Conversation
queengooborg
added
data:api
foolip
left a comment
foolip
left a comment
There was a problem hiding this comment.
I agree this isn't valuable enough to preserve, and pinning down a version for Safari would be hard. I tried to make a test case in another PR that I can't find now, but it's hard to say what exactly counts as enough smoothing.
I don't think this is an entirely clear cut case, however, so I'll defer to @ddbeck or @Elchi3 to approve and merge.
|
I'd be willing to go along with removing this as matter of practicality alone. It's a silly web koan: if a feature can't be observed, is it a feature at all? Plus the spec backs us up here. It looks like this feature came about because of this bug https://bugzilla.mozilla.org/show_bug.cgi?id=1360415 and this change to the spec whatwg/html@5854cb0. At present, the spec says:
Seems like UAs are supposed to be free to do as they wish here and I'm not sure it helps web developers to imply that this behavior is supposed to have any consistency (even if it does). |
queengooborg
deleted the
api/CanvasRenderingContext2D/smoothing_downscaling
branch
* Bump version to v4.0.6 * Add release note for #12094 * Add release note for #12367 * Add release note for #12462 * Add release note for #12490 * Add release note for #12565 * Add release note for #12632 * Add release note for #12638 * Add release note for #12639 * Add release note for #12703 * Add release note for #12718 * Add release note for #12723 * Add release note for #12768 * Add release note for #12770 * Add release note for #12772 * Add release note for #12794 * Add release note for #12796 * Add release note for #12813 * Add stats and hoist known issues
3 participants
This PR removes the
smoothing_downscalingsubfeature of thedrawImagemethod ofCanvasRenderingContext2D. This feature was a remnant from the old wiki migration. There is nothing in the docs explaining what exactly this means, so it's unclear how to test support for this. With that, plus the fact that this isn't a spec-defined behavior, I think it makes sense to simply remove this subfeature from BCD.