Intent to Implement and Ship: Support for CSS image-orientation property

153 views
Skip to first unread message

Stephen Chenney

unread,
Nov 20, 2019, 12:31:28 PM11/20/19
to blink-dev
sche...@chromium.org One-pager Specification: https://drafts.csswg.org/css-images-3/#the-image-orientation None. https://github.com/w3c/csswg-drafts/issues/4164 Enable the CSS image-orientation feature, currently hidden behind a flag. Change the default value to from-image, and apply the property to all images, both content and style. The CSS WG recently decided on the new behavior with a view to moving toward future removal of the property. The new defaults and behavior would match that if the property did not exist at all.
Safari: Shipped Mozilla: Support and indications of shipping after we do Edge/IE: Never implemented Firefox: Public support Edge: No public signals Safari: Shipped Web developers: Positive (https://bugs.chromium.org/p/chromium/issues/detail?id=158753) 153 stars for implementation of the feature, with many passionate comments. None. Trials via cluster telemetry indicate possible impact for 56/100000 sites. These sites, and others, may have content change when the feature is shipped with new defaults. https://chromestatus.com/metrics/css/popularity shows less than 0.000001% usage of the property, but this is not a good indicator of impact due the proposed default change (we might expect usage to go up as developers use image-orientation: none top maintain the current appearance of their site). None.
No additional support required. Yes No Existing WPT tests will be used as is.Some additional tests will be implemented and existing web_tests will be moved to WPT. https://bugs.chromium.org/p/chromium/issues/detail?id=158753 https://www.chromestatus.com/feature/6313474512650240
This intent message was generated by Chrome Platform Status.

Stephen Chenney

unread,
Nov 20, 2019, 2:55:02 PM11/20/19
to blink-dev
Minor update RE Tag Review. Inline.

On Wed, Nov 20, 2019 at 12:31 PM Stephen Chenney <sche...@chromium.org> wrote:
sche...@chromium.org One-pager Specification: https://drafts.csswg.org/css-images-3/#the-image-orientation No design doc (implementation already exists behind flag). Tag review has not been requested because the property has long been in the spec and has seen extensive recent discussion (leading Safari and ourselves to update the implementation). See https://github.com/w3c/csswg-drafts/issues/4164 
  Enable the CSS image-orientation feature, currently hidden behind a flag. Change the default value to from-image, and apply the property to all images, both content and style. The CSS WG recently decided on the new behavior with a view to moving toward future removal of the property. The new defaults and behavior would match that if the property did not exist at all.
Safari: Shipped Mozilla: Support and indications of shipping after we do Edge/IE: Never implemented Firefox: Public support Edge: No public signals Safari: Shipped Web developers: Positive (https://bugs.chromium.org/p/chromium/issues/detail?id=158753) 153 stars for implementation of the feature, with many passionate comments. None. Trials via cluster telemetry indicate possible impact for 56/100000 sites. These sites, and others, may have content change when the feature is shipped with new defaults. https://chromestatus.com/metrics/css/popularity shows less than 0.000001% usage of the property, but this is not a good indicator of impact due the proposed default change (we might expect usage to go up as developers use image-orientation: none to maintain the current appearance of their site). None.

Chris Harrelson

unread,
Nov 21, 2019, 3:41:43 PM11/21/19
to Stephen Chenney, blink-dev
LGTM1

I agree that no TAG review is necessary, as this is just shipping a simple feature with CSSWG review and another browser already shipped.

--
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.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/blink-dev/CAGsbWzQGD1MdahaWsKtLbRP%3D6pqo1g9xjCNvu5CGWydZK7WOfA%40mail.gmail.com.

Manuel Rego Casasnovas

unread,
Nov 26, 2019, 5:17:10 PM11/26/19
to Stephen Chenney, blink-dev


On 20/11/2019 18:31, Stephen Chenney wrote:
> Specification:
https://drafts.csswg.org/css-images-3/#the-image-orientation

What syntax are you planning to ship?

I've already added a comment on the CSSWG issue, as the spec text needs
an update; the property is marked as deprecated right now.

> Safari: Shipped

I've just tested Safari Technology Preview 96 and it says invalid
property for "image-orientation".
Also the bug is still open: https://bugs.webkit.org/show_bug.cgi?id=89052

Could you clarify if this is shipped in Safari?

Thanks,
Rego
pEpkey.asc

PhistucK

unread,
Nov 27, 2019, 2:50:16 AM11/27/19
to Manuel Rego Casasnovas, Stephen Chenney, blink-dev
I think "Shipped" here is for the default behavior (from-image), not necessarily support for the property itself.

☆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+...@chromium.org.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/blink-dev/e9a048d9-5221-d9ae-53d8-703c2a0561ba%40igalia.com.

Simon Pieters

unread,
Nov 27, 2019, 6:46:24 AM11/27/19
to Stephen Chenney, blink-dev
Nice to see this change!

Den ons 20 nov. 2019 kl 18:31 skrev Stephen Chenney <sche...@chromium.org>:
sche...@chromium.org One-pager Specification: https://drafts.csswg.org/css-images-3/#the-image-orientation None. https://github.com/w3c/csswg-drafts/issues/4164 Enable the CSS image-orientation feature, currently hidden behind a flag. Change the default value to from-image, and apply the property to all images, both content and style. The CSS WG recently decided on the new behavior with a view to moving toward future removal of the property. The new defaults and behavior would match that if the property did not exist at all.

Per https://github.com/w3c/csswg-drafts/issues/4164#issuecomment-548578471 my understanding is that the property will not be removed.
 
Safari: Shipped Mozilla: Support and indications of shipping after we do Edge/IE: Never implemented Firefox: Public support Edge: No public signals Safari: Shipped Web developers: Positive (https://bugs.chromium.org/p/chromium/issues/detail?id=158753) 153 stars for implementation of the feature, with many passionate comments. None. Trials via cluster telemetry indicate possible impact for 56/100000 sites. These sites, and others, may have content change when the feature is shipped with new defaults. https://chromestatus.com/metrics/css/popularity shows less than 0.000001% usage of the property, but this is not a good indicator of impact due the proposed default change (we might expect usage to go up as developers use image-orientation: none top maintain the current appearance of their site).

There may be web sites or apps where EXIF orientation in images are present, but the site/app assumes the browser will ignore it and apply its own rotation. This can result in double rotation. The opt-out to recommend is image-orientation: none. I think it would be nice with a blog post or something about this.

 
None.
No additional support required. Yes No Existing WPT tests will be used as is.Some additional tests will be implemented and existing web_tests will be moved to WPT.

https://github.com/web-platform-tests/wpt/issues/18549 has a testing plan for image-orientation, if that helps. (I don't think anyone has started writing tests for that issue.)

cheers

https://bugs.chromium.org/p/chromium/issues/detail?id=158753 https://www.chromestatus.com/feature/6313474512650240
This intent message was generated by Chrome Platform Status.

--
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.

Noel Gordon

unread,
Nov 27, 2019, 7:32:42 AM11/27/19
to Simon Pieters, Stephen Chenney, blink-dev
/me similarly happy to see this move forward, but we still need new spec written + vendor agreement of same, to mention the developer opt-out, which I think is { image-orientation: none; }, but might be wrong since I have no spec to refer to :)

~noel

Stephen Chenney

unread,
Nov 27, 2019, 8:25:48 AM11/27/19
to PhistucK, Manuel Rego Casasnovas, blink-dev
Yes, I should have been clearer. https://bug-201123-attachments.webkit.org/attachment.cgi?id=377807 makes all images in Safari behave like image-orientation: from-image, but does not enable the CSS property. In other words, on Safari there is no way to avoid having the EXIF orientation applied.

The proposal in this intent is to enable image-orientation by default, with allowed values "from-image" and "none".

A blog post will be forthcoming to coincide with the change.

Cheers,
Stephen.

Daniel Bratell

unread,
Nov 28, 2019, 3:45:15 PM11/28/19
to Noel Gordon, Simon Pieters, Stephen Chenney, blink-dev

LGTM2 to ship behavioural change and (temporary?) opt-out property.

Will it be possible to see, in a year or so, how many have actively opted out in a way that had any effect? If not, I think it would be good to add such a use counter so that the property can be removed if it isn't used.

/Daniel

Yoav Weiss

unread,
Nov 28, 2019, 3:48:23 PM11/28/19
to Stephen Chenney, blink-dev
On Wed, Nov 20, 2019 at 6:31 PM Stephen Chenney <sche...@chromium.org> wrote:
sche...@chromium.org One-pager Specification: https://drafts.csswg.org/css-images-3/#the-image-orientation None. https://github.com/w3c/csswg-drafts/issues/4164 Enable the CSS image-orientation feature, currently hidden behind a flag. Change the default value to from-image, and apply the property to all images, both content and style.

I remember past claims that this would be problematic for compat as some images on the web today have the "wrong" EXIF and this will result in orientation change.
The fact that Safari shipped the default "from-image" behavior helps reduce the risk, but do we have a sense of how many images would be impacted by this and change orientation once this ships?
  
The CSS WG recently decided on the new behavior with a view to moving toward future removal of the property. The new defaults and behavior would match that if the property did not exist at all.
Safari: Shipped Mozilla: Support and indications of shipping after we do Edge/IE: Never implemented Firefox: Public support Edge: No public signals Safari: Shipped Web developers: Positive (https://bugs.chromium.org/p/chromium/issues/detail?id=158753) 153 stars for implementation of the feature, with many passionate comments. None. Trials via cluster telemetry indicate possible impact for 56/100000 sites. These sites, and others, may have content change when the feature is shipped with new defaults. https://chromestatus.com/metrics/css/popularity shows less than 0.000001% usage of the property, but this is not a good indicator of impact due the proposed default change (we might expect usage to go up as developers use image-orientation: none top maintain the current appearance of their site). None.
No additional support required. Yes No Existing WPT tests will be used as is.Some additional tests will be implemented and existing web_tests will be moved to WPT. https://bugs.chromium.org/p/chromium/issues/detail?id=158753 https://www.chromestatus.com/feature/6313474512650240
This intent message was generated by Chrome Platform Status.

--

Stephen Chenney

unread,
Dec 2, 2019, 11:03:24 AM12/2/19
to Yoav Weiss, blink-dev
On Thu, Nov 28, 2019 at 3:48 PM Yoav Weiss <yo...@yoav.ws> wrote:


On Wed, Nov 20, 2019 at 6:31 PM Stephen Chenney <sche...@chromium.org> wrote:
sche...@chromium.org One-pager Specification: https://drafts.csswg.org/css-images-3/#the-image-orientation None. https://github.com/w3c/csswg-drafts/issues/4164 Enable the CSS image-orientation feature, currently hidden behind a flag. Change the default value to from-image, and apply the property to all images, both content and style.

I remember past claims that this would be problematic for compat as some images on the web today have the "wrong" EXIF and this will result in orientation change.
The fact that Safari shipped the default "from-image" behavior helps reduce the risk, but do we have a sense of how many images would be impacted by this and change orientation once this ships?

The best information we have right now is the Cluster Telemetry run I reference below, showing 56/100000 sites affected. On one of those that I explicitly checked it affected one thumbnail image out of almost 100, so the actual impact is even less in some sense. I had to insert debug info and open the page to find the image that changed, and it turns out this change will make it correct.

I'll convert the cluster telemetry patch to log UMA when an image changes/does not change orientation under the new regime. Does that address concerns on tracking impact?

Yoav Weiss

unread,
Dec 2, 2019, 11:17:36 AM12/2/19
to Stephen Chenney, blink-dev
LGTM3

On Mon, Dec 2, 2019 at 5:03 PM Stephen Chenney <sche...@chromium.org> wrote:


On Thu, Nov 28, 2019 at 3:48 PM Yoav Weiss <yo...@yoav.ws> wrote:


On Wed, Nov 20, 2019 at 6:31 PM Stephen Chenney <sche...@chromium.org> wrote:
sche...@chromium.org One-pager Specification: https://drafts.csswg.org/css-images-3/#the-image-orientation None. https://github.com/w3c/csswg-drafts/issues/4164 Enable the CSS image-orientation feature, currently hidden behind a flag. Change the default value to from-image, and apply the property to all images, both content and style.

I remember past claims that this would be problematic for compat as some images on the web today have the "wrong" EXIF and this will result in orientation change.
The fact that Safari shipped the default "from-image" behavior helps reduce the risk, but do we have a sense of how many images would be impacted by this and change orientation once this ships?

The best information we have right now is the Cluster Telemetry run I reference below, showing 56/100000 sites affected. On one of those that I explicitly checked it affected one thumbnail image out of almost 100, so the actual impact is even less in some sense. I had to insert debug info and open the page to find the image that changed, and it turns out this change will make it correct.

I'll convert the cluster telemetry patch to log UMA when an image changes/does not change orientation under the new regime. Does that address concerns on tracking impact?

Apologies for missing that reference. That usage sounds fairly low.
Reply all
Reply to author
Forward
0 new messages