1511177 - Make fallback viewport for desktop pages configurable

Status: NEW

(GeckoView :: General, defect, P3)

Description

[Chris Peterson [:cpeterson]]

STR:
1. Send yourself an email containing the following link:
https://medium.com/@RyanMarshall80/the-90s-aesthetic-188b289144b7
2. In Gmail, clink to open the email's link.

EXPECTED RESULT:
In Focus+WebView, Gmail's redirect page is zoomed to make reading and clicking the URL easier.

ACTUAL RESULT:
In Focus+GV, Gmail's redirect page has very tiny text. The user must manually zoom to read and click the URL.

This is probably a Gecko font inflation issue.

Here are screenshots of the redirect page in Focus+WebView and Focus+GV:

https://docs.google.com/presentation/d/1-9EmORROWirZ2KeRjYDPZqT-A3Ph8GzRpiJSVZkzRfo/edit#slide=id.g48451af66a_1_15

Comment 1

[Chris Peterson [:cpeterson]]

Attachment: redirect-page-screenshot.png

Here is the side-by-side screenshot comparison.

Comment 2

[Jan Henning [:JanH]]

(In reply to Chris Peterson [:cpeterson] from comment #0)
> This is probably a Gecko font inflation issue.

Yes and no. Chrome has font inflation (variously called font boosting, or text autosizing) as well, but AFAIK that feature isn't exposed to WebViews.

The difference here is down to the viewport handling for non-mobile pages: Gecko (or Blink inside Chrome for that matter) use a viewport width of 980 CSS px for pages that don't specify a viewport width themselves (and then scale the page down to fit the screen), which preserves the page layout, but leads to small text. The WebView used by Focus on the other hand doesn't do anything special [1] and just displays the page using an implicit 'width="device-width"', which keeps the font size readable, but can lead to layout breakage, as most non-mobile/non-responsive-design pages don't really handle that narrow viewports very well.

Testing with Firefox (https://www.google.com/url?q=https://example.com is an easier way to get the same kind of page), enabling font inflation *would* solve that issue, but font inflation currently isn't exposed in GeckoView.

[1] I've forgotten what it's called, but when I looked through the WebView API some time ago, I found some promising option that sounds like it would control that feature, so either Focus turned it off, or it's turned off by default and Focus never bothered to enable it.

Comment 3

[Chris Peterson [:cpeterson]]

James says Jan's explanation in comment 2 is correct.

Randall is adding a setting to control desktop viewport mode.

Comment 4

[Chris Peterson [:cpeterson]]

Vesta says this bug should be a P1 for Fenix.

Comment 5

[Chris Peterson [:cpeterson]]

Chrome and Gecko have different default viewport widths when there is to meta tag. Chrome uses the device width and Safari used 980 px if there is no meta tag.

Comment 6

[Jan Henning [:JanH]]

(In reply to Chris Peterson [:cpeterson] from comment #5)

Chrome and Gecko have different default viewport widths when there is to meta tag.

I'm not quite sure where you've got that idea from, but Chrome most certainly uses 980 CSS px for pages without a viewport tag, too. Maybe you mean WebView, where it could indeed be that getUseWideViewPort() is false by default, leading to the previous behaviour of Focus.
A wide viewport for desktop pages is also implicitly required for font inflation in both Chrome and Firefox.

(Incidentally, apparently setUseWideViewPort(false) not just turns off the wide viewport for desktop pages, but apparently completely disables <meta> viewport handling).

Comment 7

[Chris Peterson [:cpeterson]]

Adding the [layout:triage-discuss] whiteboard tag so the Layout team can triage this bug. It would be nice to fix this for Fenix MVP (GV 68) because the page layout for https://www.google.com/url?q=https://example.com is readable in Chrome/WebView, but not Gecko.

Here is the WebView/Gecko side-by-side screenshot comparison: https://bugzilla.mozilla.org/attachment.cgi?id=9028780

Comment 8

[Jan Henning [:JanH]]

I start feeling like a broken record, but turning on font inflation (which would also happens implicitly when turning on automatic font size adjustment) would obviate any need to mess around with viewport widths, which has the potential to mess up the layout of any somewhat complex desktop/non-mobile page [1].

[1] To be fair, our font inflation implementation isn't perfect and occasionally messes up certain pages, but on average I think it still fares better than using a very narrow mobile viewport width for displaying pages written with Desktop only in mind.

Comment 9

[Imanol Fernandez]

We are having some issues in Firefox Reality related to the default viewport.

We need to use a Desktop version in Youtube to get VR videos and a quality selector. No matter the window GV size is adjusted to, YouTube always displays the 980px layout making text small, hard to read and interact with.

Oculus VR Browser (based on Chrome) doesn't have a fixed desktop viewport width, so the viewport & layout adapts to the size of the browser window, and page looks better.

For Firefox Reality we need something such as defaulting to device-width or a API to dynamically change the DesktopViewportWidth when the browser resizes.

FWIW This is a test page to easily test the viewport values: https://cvan.io/d/ In Oculus VR Browser the viewport changes each time the browser is resized. In FxR the width is always stuck at 980px.

Comment 10

[Jan Henning [:JanH]]

In that case it might be an idea to provide an additional GeckoSessionSettings viewport mode flag (VIEWPORT_MODE_FORCE_MOBILE or something like that), which when enabled would treat a missing <meta viewport> tag as "width=device-width". That would presumably require a new flag in Gecko similar to the desktop mode flag and could then possibly be wired into this function, i.e. return true there when the new flag is set on the window.

Alternatively, if the above doesn't give you what you want and you want to make browser.viewport.desktopWidth configurable in GeckoView anyway, don't forget that currently that pref controls both the fallback viewport for pages without a <meta viewport> tag as well as the forced desktop mode (VIEWPORT_MODE_DESKTOP) width. Maybe for Firefox Reality that would still be fine, but for a normal browser that means that manipulating that value wouldn't really be useful, as setting a narrower width there would make Desktop mode less useful at the same time. So in that case it might be necessary to separate those two cases and provide separate preferences for them in Gecko.

However I still maintain that Firefox Reality is something of a special case here, and for a browser on a normal phone (with a correspondingly narrow screen width in portrait mode) the current values should be fine, especially in conjunction with GeckoRuntimeSettings.setAutomaticFontSizeAdjustment(). Plus like I already said, GeckoView's current viewport behaviour plus setAutomaticFontSizeAdjustment() turned is the same behaviour as in Chrome.

Comment 11

[Sean Voisen (:svoisen)]

@Hiro: Can you take a look at this and add some of your thoughts?

Comment 12

[Hiroyuki Ikezoe (:hiro)]

I don't recommend to use device-width for the fallback value since it will break some sites. For example, youtube.com will break on 360px width devices. Making the fallback value configurable sounds more reasonable but still we need to carefully choose the value, given that youtube.com will break on 360px width and the value will depend on each site. Also, the value should be changed when the device is rotated.

So, though I am not familiar with out font inflation handling, tweaking it for no viewport meta tag sites sounds the best thing to do here. But I don't quite understand the font inflation doesn't inflate the font size on the google redirect pages.

Keep NI to me, I will debug why it doesn't work there.

Comment 13

[Jan Henning [:JanH]]

(In reply to Hiroyuki Ikezoe (:hiro) from comment #12)

So, though I am not familiar with out font inflation handling, tweaking it for no viewport meta tag sites sounds the best thing to do here. But I don't quite understand the font inflation doesn't inflate the font size on the google redirect pages.

Specifically regarding the Google redirect page - depending on the length of the URL, the amount of text on the page is just below our current line threshold (font.size.inflation.lineThreshold - 400 means 4 lines of square text, so a little less than two lines given the aspect ratio of a typical normal font), so it doesn't get inflated.

For very short URLs like the https://www.google.com/url?q=https://example.com example, the line threshold would have to be reduced to around 320, i.e. around 1½ lines of text. Historically (see bug 706193), this would have been too little for something like the footer on nytimes.com to remain uninflated. Of course nowadays the NY Times has a mobile layout and is no longer relevant, but I'm not sure in how far other similar layouts might still be around that would be negatively impacted by a reduced line threshold.

Comment 14

[Hiroyuki Ikezoe (:hiro)]

Thank you, Jan! I just arrived the place where we don't set nsFontInflationData::mInflationEnabled to true due to the text threshold.

(In reply to Jan Henning [:JanH] from comment #13)

(In reply to Hiroyuki Ikezoe (:hiro) from comment #12)

So, though I am not familiar with out font inflation handling, tweaking it for no viewport meta tag sites sounds the best thing to do here. But I don't quite understand the font inflation doesn't inflate the font size on the google redirect pages.

Specifically regarding the Google redirect page - depending on the length of the URL, the amount of text on the page is just below our current line threshold (font.size.inflation.lineThreshold - 400 means 4 lines of square text, so a little less than two lines given the aspect ratio of a typical normal font), so it doesn't get inflated.

For very short URLs like the https://www.google.com/url?q=https://example.com example, the line threshold would have to be reduced to around 320, i.e. around 1½ lines of text. Historically (see bug 706193), this would have been too little for something like the footer on nytimes.com to remain uninflated. Of course nowadays the NY Times has a mobile layout and is no longer relevant, but I'm not sure in how far other similar layouts might still be around that would be negatively impacted by a reduced line threshold.

dbaron, do you think we can reduce the threshold in these days? I guess, if we reduce it, we should apply the value only for the sites which has no viewport meta tag.

Comment 15

[Hiroyuki Ikezoe (:hiro)]

Chris, now I start thinking this is more likely the decision what GeckoView wants to do. Though the default value of the lineThreshold is 400, we can expose the preference to be able to be modified in GeckoView just like font.size.inflation.minTwips value. Is it sufficient for GeckoView? I am not sure what value is preferable, but you can easily confirm the results changing each value.

Comment 16

[Jan Henning [:JanH]]

I guess, if we reduce it, we should apply the value only for the sites which has no viewport meta tag.

As a rough approximation, don't forget that we'll only ever use font inflation on pages without a meta viewport tag anyway (well, technically a page that specifies an explicit viewport width as in "width=500" instead of using "width=device-width" may also get inflated if the width is large enough that we treat the page like a desktop page, too)

Comment 17

[Jan Henning [:JanH]]

(In reply to Hiroyuki Ikezoe (:hiro) from comment #15)

Chris, now I start thinking this is more likely the decision what GeckoView wants to do. Though the default value of the lineThreshold is 400, we can expose the preference to be able to be modified in GeckoView just like font.size.inflation.minTwips value. Is it sufficient for GeckoView? I am not sure what value is preferable, but you can easily confirm the results changing each value.

I'm not sure there's much point in making that value configurable, though - it's unlikely that we want to expose it as a user-facing option and I'm not sure how much embedders would actually want to play around with it, either - best to just have a sensible default value that works across whatever non-mobile-friendly pages are still commonly around today.

Comment 18

[Chris Peterson [:cpeterson]]

(In reply to Hiroyuki Ikezoe (:hiro) from comment #15)

Chris, now I start thinking this is more likely the decision what GeckoView wants to do. Though the default value of the lineThreshold is 400, we can expose the preference to be able to be modified in GeckoView just like font.size.inflation.minTwips value. Is it sufficient for GeckoView? I am not sure what value is preferable, but you can easily confirm the results changing each value.

Sorry, I don't know. I defer to dbaron's and Jan's recommendations here.

I agree with Jan that GeckoView should have a sensible default instead of a configurable value, but I don't know if our current default lineThreshold of 400 is still appropriate.

Comment 19

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

So I agree with JanH that messing with the default viewport width probably isn't a good idea. The 980px default is a characteristic of the web platform that's shared across engines; it's a part of how the web evolved from being desktop-only to working on mobile in a backwards-compatible way.

Although it's not shared across engines, I also think the font.size.inflation.lineThreshold pref is not really something that we want to be configurable. It's a characteristic of the font inflation algorithm: its purpose is to be a heuristic that distinguishes "large" chunks of text that are likely to be things that are expected to wrap across multiple lines in a context where the page layout adapts to the number of lines (the height) of the text, versus "small" pieces of text that are expected to occupy a specific place in the page layout that is less likely to be auto-height and less likely to respond well to the text size being increased (this might be a single word in a menu, or it might be two lines of text within a fixed-height container).

That said, 400 may not be the right number. It's entirely possible that, given the content on the web, the balance between the risks of having the number too high (failing to inflate smaller pieces of text) and the risks of having the number too low (inflated text being more likely to overlap other things because the page's layout didn't adjust to the text's size) call for having a different number.

(Some of the other font size inflation preferences, particularly font.size.inflation.emPerLine and font.size.inflation.minTwips, are designed to be configurable, since they essentially behave like a text size preference.)

(Take all of this with a bit of skepticism; it's been quite a while since I worked on this and I might be misremembering things.)

Comment 20

[Jan Henning [:JanH]]

This could probably be split off into a separate bug, but just to mention it here: I've discovered that Chrome tries to detect rows containing only links, i.e. typical navbars, which in turn might make it more feasible to decrease the general line threshold.

Comment 21

[Jan Henning [:JanH]]

Ignoring any discussions regarding the merits of viewport width vs. doing font inflation., it seems that this API might be required in order to use a different (larger in this case, actually) fallback width for tablets (bug 1564253).

Comment 22

[Jan Henning [:JanH]]

Or maybe not, depending on whether it would be preferable to just hardcode a second viewport width for use in tablets without any involvement from the front-end at all, or actually let the front end configure the precise value as per this bug, which could then e.g. also be used to set a different width in tablet mode. See also bug 1564253 comment 9.

Comment 23

[Jan Henning [:JanH]]

Attachment: Bug 1511177 - Part 1: Make desktop mode viewport width configurable in GeckoRuntimeSettings. r?snorp

Comment 24

[Jan Henning [:JanH]]

Attachment: Bug 1511177 - Part 0a: Fix Javadoc comment. r?snorp

Comment 25

[Jan Henning [:JanH]]

Attachment: Bug 1511177 - Part 0b: Keep changelog order consistent. r?snorp

It seems like we can't quite decide whether the change log for a version should
grow top-down or bottom-up. hg blame and the numbering of references seems to
somewhat favour the latter, though.

Comment 26

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/23282c683f67
Part 0a: Fix Javadoc comment. r=snorp
https://hg.mozilla.org/integration/autoland/rev/13004b37170e
Part 0b: Keep changelog order consistent. r=snorp
https://hg.mozilla.org/integration/autoland/rev/bbe9d07407c9
Part 1: Make desktop mode viewport width configurable in GeckoRuntimeSettings. r=snorp

Comment 27

[Daniel Varga [:dvarga]]

Backed out 3 changesets (Bug 1511177) for web-platform-tests-reftests at /css/css-tables/percent-width-cell-dynamic.html on a CLOSED TREE

Push with failure: https://treeherder.mozilla.org/#/jobs?repo=autoland&selectedJob=260643593&resultStatus=testfailed%2Cbusted%2Cexception&classifiedState=unclassified&revision=bbe9d07407c980a0274d8a11495bf84ef7dc3298

Failure log: https://treeherder.mozilla.org/logviewer.html#/jobs?job_id=260666783&repo=autoland&lineNumber=2365
and
https://treeherder.mozilla.org/logviewer.html#/jobs?job_id=260661777&repo=autoland&lineNumber=1563

Backout link: https://treeherder.mozilla.org/#/jobs?repo=autoland&selectedJob=260661777&resultStatus=testfailed%2Cbusted%2Cexception&classifiedState=unclassified&revision=6dd4ad534942b23dd66c46f20c9014cd48eb1219

[task 2019-08-08T21:40:50.618Z] 21:40:50     INFO -  REFTEST TEST-START | data:text/plain, == about:blank
[task 2019-08-08T21:40:50.618Z] 21:40:50     INFO -  REFTEST TEST-LOAD | data:text/plain, | 1 / 7767 (0%)
[task 2019-08-08T21:40:50.620Z] 21:40:50     INFO -  REFTEST TEST-UNEXPECTED-FAIL | data:text/plain, == about:blank | image comparison, max difference: 54, number of differing pixels: 2826
[task 2019-08-08T21:40:50.623Z] 21:40:50     INFO -  REFTEST   IMAGE 1 (TEST): data:image/png;base64,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
[task 2019-08-08T21:40:50.627Z] 21:40:50     INFO -  REFTEST   IMAGE 2 (REFERENCE): data:image/png;base64,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
[task 2019-08-08T21:40:50.627Z] 21:40:50     INFO -  REFTEST INFO | Saved log: START data:text/plain,
[task 2019-08-08T21:40:50.627Z] 21:40:50     INFO -  REFTEST INFO | Saved log: [CONTENT] OnDocumentLoad triggering AfterOnLoadScripts
[task 2019-08-08T21:40:50.628Z] 21:40:50     INFO -  REFTEST INFO | Saved log: Initializing canvas snapshot
[task 2019-08-08T21:40:50.628Z] 21:40:50     INFO -  REFTEST INFO | Saved log: DoDrawWindow 0,0,800,1000
[task 2019-08-08T21:40:50.628Z] 21:40:50     INFO -  REFTEST INFO | Saved log: [CONTENT] RecordResult fired
[task 2019-08-08T21:40:50.628Z] 21:40:50     INFO -  REFTEST INFO | Saved log: RecordResult fired
[task 2019-08-08T21:40:50.628Z] 21:40:50     INFO -  REFTEST INFO | Saved log: RecordResult fired
[task 2019-08-08T21:40:50.628Z] 21:40:50     INFO -  REFTEST TEST-END | data:text/plain, == about:blank
[task 2019-08-08T21:40:50.628Z] 21:40:50     INFO -  REFTEST TEST-START | data:text/plain,HELLO != about:blank
[task 2019-08-08T21:40:50.628Z] 21:40:50     INFO -  REFTEST TEST-LOAD | data:text/plain,HELLO | 2 / 7767 (0%)
[task 2019-08-08T21:40:50.628Z] 21:40:50     INFO -  REFTEST TEST-PASS | data:text/plain,HELLO != about:blank | image comparison, max difference: 254, number of differing pixels: 2988
[task 2019-08-08T21:40:50.629Z] 21:40:50     INFO -  REFTEST TEST-END | data:text/plain,HELLO != about:blank
[task 2019-08-08T21:40:50.629Z] 21:40:50     INFO -  REFTEST TEST-START | chrome://reftest/content/reftest-sanity/test-async.xul == chrome://reftest/content/reftest-sanity/test-async-ref.xul
[task 2019-08-08T21:40:50.629Z] 21:40:50     INFO -  REFTEST TEST-LOAD | chrome://reftest/content/reftest-sanity/test-async.xul | 3 / 7767 (0%)
[task 2019-08-08T21:40:50.629Z] 21:40:50     INFO -  REFTEST TEST-LOAD | chrome://reftest/content/reftest-sanity/test-async-ref.xul | 3 / 7767 (0%)
[task 2019-08-08T21:40:50.629Z] 21:40:50     INFO -  REFTEST TEST-PASS | chrome://reftest/content/reftest-sanity/test-async.xul == chrome://reftest/content/reftest-sanity/test-async-ref.xul | image comparison, max difference: 0, number of differing pixels: 0
[task 2019-08-08T21:40:50.629Z] 21:40:50     INFO -  REFTEST TEST-END | chrome://reftest/content/reftest-sanity/test-async.xul == chrome://reftest/content/reftest-sanity/test-async-ref.xul
[task 2019-08-08T21:40:50.629Z] 21:40:50     INFO -  REFTEST TEST-START | http://10.0.2.2:8854/tests/layout/reftests/reftest-sanity/test-async.html == http://10.0.2.2:8854/tests/layout/reftests/reftest-sanity/test-async-ref.html
[task 2019-08-08T21:40:50.629Z] 21:40:50     INFO -  REFTEST TEST-LOAD | http://10.0.2.2:8854/tests/layout/reftests/reftest-sanity/test-async.html | 4 / 7767 (0%)
[task 2019-08-08T21:40:50.630Z] 21:40:50     INFO -  REFTEST TEST-LOAD | http://10.0.2.2:8854/tests/layout/reftests/reftest-sanity/test-async-ref.html | 4 / 7767 (0%)
[task 2019-08-08T21:40:50.630Z] 21:40:50     INFO -  REFTEST TEST-PASS | http://10.0.2.2:8854/tests/layout/reftests/reftest-sanity/test-async.html == http://10.0.2.2:8854/tests/layout/reftests/reftest-sanity/test-async-ref.html | image comparison, max difference: 0, number of differing pixels: 0
[task 2019-08-08T21:40:50.630Z] 21:40:50     INFO -  REFTEST TEST-END | http://10.0.2.2:8854/tests/layout/reftests/reftest-sanity/test-async.html == http://10.0.2.2:8854/tests/layout/reftests/reftest-sanity/test-async-ref.html
[task 2019-08-08T21:40:50.630Z] 21:40:50     INFO -  REFTEST TEST-START | http://10.0.2.2:8854/tests/layout/reftests/reftest-sanity/test-zoom.html == http://10.0.2.2:8854/tests/layout/reftests/reftest-sanity/test-zoom-ref.html
[task 2019-08-08T21:40:50.630Z] 21:40:50     INFO -  REFTEST TEST-LOAD | http://10.0.2.2:8854/tests/layout/reftests/reftest-sanity/test-zoom.html | 5 / 7767 (0%)
[task 2019-08-08T21:40:50.630Z] 21:40:50     INFO -  REFTEST TEST-LOAD | http://10.0.2.2:8854/tests/layout/reftests/reftest-sanity/test-zoom-ref.html | 5 / 7767 (0%)
[task 2019-08-08T21:40:50.631Z] 21:40:50     INFO -  REFTEST TEST-UNEXPECTED-FAIL | http://10.0.2.2:8854/tests/layout/reftests/reftest-sanity/test-zoom.html == http://10.0.2.2:8854/tests/layout/reftests/reftest-sanity/test-zoom-ref.html | image comparison, max difference: 54, number of differing pixels: 7020

Comment 28

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/5a02186b16a5
Removing comment lines. On a CLOSED TREE DONTBUILD

Comment 29

[Daniel Varga [:dvarga]]

Sorry, wrong bug number, backed out.

Comment 30

[Pulsebot]

Backout by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/ad88c9455b0f
Backed out changeset 5a02186b16a5 for wrong bug number. On a CLOSED TREE DONTBUILD

Comment 31

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/0d8abfb9664d
Part 0a: Fix Javadoc comment. r=snorp
https://hg.mozilla.org/integration/autoland/rev/280c58a8f609
Part 0b: Keep changelog order consistent. r=snorp

Comment 32

[Razvan Maries]

https://hg.mozilla.org/mozilla-central/rev/0d8abfb9664d
https://hg.mozilla.org/mozilla-central/rev/280c58a8f609

Comment 33

[Jan Henning [:JanH]]

I'm not 100% sure, but still almost certain that this is because reftests have been traditionally using a desktop viewport width matching the size of the reftest window (800 px wide), plus I know at least one reftest that intentionally sets yet another viewport width that is neither the default value nor the "default" reftest width.
All this is conflicting with how GeckoView currently assumes that all prefs included in GeckoRuntimeSettings are under its complete control - as such the custom values set by the test harness are ignored and the tests are run using the default viewport width included in GeckoView.