Force image inputs to be clipped as well.#12491
Conversation
Matches Blink and WebKit behavior, and Gecko behavior before https://bugzil.la/1999100 See https://bugzil.la/2038137.
annevk
left a comment
There was a problem hiding this comment.
This looks okay to me. OP needs to be filled out though.
|
Filed the relevant bugs. |
…r=firefox-style-system-reviewers,boris Matches other browsers' behavior and pre-regression behavior, see whatwg/html#12491. Differential Revision: https://phabricator.services.mozilla.com/D302639
|
WebKit doesn't support overflow-clip-margin, but I'm sure it probably makes sense. |
|
So image isn't really special here, its a "button" type. Our manual logic is: E.g. see clipping here: Something more complete which would likely allow us to remove this would be: I think gecko previously had |
|
@bfgeek but image clips to the content box, doesn't it? See the rendering of this in Chromium: |
|
I guess you clip to the content-box when is an image and to the padding-box when it's not? I'd prefer to standardize on this since Gecko used to not clip at all when not an image, and clip to the content-box when it is. |
|
Ah yeah thats what we are doing. I think i'd still prefer something like: |
|
@bfgeek That's basically this change, right? But with changing the existing |
|
@bfgeek d'oh, you're right, but that's because the default |
|
Does it even impact the computed value? Presumably that’s normalized? |
|
Yeah it is: |
|
Ah sorry I forgot that it defaulted to the padding-box. Ok you patch looks good then. I think the only weird thing then is how the |
|
Curious, why? |
|
Err, scratch that, I totally mis-grepped our UA sheet...
It seems Blink has |
|
I spun off #12521 to discuss |
|
(But ok, seems this should be fine :)) |

Matches Blink and WebKit behavior, and Gecko behavior before https://bugzil.la/1999100
See https://bugzil.la/2038137.
(See WHATWG Working Mode: Changes for more details.)
/rendering.html ( diff )