Bug 78087 – Enable the 'resolution' media query by default

• Component: CSS
• Status: RESOLVED FIXED

[Simon Fraser (smfr)]

Need to implement
http://dev.w3.org/csswg/css3-mediaqueries/#resolution0

See FIXME in MediaQueryEvaluator.cpp:
FIXME: following media features are not implemented: color_index, scan, resolution

[Simon Fraser (smfr)]

This should do the right thing when printing too (report printer output resolution).

[Simon Fraser (smfr)]

Also need to match what other browsers do when pageScale changes.

[Radar WebKit Bug Importer]

<rdar://problem/10829378>

[Eric Meyer]

If you want a quick initial test, http://meyerweb.com/eric/css/tests/mq-valueless.html includes valueless, iPhone 4S value (I think), and min/max queries.  See test 11.  If not, no harm done.

[Theresa O'Connor]

General comment: It's not clear to me why we would implement the resolution media query, considering how broken it is.

Eric, that isn't the correct iPhone 4S value per the spec, because the CSS dpi unit is "physical pixels per CSS inch" and a CSS inch is 96 CSS pixels. So "resolution: 192dpi" should match on an iPhone 4S.

[Seo Sanghyeon]

How does this relate to bug 85262?

[Simon Fraser (smfr)]

(In reply to comment #6)
> How does this relate to bug 85262?

That is about a new CSS property based on the resolution of image resources. This is about a media query based on screen resolution.

[Adele Peterson]

Should we dupe this to https://bugs.webkit.org/show_bug.cgi?id=99077?

[Kenneth Rohde Christiansen]

(In reply to comment #8)
> Should we dupe this to https://bugs.webkit.org/show_bug.cgi?id=99077?

It still need some special handling when printing as smfr pointed out, but maybe that is another bug

[Kenneth Rohde Christiansen]

(In reply to comment #2)
> Also need to match what other browsers do when pageScale changes.

According to https://bugzilla.mozilla.org/show_bug.cgi?id=662061 zooming ( and I assume pageScale) should not affect the resolution values.

It currently does in Firefox 16, at least full page zoom, though not text size zoom.

[Simon Fraser (smfr)]

Support for the 'resolution'  media query was added behind ENABLE(RESOLUTION_MEDIA_QUERY) via bug 99077.

[Albert Scheiner]

Are there any plans to enable it by default any time soon? Once dppx support lands in IE (it's currently in preview built) all other major browser have added full support.

http://dabblet.com/gist/7166616
https://status.modern.ie/dppxunitfortheresolutionmediaquery?term=dppx

[Chris Rebert]

MS Edge has shipped this (see http://caniuse.com/#feat=css-media-resolution ), so now only WebKit doesn't support this.

[Mattia Astorino]

This bug was filed 8 years ago, even IE 9 supports this important media query. What happened here? 😕

[Brent Fulgham]

(In reply to Mattia Astorino from comment #14)
> This bug was filed 8 years ago, even IE 9 supports this important media
> query. What happened here? 😕

The media query has existed in WebKit since 2012. It looks like some details specific to printing were not completed

@Mattia: What issues have you had when you used the 'resolution' media query in Webkit? Were you attempting printing?

[Brent Fulgham]

(In reply to Brent Fulgham from comment #15)
> (In reply to Mattia Astorino from comment #14)
> > This bug was filed 8 years ago, even IE 9 supports this important media
> > query. What happened here? 😕
> 
> The media query has existed in WebKit since 2012. It looks like some details
> specific to printing were not completed
> 
> @Mattia: What issues have you had when you used the 'resolution' media query
> in Webkit? Were you attempting printing?

I see - it's present but apparently not enabled by default? I'm trying to see what the correct state of this flag should be.

[Sam Sneddon [:gsnedders]]

(In reply to Brent Fulgham from comment #16)
> (In reply to Brent Fulgham from comment #15)
> > (In reply to Mattia Astorino from comment #14)
> > > This bug was filed 8 years ago, even IE 9 supports this important media
> > > query. What happened here? 😕
> > 
> > The media query has existed in WebKit since 2012. It looks like some details
> > specific to printing were not completed
> > 
> > @Mattia: What issues have you had when you used the 'resolution' media query
> > in Webkit? Were you attempting printing?
> 
> I see - it's present but apparently not enabled by default? I'm trying to
> see what the correct state of this flag should be.

Correct; ENABLE_RESOLUTION_MEDIA_QUERY is not defined for any current port, as far as I can tell (and that's likely been the case since the Blink fork). It doesn't look like there's anything too likely to have bitrotted or is missing there, at least.

(In reply to Theresa O'Connor from comment #5)
> General comment: It's not clear to me why we would implement the resolution
> media query, considering how broken it is.

The thread starting at https://lists.w3.org/Archives/Public/www-style/2012Apr/0674.html is about this brokenness, and summarised in Maciej's response in https://lists.w3.org/Archives/Public/www-style/2012Apr/0747.html:

> For comparison, device-pixel-ratio was first implemented as a vendor extension in 2006[1] and exists (prefixed) in at least WebKit, Presto and Gecko. So I wondered why device-pixel-ratio counts as "add[ing] anything new" but the [first defined in late 2011/early 2012] dppx unit does not. I don't follow closely, so maybe there is some reason for diverging from content and implementations here.

Regardless, while it might be the case that resolution: [n]dppx (and the other <resolution> units, aside from dppx, are confusing) is redundant with -webkit-device-pixel-ratio: n (which is itself now standardised in https://compat.spec.whatwg.org/#css-media-queries-webkit-device-pixel-ratio), I think it's pretty clear the ship here has sailed.

[Sam Sneddon [:gsnedders]]

Created attachment 446755 [details]
Patch

[EWS]

Committed r286874 (245104@main): <https://commits.webkit.org/245104@main>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 446755 [details].

[WebKit Commit Bot]

Re-opened since this is blocked by bug 236734

[Sam Sneddon [:gsnedders]]

bug 236804 reverted the revert; this shouldn't have been reverted on trunk for that breakage

I'm having trouble understanding the recent commits / reverts on this.

Is there an intent to *enable* the 'resolution' media query by default in Webkit (Safari)?

Or when we say "that ship has sailed," are we saying that Safari will continue to *not* support the 'resolution' media query in lieu of the '-webkit-device-pixel-ratio' feature?

[Sam Sneddon [:gsnedders]]

(In reply to matthew.wade from comment #22)
> I'm having trouble understanding the recent commits / reverts on this.
> 
> Is there an intent to *enable* the 'resolution' media query by default in
> Webkit (Safari)?
> 
> Or when we say "that ship has sailed," are we saying that Safari will
> continue to *not* support the 'resolution' media query in lieu of the
> '-webkit-device-pixel-ratio' feature?

It is enabled by default, and has been on trunk since December. It was disabled on one Safari release branch (and, inadvertently also on trunk briefly) due to site breakage caused by enabling it (mostly sites using `@media not all and (min-resolution:.001dpcm)` to specifically target WebKit, which inevitably stops matching with support for the resolution feature).

As for Safari, Apple does not comment on future releases.

(In reply to Sam Sneddon [:gsnedders] from comment #23)
> (In reply to matthew.wade from comment #22)
> > I'm having trouble understanding the recent commits / reverts on this.
> > 
> > Is there an intent to *enable* the 'resolution' media query by default in
> > Webkit (Safari)?
> > 
> > Or when we say "that ship has sailed," are we saying that Safari will
> > continue to *not* support the 'resolution' media query in lieu of the
> > '-webkit-device-pixel-ratio' feature?
> 
> It is enabled by default, and has been on trunk since December. It was
> disabled on one Safari release branch (and, inadvertently also on trunk
> briefly) due to site breakage caused by enabling it (mostly sites using
> `@media not all and (min-resolution:.001dpcm)` to specifically target
> WebKit, which inevitably stops matching with support for the resolution
> feature).
> 
> As for Safari, Apple does not comment on future releases.

Thanks, this is very helpful! Also, you practically read my mind with the `@media not all and (min-resolution: .001dpcm)` hack, I was essentially asking how "safe" that is to continue using as a way to detect Safari. Based on all this, my guess is the answer is basically "it still works in the latest Safari, but could stop working at any time."

I didn't realize the split between Webkit and how Apple uses it in Safari releases, so thanks for that context as well.