it's due to this patch though.

[Mike West, 2015-12-08 11:59:54]

On 2015/12/08 at 11:47:06, noel wrote:
> Bots seems unhappy: not sure it's due to this patch though.

The bots are on fire today. Happily I get to be sheriff, so... bleh. The tree is
going to be closed for a while. Sorry in advance.

The test you've added has ~40 "timeout" tests. I hope I'm correct in assuming
that these will end up executing in parallel, so we won't be waiting a minute
for results?

Am I also correct in assuming that you're upstreaming this to
w3c/web-platform-tests?

LGTM if both of those assumptions hold. :)

[dstockwell, 2015-12-08 12:02:37]

sorry, drive-by nits.

https://codereview.chromium.org/1511523002/diff/20001/third_party/WebKit/Sour...
File third_party/WebKit/Source/core/dom/ContainerNode.h (right):

https://codereview.chromium.org/1511523002/diff/20001/third_party/WebKit/Sour...
third_party/WebKit/Source/core/dom/ContainerNode.h:250: void
notifyNodeRemoved(Node&, Node*);
these need parameter names

https://codereview.chromium.org/1511523002/diff/20001/third_party/WebKit/Sour...
File third_party/WebKit/Source/core/dom/Element.h (right):

https://codereview.chromium.org/1511523002/diff/20001/third_party/WebKit/Sour...
third_party/WebKit/Source/core/dom/Element.h:558: void
removedFrom(ContainerNode*, Node*) override;
ditto

[Yoav Weiss, 2015-12-08 12:07:27]

On 2015/12/08 11:59:54, Mike West wrote:
> On 2015/12/08 at 11:47:06, noel wrote:
> > Bots seems unhappy: not sure it's due to this patch though.
> 
> The bots are on fire today. Happily I get to be sheriff, so... bleh. The tree
is
> going to be closed for a while. Sorry in advance.
> 
> The test you've added has ~40 "timeout" tests. I hope I'm correct in assuming
> that these will end up executing in parallel, so we won't be waiting a minute
> for results?

Yeah, they run in parallel, so the test runs fairly fast.

> 
> Am I also correct in assuming that you're upstreaming this to
> w3c/web-platform-tests?

The test was downstreamed from web-platform-tests (I just changed paths so it
would work in our setup). I'll change the description to make it clearer.

> 
> LGTM if both of those assumptions hold. :)

[Noel Gordon, 2015-12-08 12:17:47]

Description was changed from

==========
Align HTMLImageElement relevant mutations to spec

This CL (almost) aligns the HTMLImageElement's implementation of
relevant mutations with the spec [1] and ensures that the "update the
image data" algorithm fires when it should, and that it doesn't fire
when it shouldn't.

Specific changes:
* Avoid firing update when srcset and sizes are set to current value
* Fire an update when an image is inserted into a <picture>
* Fire a (null) update when an image is removed from a <picture>
* Avoid firing update when a source that's after the <img> was removed,
  inserted or changed

The part where this CL diverges from the spec is upon removal of an
HTMLImageElement from an HTMLPictureElement parent. Instead of running
a full "update the image data", we run it with a null "selected
source" to avoid spurious downloads. See [2] for details.

Compatibility:
* Firefox agrees with the spec on all these points, so this change
  increases compat there.
* The relevant mutations parts we change here are not yet implemented
  in WebKit.
* Latest MS Edge seems to be failing most tests that this CL fixes.

[1]
https://html.spec.whatwg.org/multipage/embedded-content.html#relevant-mutations
[2] https://github.com/ResponsiveImagesCG/picture-element/issues/274

BUG=418903
==========

to

==========
Align HTMLImageElement relevant mutations to spec

This CL (almost) aligns the HTMLImageElement's implementation of
relevant mutations with the spec [1] and ensures that the "update the
image data" algorithm fires when it should, and that it doesn't fire
when it shouldn't.

Specific changes:
* Avoid firing update when srcset and sizes are set to current value
* Fire an update when an image is inserted into a <picture>
* Fire a (null) update when an image is removed from a <picture>
* Avoid firing update when a source that's after the <img> was removed,
  inserted or changed

The part where this CL diverges from the spec is upon removal of an
HTMLImageElement from an HTMLPictureElement parent. Instead of running
a full "update the image data", we run it with a null "selected
source" to avoid spurious downloads. See [2] for details.

Compatibility:
* Firefox agrees with the spec on all these points, so this change
  increases compat there.
* The relevant mutations parts we change here are not yet implemented
  in WebKit.
* Latest MS Edge seems to be failing most tests that this CL fixes.

Test:
  TODO: add details about the source of the new test.


[1]
https://html.spec.whatwg.org/multipage/embedded-content.html#relevant-mutations
[2] https://github.com/ResponsiveImagesCG/picture-element/issues/274

BUG=418903
==========

[Noel Gordon, 2015-12-08 12:18:02]

Description was changed from

==========
Align HTMLImageElement relevant mutations to spec

This CL (almost) aligns the HTMLImageElement's implementation of
relevant mutations with the spec [1] and ensures that the "update the
image data" algorithm fires when it should, and that it doesn't fire
when it shouldn't.

Specific changes:
* Avoid firing update when srcset and sizes are set to current value
* Fire an update when an image is inserted into a <picture>
* Fire a (null) update when an image is removed from a <picture>
* Avoid firing update when a source that's after the <img> was removed,
  inserted or changed

The part where this CL diverges from the spec is upon removal of an
HTMLImageElement from an HTMLPictureElement parent. Instead of running
a full "update the image data", we run it with a null "selected
source" to avoid spurious downloads. See [2] for details.

Compatibility:
* Firefox agrees with the spec on all these points, so this change
  increases compat there.
* The relevant mutations parts we change here are not yet implemented
  in WebKit.
* Latest MS Edge seems to be failing most tests that this CL fixes.

Test:
  TODO: add details about the source of the new test.


[1]
https://html.spec.whatwg.org/multipage/embedded-content.html#relevant-mutations
[2] https://github.com/ResponsiveImagesCG/picture-element/issues/274

BUG=418903
==========

to

==========
Align HTMLImageElement relevant mutations to spec

This CL (almost) aligns the HTMLImageElement's implementation of
relevant mutations with the spec [1] and ensures that the "update the
image data" algorithm fires when it should, and that it doesn't fire
when it shouldn't.

Specific changes:
* Avoid firing update when srcset and sizes are set to current value
* Fire an update when an image is inserted into a <picture>
* Fire a (null) update when an image is removed from a <picture>
* Avoid firing update when a source that's after the <img> was removed,
  inserted or changed

The part where this CL diverges from the spec is upon removal of an
HTMLImageElement from an HTMLPictureElement parent. Instead of running
a full "update the image data", we run it with a null "selected
source" to avoid spurious downloads. See [2] for details.

Compatibility:
* Firefox agrees with the spec on all these points, so this change
  increases compat there.
* The relevant mutations parts we change here are not yet implemented
  in WebKit.
* Latest MS Edge seems to be failing most tests that this CL fixes.

Test:
  TODO: add details about the source of the new test.

[1]
https://html.spec.whatwg.org/multipage/embedded-content.html#relevant-mutations
[2] https://github.com/ResponsiveImagesCG/picture-element/issues/274

BUG=418903
==========

[Noel Gordon, 2015-12-08 12:21:58]

Description was changed from

==========
Align HTMLImageElement relevant mutations to spec

This CL (almost) aligns the HTMLImageElement's implementation of
relevant mutations with the spec [1] and ensures that the "update the
image data" algorithm fires when it should, and that it doesn't fire
when it shouldn't.

Specific changes:
* Avoid firing update when srcset and sizes are set to current value
* Fire an update when an image is inserted into a <picture>
* Fire a (null) update when an image is removed from a <picture>
* Avoid firing update when a source that's after the <img> was removed,
  inserted or changed

The part where this CL diverges from the spec is upon removal of an
HTMLImageElement from an HTMLPictureElement parent. Instead of running
a full "update the image data", we run it with a null "selected
source" to avoid spurious downloads. See [2] for details.

Compatibility:
* Firefox agrees with the spec on all these points, so this change
  increases compat there.
* The relevant mutations parts we change here are not yet implemented
  in WebKit.
* Latest MS Edge seems to be failing most tests that this CL fixes.

Test:
  TODO: add details about the source of the new test.

[1]
https://html.spec.whatwg.org/multipage/embedded-content.html#relevant-mutations
[2] https://github.com/ResponsiveImagesCG/picture-element/issues/274

BUG=418903
==========

to

==========
Align HTMLImageElement relevant mutations to spec

This CL (almost) aligns the HTMLImageElement's implementation of
relevant mutations with the spec [1] and ensures that the "update the
image data" algorithm fires when it should, and that it doesn't fire
when it shouldn't.

Specific changes:
* Avoid firing update when srcset and sizes are set to current value
* Fire an update when an image is inserted into a <picture>
* Fire a (null) update when an image is removed from a <picture>
* Avoid firing update when a source that's after the <img> was removed,
  inserted or changed

The part where this CL diverges from the spec is upon removal of an
HTMLImageElement from an HTMLPictureElement parent. Instead of running
a full "update the image data", we run it with a null "selected
source" to avoid spurious downloads. See [2] for details.

Compatibility:
* Firefox agrees with the spec on all these points, so this change
  increases compat there.
* The relevant mutations parts we change here are not yet implemented
  in WebKit.
* Latest MS Edge seems to be failing most tests that this CL fixes.

Test:
  TODO: add details about the new test.

[1]
https://html.spec.whatwg.org/multipage/embedded-content.html#relevant-mutations
[2] https://github.com/ResponsiveImagesCG/picture-element/issues/274

BUG=418903
==========

[Yoav Weiss, 2015-12-08 12:40:52]

Description was changed from

==========
Align HTMLImageElement relevant mutations to spec

This CL (almost) aligns the HTMLImageElement's implementation of
relevant mutations with the spec [1] and ensures that the "update the
image data" algorithm fires when it should, and that it doesn't fire
when it shouldn't.

Specific changes:
* Avoid firing update when srcset and sizes are set to current value
* Fire an update when an image is inserted into a <picture>
* Fire a (null) update when an image is removed from a <picture>
* Avoid firing update when a source that's after the <img> was removed,
  inserted or changed

The part where this CL diverges from the spec is upon removal of an
HTMLImageElement from an HTMLPictureElement parent. Instead of running
a full "update the image data", we run it with a null "selected
source" to avoid spurious downloads. See [2] for details.

Compatibility:
* Firefox agrees with the spec on all these points, so this change
  increases compat there.
* The relevant mutations parts we change here are not yet implemented
  in WebKit.
* Latest MS Edge seems to be failing most tests that this CL fixes.

Test:
  TODO: add details about the new test.

[1]
https://html.spec.whatwg.org/multipage/embedded-content.html#relevant-mutations
[2] https://github.com/ResponsiveImagesCG/picture-element/issues/274

BUG=418903
==========

to

==========
Align HTMLImageElement relevant mutations to spec

This CL (almost) aligns the HTMLImageElement's implementation of
relevant mutations with the spec [1] and ensures that the "update the
image data" algorithm fires when it should, and that it doesn't fire
when it shouldn't.

Specific changes:
* Avoid firing update when srcset and sizes are set to current value
* Fire an update when an image is inserted into a <picture>
* Fire a (null) update when an image is removed from a <picture>
* Avoid firing update when a source that's after the <img> was removed,
  inserted or changed

The part where this CL diverges from the spec is upon removal of an
HTMLImageElement from an HTMLPictureElement parent. Instead of running
a full "update the image data", we run it with a null "selected
source" to avoid spurious downloads. See [2] for details.

Compatibility:
* Firefox agrees with the spec on all these points, so this change
  increases compat there.
* The relevant mutations parts we change here are not yet implemented
  in WebKit.
* Latest MS Edge seems to be failing most tests that this CL fixes.

Test:
 Imported from
https://github.com/w3c/web-platform-tests/blob/master/html/semantics/embedded...

[1]
https://html.spec.whatwg.org/multipage/embedded-content.html#relevant-mutations
[2] https://github.com/ResponsiveImagesCG/picture-element/issues/274

BUG=418903
==========

[Yoav Weiss, 2015-12-08 13:13:21]

On 2015/12/08 12:02:37, dstockwell wrote:
> sorry, drive-by nits.
> 
>
https://codereview.chromium.org/1511523002/diff/20001/third_party/WebKit/Sour...
> File third_party/WebKit/Source/core/dom/ContainerNode.h (right):
> 
>
https://codereview.chromium.org/1511523002/diff/20001/third_party/WebKit/Sour...
> third_party/WebKit/Source/core/dom/ContainerNode.h:250: void
> notifyNodeRemoved(Node&, Node*);
> these need parameter names
> 
>
https://codereview.chromium.org/1511523002/diff/20001/third_party/WebKit/Sour...
> File third_party/WebKit/Source/core/dom/Element.h (right):
> 
>
https://codereview.chromium.org/1511523002/diff/20001/third_party/WebKit/Sour...
> third_party/WebKit/Source/core/dom/Element.h:558: void
> removedFrom(ContainerNode*, Node*) override;
> ditto

Added parameter names

[Yoav Weiss, 2015-12-08 13:14:08]

The CQ bit was checked by yoav@yoav.ws

[Yoav Weiss, 2015-12-08 13:14:08]

The patchset sent to the CQ was uploaded after l-g-t-m from noel@chromium.org,
mkwst@chromium.org
Link to the patchset: https://codereview.chromium.org/1511523002/#ps60001
(title: "Added parameter names")

[commit-bot: I haz the power, 2015-12-08 13:14:24]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1511523002/60001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1511523002/60001

[commit-bot: I haz the power, 2015-12-08 13:35:09]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2015-12-08 13:35:10]

Try jobs failed on following builders:
  chromium_presubmit on tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/chromium_presub...)
  linux_chromium_clobber_rel_ng on tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)

[Yoav Weiss, 2015-12-08 13:37:41]

The CQ bit was checked by yoav@yoav.ws

[commit-bot: I haz the power, 2015-12-08 13:37:59]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1511523002/60001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1511523002/60001

[commit-bot: I haz the power, 2015-12-08 14:01:07]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2015-12-08 14:01:08]

Try jobs failed on following builders:
  linux_chromium_chromeos_rel_ng on tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)

[tkent, 2015-12-08 14:39:15]

tkent@chromium.org changed reviewers:
+ tkent@chromium.org

[tkent, 2015-12-08 14:39:16]

https://codereview.chromium.org/1511523002/diff/60001/third_party/WebKit/Layo...
File
third_party/WebKit/LayoutTests/imported/web-platform-tests/html/semantics/embedded-content/the-img-element/relevant-mutations.html
(right):

https://codereview.chromium.org/1511523002/diff/60001/third_party/WebKit/Layo...
third_party/WebKit/LayoutTests/imported/web-platform-tests/html/semantics/embedded-content/the-img-element/relevant-mutations.html:1:
<!doctype html>
It's not proper way to add a test to LayoutTest/imported/.  Please read
LayoutTests/imported/README.

relevant-mutations.html is explicitly black-listed in
LayoutTests/W3CImportExpectations due to crbug.com/498120.  Fixing that bug only
for /images/ is not hard, I think.

[Yoav Weiss, 2015-12-08 22:29:05]

On 2015/12/08 14:39:16, tkent wrote:
>
https://codereview.chromium.org/1511523002/diff/60001/third_party/WebKit/Layo...
> File
>
third_party/WebKit/LayoutTests/imported/web-platform-tests/html/semantics/embedded-content/the-img-element/relevant-mutations.html
> (right):
> 
>
https://codereview.chromium.org/1511523002/diff/60001/third_party/WebKit/Layo...
>
third_party/WebKit/LayoutTests/imported/web-platform-tests/html/semantics/embedded-content/the-img-element/relevant-mutations.html:1:
> <!doctype html>
> It's not proper way to add a test to LayoutTest/imported/.  Please read
> LayoutTests/imported/README.
> 
> relevant-mutations.html is explicitly black-listed in
> LayoutTests/W3CImportExpectations due to crbug.com/498120.  Fixing that bug
only
> for /images/ is not hard, I think.

Hmm, that seems like an orthogonal issue... Would it be OK if I just add that
test to a separate directory with a TODO that it should be moved to /imported
once that but is fixed (at least for /images)?

[Noel Gordon, 2015-12-09 01:13:00]

On 2015/12/08 13:13:21, Yoav Weiss wrote:
> Added parameter names

Thanks. Let's cq-dry-run to see if the bots have improved.

[Noel Gordon, 2015-12-09 01:13:16]

The CQ bit was checked by noel@chromium.org to run a CQ dry run

[tkent, 2015-12-09 01:13:16]

On 2015/12/08 at 22:29:05, yoav wrote:
> Hmm, that seems like an orthogonal issue... Would it be OK if I just add that
test to a separate directory with a TODO that it should be moved to /imported
once that but is fixed (at least for /images)?

It sounds ok.

[Noel Gordon, 2015-12-09 01:14:15]

The CQ bit was checked by noel@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2015-12-09 01:35:53]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1511523002/60001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1511523002/60001

[commit-bot: I haz the power, 2015-12-09 03:20:20]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2015-12-09 03:20:21]

Dry run: Try jobs failed on following builders:
  win_chromium_rel_ng on tryserver.chromium.win (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.win/builders/win_chromium_rel_...)

[Yoav Weiss, 2015-12-09 08:36:18]

The CQ bit was checked by yoav@yoav.ws to run a CQ dry run

[commit-bot: I haz the power, 2015-12-09 08:38:02]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1511523002/80001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1511523002/80001

[commit-bot: I haz the power, 2015-12-09 10:16:11]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2015-12-09 10:16:11]

Dry run: Try jobs failed on following builders:
  mac_chromium_rel_ng on tryserver.chromium.mac (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.mac/builders/mac_chromium_rel_...)

[Yoav Weiss, 2015-12-09 21:58:16]

Test failures were a function of Node traversal skipping text nodes (so the
sourceOrMdeiaChanged() missed the fact that <source> was before <img> because
the next sibling was a text node).

Fixed it by switching to go over all child nodes.

[Yoav Weiss, 2015-12-09 23:14:36]

The CQ bit was checked by yoav@yoav.ws

[Yoav Weiss, 2015-12-09 23:14:36]

The patchset sent to the CQ was uploaded after l-g-t-m from mkwst@chromium.org,
noel@chromium.org
Link to the patchset: https://codereview.chromium.org/1511523002/#ps100001
(title: "Fix removal issue when a text node is between <source> and <img>")

[commit-bot: I haz the power, 2015-12-09 23:16:23]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1511523002/100001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1511523002/100001

[Noel Gordon, 2015-12-09 23:30:13]

On 2015/12/09 21:58:16, Yoav Weiss wrote:
> Test failures were a function of Node traversal skipping text nodes (so the
> sourceOrMdeiaChanged() missed the fact that <source> was before <img> because
> the next sibling was a text node).

Tests for the win! The HTMLPictureElement::sourceOrMediaChanged fix looks fine.

Still LGTM.

[commit-bot: I haz the power, 2015-12-09 23:42:33]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2015-12-09 23:42:34]

Try jobs failed on following builders:
  linux_android_rel_ng on tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_android_r...)

[Yoav Weiss, 2015-12-10 05:38:21]

The CQ bit was checked by yoav@yoav.ws

[commit-bot: I haz the power, 2015-12-10 05:39:31]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1511523002/100001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1511523002/100001

[commit-bot: I haz the power, 2015-12-10 07:15:57]

Description was changed from

==========
Align HTMLImageElement relevant mutations to spec

This CL (almost) aligns the HTMLImageElement's implementation of
relevant mutations with the spec [1] and ensures that the "update the
image data" algorithm fires when it should, and that it doesn't fire
when it shouldn't.

Specific changes:
* Avoid firing update when srcset and sizes are set to current value
* Fire an update when an image is inserted into a <picture>
* Fire a (null) update when an image is removed from a <picture>
* Avoid firing update when a source that's after the <img> was removed,
  inserted or changed

The part where this CL diverges from the spec is upon removal of an
HTMLImageElement from an HTMLPictureElement parent. Instead of running
a full "update the image data", we run it with a null "selected
source" to avoid spurious downloads. See [2] for details.

Compatibility:
* Firefox agrees with the spec on all these points, so this change
  increases compat there.
* The relevant mutations parts we change here are not yet implemented
  in WebKit.
* Latest MS Edge seems to be failing most tests that this CL fixes.

Test:
 Imported from
https://github.com/w3c/web-platform-tests/blob/master/html/semantics/embedded...

[1]
https://html.spec.whatwg.org/multipage/embedded-content.html#relevant-mutations
[2] https://github.com/ResponsiveImagesCG/picture-element/issues/274

BUG=418903
==========

to

==========
Align HTMLImageElement relevant mutations to spec

This CL (almost) aligns the HTMLImageElement's implementation of
relevant mutations with the spec [1] and ensures that the "update the
image data" algorithm fires when it should, and that it doesn't fire
when it shouldn't.

Specific changes:
* Avoid firing update when srcset and sizes are set to current value
* Fire an update when an image is inserted into a <picture>
* Fire a (null) update when an image is removed from a <picture>
* Avoid firing update when a source that's after the <img> was removed,
  inserted or changed

The part where this CL diverges from the spec is upon removal of an
HTMLImageElement from an HTMLPictureElement parent. Instead of running
a full "update the image data", we run it with a null "selected
source" to avoid spurious downloads. See [2] for details.

Compatibility:
* Firefox agrees with the spec on all these points, so this change
  increases compat there.
* The relevant mutations parts we change here are not yet implemented
  in WebKit.
* Latest MS Edge seems to be failing most tests that this CL fixes.

Test:
 Imported from
https://github.com/w3c/web-platform-tests/blob/master/html/semantics/embedded...

[1]
https://html.spec.whatwg.org/multipage/embedded-content.html#relevant-mutations
[2] https://github.com/ResponsiveImagesCG/picture-element/issues/274

BUG=418903
==========

[commit-bot: I haz the power, 2015-12-10 07:15:59]

Committed patchset #6 (id:100001)

[commit-bot: I haz the power, 2015-12-10 07:16:49]

Description was changed from

==========
Align HTMLImageElement relevant mutations to spec

This CL (almost) aligns the HTMLImageElement's implementation of
relevant mutations with the spec [1] and ensures that the "update the
image data" algorithm fires when it should, and that it doesn't fire
when it shouldn't.

Specific changes:
* Avoid firing update when srcset and sizes are set to current value
* Fire an update when an image is inserted into a <picture>
* Fire a (null) update when an image is removed from a <picture>
* Avoid firing update when a source that's after the <img> was removed,
  inserted or changed

The part where this CL diverges from the spec is upon removal of an
HTMLImageElement from an HTMLPictureElement parent. Instead of running
a full "update the image data", we run it with a null "selected
source" to avoid spurious downloads. See [2] for details.

Compatibility:
* Firefox agrees with the spec on all these points, so this change
  increases compat there.
* The relevant mutations parts we change here are not yet implemented
  in WebKit.
* Latest MS Edge seems to be failing most tests that this CL fixes.

Test:
 Imported from
https://github.com/w3c/web-platform-tests/blob/master/html/semantics/embedded...

[1]
https://html.spec.whatwg.org/multipage/embedded-content.html#relevant-mutations
[2] https://github.com/ResponsiveImagesCG/picture-element/issues/274

BUG=418903
==========

to

==========
Align HTMLImageElement relevant mutations to spec

This CL (almost) aligns the HTMLImageElement's implementation of
relevant mutations with the spec [1] and ensures that the "update the
image data" algorithm fires when it should, and that it doesn't fire
when it shouldn't.

Specific changes:
* Avoid firing update when srcset and sizes are set to current value
* Fire an update when an image is inserted into a <picture>
* Fire a (null) update when an image is removed from a <picture>
* Avoid firing update when a source that's after the <img> was removed,
  inserted or changed

The part where this CL diverges from the spec is upon removal of an
HTMLImageElement from an HTMLPictureElement parent. Instead of running
a full "update the image data", we run it with a null "selected
source" to avoid spurious downloads. See [2] for details.

Compatibility:
* Firefox agrees with the spec on all these points, so this change
  increases compat there.
* The relevant mutations parts we change here are not yet implemented
  in WebKit.
* Latest MS Edge seems to be failing most tests that this CL fixes.

Test:
 Imported from
https://github.com/w3c/web-platform-tests/blob/master/html/semantics/embedded...

[1]
https://html.spec.whatwg.org/multipage/embedded-content.html#relevant-mutations
[2] https://github.com/ResponsiveImagesCG/picture-element/issues/274

BUG=418903

Committed: https://crrev.com/cb0e14ed4b021eff0699175f0fe3ac222ae3978f
Cr-Commit-Position: refs/heads/master@{#364307}
==========

[commit-bot: I haz the power, 2015-12-10 07:16:50]

Patchset 6 (id:??) landed as
https://crrev.com/cb0e14ed4b021eff0699175f0fe3ac222ae3978f
Cr-Commit-Position: refs/heads/master@{#364307}

[Noel Gordon, 2015-12-10 09:36:22]

ʕ•ᴥ•ʔ Happy Koala to you. cr-blink-image bugs -= 1;

[Yoav Weiss, 2015-12-10 10:32:53]

On 2015/12/10 09:36:22, noel gordon wrote:
> ʕ•ᴥ•ʔ Happy Koala to you. cr-blink-image bugs -= 1;

Yay Koalas!

[esprehn, 2015-12-14 22:49:40]

esprehn@chromium.org changed reviewers:
+ esprehn@chromium.org

[esprehn, 2015-12-14 22:49:42]

Please revert the change that plumbs next through every removedFrom call.

https://codereview.chromium.org/1511523002/diff/100001/third_party/WebKit/Sou...
File third_party/WebKit/Source/core/dom/ContainerNode.cpp (right):

https://codereview.chromium.org/1511523002/diff/100001/third_party/WebKit/Sou...
third_party/WebKit/Source/core/dom/ContainerNode.cpp:862: node.removedFrom(this,
node == root ? next : node.nextSibling());
Please revert this change, plumbing next through the whole stack is not correct.
You should just mark the parent <picture> as dirty and async walk the kids. This
puts a branch in the middle of the remove loop just for picture which is not
right.

[esprehn, 2015-12-14 22:52:42]

https://codereview.chromium.org/1511523002/diff/100001/third_party/WebKit/Sou...
File third_party/WebKit/Source/core/html/HTMLPictureElement.cpp (right):

https://codereview.chromium.org/1511523002/diff/100001/third_party/WebKit/Sou...
third_party/WebKit/Source/core/html/HTMLPictureElement.cpp:31:
getChildNodes(*this, potentialSourceNodes);
This copies every kid into a vector, if you're removing every child of a vector
you now copy the kids N times into an on stack vector.

You only need to use getChildNodes() in the cases where your kids will get
mutated out from under you, why do you think that could happen here?

[Yoav Weiss, 2015-12-14 22:59:57]

On 2015/12/14 22:49:42, esprehn wrote:
> Please revert the change that plumbs next through every removedFrom call.
> 
>
https://codereview.chromium.org/1511523002/diff/100001/third_party/WebKit/Sou...
> File third_party/WebKit/Source/core/dom/ContainerNode.cpp (right):
> 
>
https://codereview.chromium.org/1511523002/diff/100001/third_party/WebKit/Sou...
> third_party/WebKit/Source/core/dom/ContainerNode.cpp:862:
node.removedFrom(this,
> node == root ? next : node.nextSibling());
> Please revert this change, plumbing next through the whole stack is not
correct.
> You should just mark the parent <picture> as dirty and async walk the kids.
This
> puts a branch in the middle of the remove loop just for picture which is not
> right.

OK, I'll revert, but I'm not sure how to resolve this without having some
indication of where source was placed in the parent picture.

Basically, I need to figure out if <source> was before or after the encountered
<img>, after <source> was removed. Advice on how to accomplish that would be
appreciated.

[Yoav Weiss, 2015-12-14 23:01:08]

A revert of this CL (patchset #6 id:100001) has been created in
https://codereview.chromium.org/1522253002/ by yoav@yoav.ws.

The reason for reverting is: Reverting due to esprehn's request.