Weekly Edition
Return to the Development page
| |||||||||||||||||||
Firefox 18 is now available
Firefox 18 is now available as a free download for Windows, Mac, and Linux from http://www.mozilla.org/firefox/new/. As always, we recommend that users keep up to date with the newest version of Firefox for the latest features and fixes. The release notes for Firefox 18 are available at http://www.mozilla.org/en-US/firefox/18.0/releasenotes/. Firefox 18.0 is also now available for Android. The associated release notes are available at https://www.mozilla.org/en-US/mobile/18.0/releasenotes/. - Bhavana Bajaj Release Manager at Mozilla _______________________________________________ announce mailing list announce@lists.mozilla.org https://lists.mozilla.org/listinfo/announce (Log in to post comments)
Retina display support Posted Jan 9, 2013 9:52 UTC (Wed) by epa (subscriber, #39769) [Link]
The "what's new" talks about support for Apple's high-resolution "Retina displays" but doesn't go into detail about what this involves. Does it do anything interesting, and if so is there a way to enable this on operating systems other than Mac OS X?
Retina display support Posted Jan 9, 2013 10:30 UTC (Wed) by roc (subscriber, #30627) [Link]
It means we render everything at the device resolution, so CSS 1px x 1px renders as 2x2 device pixels. It's kinda like having a default zoom.
We have some support for that for Windows, but it's currently disabled because of a few major bugs.
It would be pretty easy to extend to Linux/X11 too, except we don't know exactly which system setting(s) to observe to know when to automatically zoom.
Retina display support Posted Jan 9, 2013 11:03 UTC (Wed) by epa (subscriber, #39769) [Link] It means we render everything at the device resolution, so CSS 1px x 1px renders as 2x2 device pixels. It's kinda like having a default zoom.Thanks for the explanation. At work I run Firefox on Windows on a high-resolution display, and use Ctrl-+ to zoom in, which magnifies both the text and the images. Is that equivalent?
Retina display support Posted Jan 9, 2013 12:50 UTC (Wed) by roc (subscriber, #30627) [Link]
Yes, that's pretty close.
Retina display support Posted Jan 9, 2013 17:45 UTC (Wed) by nye (guest, #51576) [Link]
Wait, so by default does that mean that images are shown at 200% zoom? Isn't that going to look terrible?
Is there any way for a site author to 'opt-out' of this? Maybe something involving a media-dependent stylesheet or something.
Retina display support Posted Jan 9, 2013 18:41 UTC (Wed) by Cyberax (✭ supporter ✭, #52523) [Link]
> Wait, so by default does that mean that images are shown at 200% zoom? Isn't that going to look terrible?
Yup. It's not that terrible in practice, but definitely not optimal.
> Is there any way for a site author to 'opt-out' of this? Maybe something involving a media-dependent stylesheet or something.
Retina display support Posted Jan 9, 2013 21:40 UTC (Wed) by roc (subscriber, #30627) [Link]
Terrible compared to what alternatives? Showing images at half their expected size? On a Retina screen, images have to be scaled up somehow.
For non-Retina-aware apps, MacOS just scales the entire window up by a factor of 2. Being Retina-aware means for things we can scale intelligently (text, vector graphics) we do much better. For images we usually can't do much better although at least we give the author some control (choosing between nearest-neighbour and bilinear scaling via the image-rendering CSS property).
Retina display support Posted Jan 10, 2013 9:52 UTC (Thu) by epa (subscriber, #39769) [Link]
Ah, so the site author can choose nearest-neighbour scaling using a CSS property. It would be lovely if Firefox also supported hqx or other intelligent scaling methods for bitmap graphics.
Retina display support Posted Jan 10, 2013 13:08 UTC (Thu) by nye (guest, #51576) [Link]
>Terrible compared to what alternatives? Showing images at half their expected size?
No, compared to showing it at the *expected* size. If my image is 800px wide, the expected size is 800px. Stretching it to 1600px mean showing it at *double* its expected size.
>On a Retina screen, images have to be scaled up somehow.
I have here a number of displays, ranging from around 100ppi to around 250ppi. Looking at the same image on the lowest and highest density displays, it looks far better on the higher one when displayed at its native size. If that gets rescaled to make it the same *physical* size, the theoretical best case scenario is that it looks like it does on the low density display, except in practice the better display means that you can really see the deficiency of the scaling.
I recognise that there is a real problem to be solved here, where some genius has specified a text height of 12px and some background image to match (or whatever), but the solution given doesn't seem like a good one at all.
Retina display support Posted Jan 10, 2013 13:29 UTC (Thu) by nye (guest, #51576) [Link]
>I have here a number of displays, ranging from around 100ppi to around 250ppi. Looking at the same image on the lowest and highest density displays, it looks far better on the higher one when displayed at its native size. If that gets rescaled to make it the same *physical* size, the theoretical best case scenario is that it looks like it does on the low density display, except in practice the better display means that you can really see the deficiency of the scaling.
To be more specific, when looking at the same image at the same physical size and distance on both screens, the difference is that the high resolution display looks blurred - this is with whatever scaler the Android browser uses.
I *hate* blurry visuals. They're tiring to look at because my eyes keep trying harder to focus until they've had a *long* time to get used to the knowledge that they simply can't; that's something that never happens in the physical world, so it makes an unpleasant transition between looking at a screen and looking at real physical objects.
Related: the default font smoothing settings on Ubuntu (picked that only because it's the only distro I've used recently with an out-of-the-box config - maybe others are different) looks like somebody took a stick of lard and smeared it over the monitor; it disgusts me, and it's uncomfortable to read. In this area at least, high PPI displays are a clear win regardless of how images are treated.
I can see that the image-doubling behaviour would be desirable if you value large images over sharp images, but that's a position I can't even begin to empathise with in any way because my eyes find it physically unpleasant.
Retina display support Posted Jan 12, 2013 6:55 UTC (Sat) by alankila (subscriber, #47141) [Link]
Notice that pixel-doubling scaling preserves the image contrast and every other property, at the cost of not utilizing the additional resolution in any way. Things look the same despite you have double the pixels!
Interpolation filters produce errors when encountering high-frequency content of the image. These errors follow from ignored gamma correction (= scaled images tend to be too dark at high-frequency features) and modeling of image as infinitesimal points of specific color and intensity at particular points of a grid, rather than interpreting the image as an area-averaged sampling of an unknown light function. If you had a theory of the light function whose gridded area-average produced the current image, you would be able to resample such function to any resolution, and likely the results would be more natural too. In any case the upshot of the latter effect is reduction of contrast of high-frequency images. I studied the problem slightly here:
https://bel.fi/alankila/scaling/
Much more work should be done to turn this into something practical, though.
Retina display support Posted Jan 12, 2013 19:02 UTC (Sat) by jimparis (subscriber, #38647) [Link]
> If you had a theory of the light function whose gridded area-average produced the current image, you would be able to resample such function to any resolution, and likely the results would be more natural too.
There's a really simple light function that produces the current image: solid colored squares, each exactly the size of the image's pixels. Which leads exactly to the "pixel-doubling scaling" you describe. You might manage to come up with an alternate light function that also fits, but there's no way of knowing that it's any more correct.
Retina display support Posted Jan 13, 2013 0:22 UTC (Sun) by alankila (subscriber, #47141) [Link]
Acceptable, but I think some kind of gaussian blurring combined with edge enhance would be a reasonable choice for at least natural photographs. Ultimately we DO know that the image is almost certainly not composed of tiny squares of colors in the real world, perfectly aligned to fall into the camera's pixel grid, after all.
So it is my faith that we can do better than that, and not even try very hard; the trouble may be in the corner cases where it would appear that we do worse. For instance, over/undersaturation is a risk for edge enhancements that are somehow tangentially involved in algorithm like this. How would you represent color that has negative luminosity? You can't, so for some cases the algorithm would have to degrade naturally.
Retina display support Posted Jan 14, 2013 7:59 UTC (Mon) by ekj (guest, #1524) [Link]
When the image is a result of capturing the real world on a camera-sensor, then this is indeed true, and as you say, algorithms that do more funky things than pixel-doubling perform well.
But we don't know that, generally speaking. (though we can attempt to detect it) Lots of images are the result of computation, they might be output from a graphing-library, for example.
What works well for a photo of a face, doesn't always work well for output from gnuplot.
Retina display support Posted Jan 10, 2013 13:34 UTC (Thu) by mpr22 (subscriber, #60784) [Link] Rendering at 1px=1px on high-pixel-density high-pixel-count displays does not, on first inspection, look like it is likely to perform well on accessibility criteria. Converting a CSS distance of 1px into a display distance of two pixels looks rather more likely to perform well on this score. As long as those who crave tiny images can easily turn off the scaling, the default behaviour should be chosen to be kinder to those for whom tiny images are not sensible.
Retina display support Posted Jan 9, 2013 22:26 UTC (Wed) by redden0t8 (guest, #72783) [Link]
I'd go for blurry-but-visible over sharp-but-illegibly-tiny any day.
Even if all the browser did was plain pixel-doubling, you'd still end up with your image exactly as it would appear on a low-dpi monitor - the only reason it looks worse is because the rest of the screen is high res so the pixelation is more obvious.
Once you add in good anti-aliasing, IMHO even a low-dpi image scaled up on a hi-dpi display looks better than the same low-dpi image displayed natively on a low-dpi display.
Retina display support Posted Jan 9, 2013 11:39 UTC (Wed) by ewan (subscriber, #5533) [Link]
Isn't that pretty much the exact opposite of supporting a retina display? AIUI the default behaviour of MacOS X is to do exactly as you describe for any application that doesn't specifically ask about retina displays; that way legacy applications look OK (but don't benefit) and retina aware ones get to draw things using the full resolution.
Retina display support Posted Jan 9, 2013 12:53 UTC (Wed) by roc (subscriber, #30627) [Link]
We're rendering everything, especially vector graphics, at the higher resolution. E.g. text that's 20 CSS pixels high gets rendered by rasterizing glyph curves at 40 device pixels high. We don't rasterize at 20 device pixels and then scale the result.
Retina display support Posted Jan 9, 2013 12:16 UTC (Wed) by niner (subscriber, #26151) [Link]
I thought we were slowly moving past browser differences. But suddenly not even the very basic and most reliable distance unit in CSS is the same anymore? Just great...
Retina display support Posted Jan 9, 2013 12:34 UTC (Wed) by NAR (subscriber, #1313) [Link]
A Hungarian news portal recently redesigned its look. It doesn't worth mentioning that it works with some browsers and doesn't with others. Even the same browser on the same operating system could lead to different (working vs non-working) results. I my opinion this "web application thing" is a really stupid idea, just doesn't work. It is probably no wonder that there's so much demand for apps on mobile - the browsers just don't work.
Retina display support Posted Jan 12, 2013 16:14 UTC (Sat) by intgr (subscriber, #39733) [Link]
The designers at one website got it wrong and you're saying that the whole web is broken because of it?
Retina display support Posted Jan 14, 2013 10:38 UTC (Mon) by NAR (subscriber, #1313) [Link]
Not one - I have 4 browsers installed and running, because some sites work only with certain browsers. Probably LWN is the only site I use that works with all browsers, even gmail is broken in Opera.
Retina display support Posted Jan 21, 2013 23:30 UTC (Mon) by oak (subscriber, #2786) [Link]
At least last year Gmail seemed frequently broken also in Debian Stable's old version of Google's Chromium browser (Gmail tab freezes after the page has loaded). If Google's mail web service doesn't work in Google's web browser, I think it a good indication that web indeed is broken.
Retina display support Posted Jan 9, 2013 12:56 UTC (Wed) by roc (subscriber, #30627) [Link]
CSS pixels have always been defined to be a resolution-independent unit (since CSS1 in 1995, at least). They were never defined to be screen pixels. The new Firefox behavior is completely consistent with the spec. It's also consistent with other Mac browsers. And it doesn't really impact Web sites at all; for almost all sites, the only impact is that they look nicer.
Retina display support Posted Jan 9, 2013 13:18 UTC (Wed) by robert_s (subscriber, #42402) [Link]
"CSS pixels have always been defined to be a resolution-independent unit (since CSS1 in 1995, at least)."
_What_
"They were never defined to be screen pixels. The new Firefox behavior is completely consistent with the spec."
Thing is, they _have_ effectively been screen pixels for the last ~20 years, and that's how people have been using them.
Being able to describe things in terms of screen pixels is fairly critical in order to avoid fuzzy graphics.
Retina display support Posted Jan 9, 2013 13:35 UTC (Wed) by Kit (guest, #55925) [Link]
Zooming. That already caused screen pixels to not be the same as device pixels, and most browsers have supported it for many years now (whole page zooming, not just text size adjustment). Hell, zooming is a major portion of what makes modern mobile browsers usable without special coding.
I wouldn't be surprised if it hadn't been true in practice even before that.
Retina display support Posted Jan 9, 2013 13:36 UTC (Wed) by roc (subscriber, #30627) [Link]
Thanks to browser zoom features, and mobile browsers, CSS pixels haven't reliably corresponded to screen pixels for quite some time now.
We've found ways to avoid almost all problems when zooming Web pages.
Retina display support Posted Jan 9, 2013 15:02 UTC (Wed) by robert_s (subscriber, #42402) [Link]
"CSS pixels haven't reliably corresponded to screen pixels for quite some time now."
Depends what you mean by reliably. I'd say it's a good 90% of the time people aren't zooming or using a mobile browser and they do correspond. And in a world where nearly half of visitors use IE so nothing really works properly anyway, I call that relatively reliable.
I'm just quite depressed by the whole way "retina" has been handled by everyone. There have been many efforts at resolution independence over the last ten years - now when significantly different resolution devices come into the market all that seems to have been thrown out the window and everyone seems to need to go round adding "retina"-specific hacks, "retina"-specific icons and whatnot to everything under the sun.
"We've found ways to avoid almost all problems when zooming Web pages."
As a web developer I look at that statement with eyebrows askew.
Retina display support Posted Jan 9, 2013 15:40 UTC (Wed) by Cyberax (✭ supporter ✭, #52523) [Link]
There really were not any serious attempts at resolution independence earlier. Most of available displays had DPI in range of 96-140 and that allowed to simply ignore occasional problems.
Retina displays allow no such slack. So everyone suddenly realized that their code is not actually that resolution independent and they have to do something about it.
Retina display support Posted Jan 9, 2013 18:59 UTC (Wed) by redden0t8 (guest, #72783) [Link]
I'm not really sure what you mean by Retina-specific hacks. As a notebook Retina-display user, I can say in 6 months of daily browsing, I have yet to find a website that doesn't render correctly.
The browser still lays things out identically, it just renders extra pixels for text+vector graphics, and anti-aliases raster images. That's the whole reason why it's implemented using a pixel-ratio. Everything automatically works.
Coding webpages in real units is impractical anyways, as you don't know the device it'll be displayed on. No one on a phone wants a webpage to be initially zoomed to be 10 inches wide when their screen is only 2-3 inches wide. The first thing they'll do is zoom out to figure out where to scroll to before zooming back in. Same idea for a tablet. Laptops/desktops are a little easier, but you still don't know the size of the display (9" netbook or 30" desktop?), how far away the user is sitting (HTPC or netbook?), and how good their eyesight is. Ultimately, it's up to the device-maker and end user.
All the web-developer needs is a reference unit. The "pixel", while poorly named, already provides this. Each device maker can decide how to handle the scaling ratio for a given form-factor. Ideally one day, once desktop OS's are truly resolution independent, they'll let end-users pick their "ideal" ratio with a slider somewhere that affects the entire UI.
Yes you'll need to start serving up higher resolution images soon or later, but this is caused by hi-dpi displays in general and not just Retina(TM) displays. It's already possible to query the device pixel ratio in a browser-independent way, or alternately blindly serve up high-res images to everyone.
Retina display support Posted Jan 9, 2013 20:09 UTC (Wed) by Cyberax (✭ supporter ✭, #52523) [Link]
HTML actually has an abstract 'em' unit which is linked to the text size.
I've actually long wished that we had some kind of 'pixels' with fixed angular size. But I kinda doubt that most web-developers would like to work with steradians :)
Retina display support Posted Jan 9, 2013 21:41 UTC (Wed) by roc (subscriber, #30627) [Link]
CSS pixels are exactly "some kind of 'pixels' with fixed angular size".
Retina display support Posted Jan 9, 2013 21:59 UTC (Wed) by Cyberax (✭ supporter ✭, #52523) [Link]
Then we lose ability to use sub-pixel positioning (last time I checked you can't use fractional pixels).
That's not a problem right now, since browser does all the drawing. But once we move into the brave new Canvas world, we'll start to get this kinds of problems. 1 pixel on a Retina display might be below eye's resolution, but 1px misalignment is easily noticeable.
Retina display support Posted Jan 10, 2013 10:35 UTC (Thu) by roc (subscriber, #30627) [Link]
You absolutely can use fractional CSS pixels.
Retina display support Posted Jan 11, 2013 10:02 UTC (Fri) by jezuch (subscriber, #52988) [Link]
Speaking of resolution-independence, I remember being excited about promises made by KDE people regarding resolution-independent desktop in KDE4 - things like vector icons and such. I don't think I'm seeing it in the final result. On my work laptop I had to increase font sizes in the GUI and now it looks somewhat out of place, like Frankenstein's monster living in a doll house. I suspect I'm missing something here...
Retina display support Posted Jan 13, 2013 11:01 UTC (Sun) by jospoortvliet (subscriber, #33164) [Link]
i guess it turned out to be harder than expected. The Oxygen style can scale only partially and while there are vector icons (and a high resolution Oxygen icon set), you simply cant make one icon which.looks good in all sizes. Note that retina in the way apple did it it, as far as i know at least, isn't resolution independent either - it is mostly just a doubling of the resolution of everything. Having a scaling of 125% for example is much harder, let alone arbitrary %s.
Retina display support Posted Jan 9, 2013 16:08 UTC (Wed) by niner (subscriber, #26151) [Link]
Actually the behaviour seems not to be consitent with the spec. To quote the spec:
"For a CSS device, these dimensions are either anchored (i) by relating the physical units to their physical measurements, or (ii) by relating the pixel unit to the reference pixel. For print media and similar high-resolution devices, the anchor unit should be one of the standard physical units (inches, centimeters, etc). For lower-resolution devices, and devices with unusual viewing distances, it is recommended instead that the anchor unit be the pixel unit."
So for example a 3rd generation Apple iPad with so called "Retina display" has a screen resolution of 263.92 PPI which is much closer to a 300 DPI print than to the usual 90 DPI of a computer screen. Google Nexus 10 features 300.24 PPI. So I would say these devices are what the spec call "print media and similiar high-resolution devices" and 1px sould be 0.75pt or 0.26458333mm.
Using 2x2 device pixels seems to be just as wrong as using 1x1. 2 display pixels on the Nexus 10 for example are just 0.1692mm wide compared to ~ 0.26mm width of the reference pixel the CSS spec uses.
Retina display support Posted Jan 9, 2013 21:39 UTC (Wed) by butlerm (subscriber, #13312) [Link]
96 CSS pixels per inch is normal (or at least nominal) for a desktop computer screen. A handheld device is typically going to have a substantially different viewing distance, and as a consequence it isn't at all natural to display the same content at the same number of CSS pixels per inch. A CSS pixel should (roughly) encompass the same angle, not the same physical size.
As a result, 2 x 2 device pixels per CSS pixel is just about ideal for a 192 DPI desktop display. Ideal as in minimizes aliasing for pixel oriented page designs, which are commonplace. If you have a handheld device, the DPI should be higher because the viewing distance is going to be less. 384 or even 400 dpi sounds about right for 2 x 2 scaling (2 dppx) on a device like that. 4x4 scaling someday (~768 DPI handheld, 384 DPI desktop) sounds well within the retina limit to me, at least if you want to avoid eyestrain.
Retina display support Posted Jan 9, 2013 21:45 UTC (Wed) by roc (subscriber, #30627) [Link]
You're not taking into account viewing distance. Tablets and phones are viewed at a shorter distance than desktop displays. For a desktop display viewed at arm's length, a CSS pixel should be one device pixel on a 96dpi screen, but for a phone viewed at half that distance, a CSS pixel should be one device pixel on a 192dpi screen (assuming default zoom).
Retina display support Posted Jan 10, 2013 2:09 UTC (Thu) by Trelane (subscriber, #56877) [Link]
> It would be pretty easy to extend to Linux/X11 too, except we don't know exactly which system setting(s) to observe to know when to automatically zoom.
What are you looking for exactly?
xrandr --query
is this what you need (modulo actually having to run xrandr instead)?
http://developer.gnome.org/gdk3/stable/GdkScreen.html
Retina display support Posted Jan 10, 2013 10:36 UTC (Thu) by roc (subscriber, #30627) [Link]
What we're looking for is a system setting that tells us exactly how much to scale by default. Just guessing based on the characteristics of the screen as reported by xrandr isn't enough information to ensure that, for example, our standard UI controls are the same size as the controls in other applications.
Retina display support Posted Jan 11, 2013 1:09 UTC (Fri) by FedericoMenaQuintero (guest, #81416) [Link]
GTK+ doesn't have a special API for high-DPI devices yet. So, for now, don't do anything special - use a scale of 1.
(Your controls *will* be the same size as those in other applications, because you are using the same toolkit.)
Firefox 18 is now available Posted Jan 9, 2013 14:39 UTC (Wed) by mirabilos (subscriber, #84359) [Link]
It should be noted that this is the first version whose binaries do not run on Kubuntu Hardy and Debian Lenny and possibly Squeeze any more, due to reliance on pretty recent GNU libc features. There’s currently an open bug about this, but the 17 ESR series continues to work for now (although 17.0.2 builds before 3 were also broken).
Firefox 18 is now available Posted Jan 9, 2013 16:38 UTC (Wed) by dskoll (subscriber, #1630) [Link] Runs fine for me on Squeeze 64-bit. Posting from FFox 18 now.
Firefox 18 is now available Posted Jan 9, 2013 18:28 UTC (Wed) by iarenaza (subscriber, #4812) [Link]
Same here (works as expected).
Firefox 18 is now available Posted Jan 9, 2013 18:38 UTC (Wed) by nteon (subscriber, #53899) [Link]
do you have any more information or a link to the bug? I'm curious what these new libc features are
Firefox 18 is now available Posted Jan 10, 2013 10:14 UTC (Thu) by khim (subscriber, #9252) [Link] "newer feature" is overstatement. Most likely this is the result of the old tempest: when GLibC developers changed Programs are incorrect (since
Firefox 18 is now available Posted Jan 10, 2013 19:42 UTC (Thu) by mirabilos (subscriber, #84359) [Link]
I don’t think it was that (but considering http://blog.fefe.de/?ts=ae12497b nothing in that source code surprises me any more; IMHO they should’ve thrown away the old Netscape code and started anew).
https://bugzilla.mozilla.org/show_bug.cgi?id=723487 is the bugreport.
|
|||||||||||||||||||