Open Bug 975444 Opened 11 years ago Updated 1 year ago

google.com - Firefox Android doesn't receive the tier1 sent to Chrome Android

Categories

(Web Compatibility :: Mobile, defect)

Other
Android
defect

Tracking

(platform-rel +, firefox100 affected)

Tracking Status
platform-rel --- +
firefox100 --- affected

People

(Reporter: Swarnava, Unassigned)

References

(Depends on 1 open bug, Blocks 1 open bug, )

Details

(Keywords: webcompat:site-wait, Whiteboard: [testday-20140221][serversniff][sitewait][platform-rel-Google])

Attachments

(12 files)

Some searches is not showing quick info on google search result "Weather in san fransisco" "9w 617" "Go indigo 277" Works fine on chrome and uc browser.
Screenshot?
Status: NEW → RESOLVED
Closed: 11 years ago
Resolution: --- → WORKSFORME
Attached image Firefox version (deleted) —
Attached image Uc browser version (deleted) —
Thanks. Let's move this over to Evangelism.
Status: RESOLVED → REOPENED
Component: General → Mobile
Product: Firefox for Android → Tech Evangelism
Resolution: WORKSFORME → ---
Summary: Google.com is not showing properly in Firefox → Google flight status cards not shown in Firefox for Android
Blocks: google.com
Attached image 2014-02-24 19.49.15.png (deleted) —
Screenshot while spoofing as Android.
So yeah, if we pretend to be a webkit browser we get the interactive cards (which work nicely). However, as you can see from the screenshot the layout of the search box/button is a little off. Looking at the inline CSS, it's mostly -webkit- only stuff as you might expect. Karl, if we gave CSS suggestions to Google do you think they would incorporate them (and then flip the bit that allows us to have the fancier search page)?
Whiteboard: [testday-20140221] → [testday-20140221][serversniff]
Mike, we can try, but probably not.
Flags: needinfo?
Status: REOPENED → NEW
Whiteboard: [testday-20140221][serversniff] → [testday-20140221][serversniff][contactready]
Contacted Alex.
Assignee: nobody → kdubost
Whiteboard: [testday-20140221][serversniff][contactready] → [testday-20140221][serversniff][sitewait]
All cards in general not really shown
Summary: Google flight status cards not shown in Firefox for Android → Google status cards not shown in Firefox for Android
Summary: Google status cards not shown in Firefox for Android → Google cards not shown in Firefox for Android
Same for Firefox OS.
Status: NEW → ASSIGNED
Attached image chrome-card.png (deleted) —
new screenshot with spoofing as chrome mobile on Firefox No issue anymore. Basically Google could just send what they send to Chrome Mobile to Firefox Android and Firefox OS.
No change on this one with Android version number. We don't get the same layout than chrome. Irony. Spoofing as Chrome, we get a better working site.
Attached image firefox-spoofing-as-chrome.png (deleted) —
What it looks like when spoofing as chrome.
putting myself in ni, for coming back to this bug ASAP.
Flags: needinfo?(kdubost)
Priority: -- → P3
Priority: P3 → P1
Priority: P1 → P2
OK I spent a bit of time with this issue today. Looking at 1. Google.com on Chrome Android 2. Google.com on Firefox Android 3. Google.com on Firefox Android with Chrome Android UA string. I didn't see anything which would prevent Google to send us the same thing that they are sending to Chrome. Let me contact Google again about this issue.
Flags: needinfo?(kdubost)
Contacted Google today about it.
platform-rel: --- → ?
Whiteboard: [testday-20140221][serversniff][sitewait] → [testday-20140221][serversniff][sitewait][platform-rel-Google]
platform-rel: ? → +
* images.google.com https://webcompat.com/issues/1358 Tested and working with a Chrome UA on a mobile device with Firefox Android. I will ask Google to send us again tier1 on Firefox Android. If there are any current bugs on Bugzilla which are related to a sub-par User experience, add this bug as a blocker.
Summary: Google cards not shown in Firefox for Android → google.com - Firefox Android doesn't receive the tier1 sent to Chrome Android
sent an email to ask if we could discuss f2f about Tier1 blockers for Firefox Android. Summary: * Google is sending to Firefox Android an inferior version (not the one Chrome receives) * The first contact was made with Google on April 22, 2014 (3 years ago) to ask them to send a compatible version. Alex Komoroske (Google) routed this internally. * During 2015/2016, we started implementing and aliasing some -webkit- and WebKit properties to be compatible with the Web. * February 2016 Testing http://www.otsukare.info/2016/02/01/testing-google-on-firefox-mobile * On April 27, 2016, Dru Knox (Google) proposed we wait on the final implementation of -webkit- prefixes in Firefox. * On January 11, 2017, I asked Dru if we could receive the tier1 sent to Chrome now that we had a solid implementation. no answer * On February 15, 2017, I asked again for the status to Dru and Rick Byers (Google) no answer * On March 17, 2017, Andreas Bovens (Mozilla) asked if he could help with contacts. (sure welcome) * On April 11, 2017, I'm proposing Dru and Rick to meet in June/July for this issue so we better understand the current blockers.
Attached image Before Google site specific UA change (deleted) —
To be honest though, it doesn't look like Google really cares at all. The end result being that users get a sub-par experience. I propose that while Google (possibly) solves this on their end (if ever), we can ship Firefox with a UA override preference for Google websites only. Here's an example: 1. Go to about:config 2. Create new entry 3. Set name to "general.useragent.override.encrypted.google.com" (can be google.com too) 4. Select type as String 5. Enter Chrome Android's UserAgent: Mozilla/5.0 (Linux; <Android Version>; <Build Tag etc.>) AppleWebKit/<WebKit Rev> (KHTML, like Gecko) Chrome/<Chrome Rev> Mobile Safari/<WebKit Rev> Done! I am attaching before and after results. Note that there are add-ons for Android on the add-ons store like these (https://addons.mozilla.org/en-US/firefox/addon/chrome-ua-on-google-android/) that specifically do exactly the same thing for Google only websites. We should also note that it's not just Google search, but other Google websites also get sub-par experiences on Firefox because of Google's behaviour. What do you guys think?
Attached image After Google site specific UA change (deleted) —
(In reply to uncertainquark from comment #26) > To be honest though, it doesn't look like Google really cares at all. The > end result being that users get a sub-par experience. I propose that while > Google (possibly) solves this on their end (if ever), we can ship Firefox > with a UA override preference for Google websites only. > What do you guys think? So, this is a double-edged sword -- it gets it fixed, but then we disappear entirely from Google's radar. We're in discussions with Google to see what we can do to get this fixed. Hopefully we'll have some news in the near future. However, if they're unwilling to serve us tier 1 content, going the override route is on the table.
Priority: P2 → P1
Changing the User-Agent causes problems on Google+ on Firefox Android: image loading is broken.
(In reply to pedrogfrancisco from comment #29) > Changing the User-Agent causes problems on Google+ on Firefox Android: image > loading is broken. This is likely because they're sending WebP images while Firefox does not support that currently. If Safari ships WebP, Firefox probably will too. https://www.cnet.com/news/apple-ios-macos-tests-googles-webp-graphics-to-speed-up-web/
What is the current status?
Flags: needinfo?(kdubost)
Marco, for now, unchanged. We (Mozilla+Softvision) have tested plenty of Google properties with Gecko and a Chrome UA on Android (so receiving tier1 on Gecko). And this creates multiple issues. You can see some of these in https://webcompat.com/issues?page=1&per_page=50&state=open&stage=all&sort=created&direction=desc&q=label%3Atype-uaoverride+google+in%3Atitle
Flags: needinfo?(kdubost)
(In reply to Karl Dubost :karlcow from comment #32) > Marco, for now, unchanged. We (Mozilla+Softvision) have tested plenty of > Google properties with Gecko and a Chrome UA on Android (so receiving tier1 > on Gecko). And this creates multiple issues. > > > You can see some of these in > https://webcompat.com/ > issues?page=1&per_page=50&state=open&stage=all&sort=created&direction=desc&q= > label%3Atype-uaoverride+google+in%3Atitle Thanks Karl, so I guess no luck with the meeting with Google. Do we know what the blockers are? Have you considered a UA override restricted to sites that are pretty bad in Firefox (such as images.google.com)?
Marco, not that much a question of luck. The Webcompat meeting with Google was good. We had fruitful discussions and a better understanding. The next step I would say is on us to better understand the full categories of issues. Just send this email about it: https://groups.google.com/forum/#!topic/mozilla.compatibility/ecCx7Ht-zFw
Depends on: 1421481
https://webcompat.com/issues/10398 This one is more about tier1 tablet instead of mobile but same issue.
Depends on: 976749
Depends on: 1200652
Depends on: 976753
Depends on: 1437808
Firefox Focus can display full user interface for Google search. But firefox mobile can't! I ask to Google directly if they can change this they reply : ""Steven Leys Google Expert : Hi Alexander Depending on the webengine a browser uses, it will display the same web page, in a visually differently way. That’s largely a choice made by the developer of a browser - and has less to do with Google itself. Best Steven"" I think it's very important that firefox mobile can display full user interface in google serch. Due to the popularity of Google lot of user can prefer to use chrome due to degraded user interface of Firefox Mobile on Google Product.
Please also read this post on Google Forum : https://productforums.google.com/forum/#!topic/websearch/M6FHhgolYkg So it's a problem from Firefox Mobile !! Please also see attachement : https://bugzilla.mozilla.org/attachment.cgi?id=8957518
(In reply to Orlandini Alexandre from comment #39) > I think it's very important that firefox mobile can display full user > interface in google serch. Due to the popularity of Google lot of user can > prefer to use chrome due to degraded user interface of Firefox Mobile on > Google Product. People are well aware of the issue. It's google who doesn't send us the tier1 page for a variety of web-compatibility issues, it seems. Firefox focus isn't affected because it uses Android's webview instead of Gecko, that is, it's a WebKit-based browser.
Maybe is it possible to spoof user agent to Firefox Focus just for google search to have full functionality ?
Attached image Issue resolved in Nightly 63? (deleted) —
Hello, it seems that Nightly 63 successfully load the new Google Search page. (From the screenshot you can see that I installed an user agent switcher add-on https://bugzilla.mozilla.org/show_bug.cgi?id=975444 which is currently inactive).
(In reply to reinhart_previano from comment #43) > Created attachment 8989043 [details] > Issue resolved in Nightly 63? > > Hello, it seems that Nightly 63 successfully load the new Google Search page. > (From the screenshot you can see that I installed an user agent switcher > add-on https://bugzilla.mozilla.org/show_bug.cgi?id=975444 which is > currently inactive). Yeah, that's due to bug 1453691, we're spoofing the UA IIUC. So no completely fixed, it was some sort of Nightly experiment.
(In reply to Emilio Cobos Álvarez (:emilio) from comment #44) > Yeah, that's due to bug 1453691, we're spoofing the UA IIUC. So no > completely fixed, it was some sort of Nightly experiment. Exactly. We're fixing some platform / interop bugs, and running a Nightly experiment to see what else surfaces.
https://bugzilla.mozilla.org/show_bug.cgi?id=1427500 Posting this issue here because, although Firefox should be intelligent enough to prioritize/default to the HTTPS version of google.com, it's quite shocking that google doesn't automatically redirect to the HTTPS version of google.com. I'm surprised there is still a non-secure google.com in existence! I'm sure there is a decent amount of Firefox for Android searches going through the insecure HTTP version of google.com and that is a problem. I really hope this gets fixed soon
I’ve been enjoying the UA spoof with 63.0a1, but as of today the search results page is doing some new nonsense. Instead of search results, it only displays this message: > You’re offline > Check your connection and try again > [⮠ Try again] Pressing “Try again” does nothing, sadly.
Thanks for the report, tigt! We've seen at least one report of this on webcompat.com, but so far haven't been able to reproduce it (the closest we've been able to reproduce it so far has been to see a "not secure" page that went away with a quick page refresh or two, but didn't seem to be related). A couple of questions: - if you haven't, could you try disabling the spoofing experiment to confirm if it's the issue? (about:config flag extensions.gws-and-facebook-chrome-spoof.enabled) - does a hard-refresh of the page help? (long-tap the refresh button in the Firefox menu). - do you use any other addons, and if so does disabling for a quick test seem to help? We will probably want to file a new bug for this one as well.
ni? tigt for the questions in Comment #48.
Flags: needinfo?(tigt)
A hard refresh did indeed fix it. I would assume disabling the spoofing experiment also would, since the HTML and JS sent down when it's off are completely different. However, I can get back in the state by changing the value in the on-page search field and submitting again. Disabling all my add-ons provides no change. I can hook up Web Inspector if you like and start poking around in the values stored by what I assume is a Service Worker, otherwise I'm not sure what their offline indicator is supposed to accomplish.
I have the same issue. Disabling the experiment does not help. Initiating a search (whether through the address bar or the result page's form) shows a "You're offline" error in the new design. Hard-refreshing shows the proper result page in the old design. The error is also shown when I'm actually offline so maybe something bad is stuck in the cache?
Flags: needinfo?(tigt)
(In reply to Jan Alexander Steffens [:heftig] from comment #51) > I have the same issue. Disabling the experiment does not help. Initiating a > search (whether through the address bar or the result page's form) shows a > "You're offline" error in the new design. Hard-refreshing shows the proper > result page in the old design. Thanks for the info heftig! Do you have reliable STR? Or it happened once, and a hard-refresh fixed it? With reliable STR, it would be super interesting to run mozregression to see if a Gecko change or a google.com change is the likely culprit.
Flags: needinfo?(jan.steffens)
Unfortunately, no. The issue persistently affected all searches, with a hard-refresh being needed each time. I switched to a guest session, which was unaffected. After leaving the guest session, the issue also disappeared on the normal profile and I was getting proper results in the old design. The issue did not reappear after reenabling the spoofing. I can no longer reproduce.
Flags: needinfo?(jan.steffens)
Depends on: 1479882
I was getting the same "You're offline" error. Clearing 'Offline website data' fixed it in my case. I'm guessing google puts something in web storage that works on Chrome but not ff. Unfortunately since I'm no longer seeing it I can't check what it was in the storage.
I wonder for all the people where it failed if they were logged in or not (on their Google Account).
In case this helps: my mobile devices do indeed receive the tier1 page but none of my tablets do.

GMail is also getting a different tier on Fennec, as per https://webcompat.com/issues/24626

Product: Tech Evangelism → Web Compatibility

Firefox 67al for Android indicates old google search.

What is the current status of this situation? In July 2018, Google told ZDNet that it is in the process of testing a "new experience" for Firefox users that will take care of the issues. See here https://www.zdnet.com/article/firefox-and-the-4-year-battle-to-have-google-to-treat-it-as-a-first-class-citizen/ Surely this should be resolved soon?

(In reply to Will from comment #60)

What is the current status of this situation? In July 2018, Google told ZDNet that it is in the process of testing a "new experience" for Firefox users that will take care of the issues. See here https://www.zdnet.com/article/firefox-and-the-4-year-battle-to-have-google-to-treat-it-as-a-first-class-citizen/ Surely this should be resolved soon?

They did ship a new experience called "Hierloom". It's not bad... just not as nice as the Chrome Mobile experience. IMO, the real issue is that beyond search, the UI is really bad -- for example, searching for images. We'll keep talking with Google, but there's not a lot we can do to force the UI without potentially breaking search for our users.

But isn't Google the default search engine for Firefox in most regions? You would think this bug/issue would have been brought up when that deal was being worked out.

(Also, the google.com homepage still looks like it's from 2009 in Firefox for Android even though the "ALL" web search results tab does look modern)

See bug 1547409. Moving webcompat whiteboard tags to keywords.

It seems that the image search for Fennec/mobile Gecko has been improved a little, in that instead of the blurry, cropped-to-a-fixed-aspect-ratio results I'm now seeing the images in a reasonable resolution and in their proper aspect ratio, plus we now seem to have infinite scrolling, too.

It's still not the fully-featured version, though, because the additional search filtering bar (most recent, GIF, HD, by colour, etc.) is still missing and if you're searching an image URL, you're not offered the choice of reverse image search, either.

Apologies if it's already submitted (I can't check all the webcompats listed here) but I'm seeing dark theme on Chrome Canary, Dev and Beta, but not in any version of Fenix. Can anyone else repro?

Attached image Screenshot_20210307-212747.png (deleted) —
Attached image advanced search screen.png (deleted) —

This one really hurts :( no "Advanced search" screen when using Fenix. Available in Chrome of course. (you can copy the URL from Chrome and paste it in Fenix, but the webpages bottom bar is for some reason located in the middle of the page covering the "region" and "last update" fields. Also, its zoomed out a lot and you have to zoom in). Apologies if this is a duplicate

Another nice improvement coming to google.com for mobile browsers was announced today, I very much doubt Firefox mobile users will get it though :( https://blog.google/products/search/continuous-scrolling-mobile/

The issue is still reproducible:

https://prnt.sc/q7M7gKwGMjDW

Tested with:

Browser / Version: Firefox Nightly 100.0a1 (2015869419 -🦎100.0a1-20220317092857🦎)/ Chrome Mobile Version 99.0.4844.73
Operating System: Samsung A51 (Android 11) -1080 × 2400 pixels 20:9 aspect ratio (~405 ppi density)

Sorry for the ping but I read someone say this issue finally will be addressed by the new EU regulations the will almost certainly be coming into effect soon. Is this true? I hope so...

Flags: needinfo?(dschubert)

Unassigning the priority here to reflect reality.

Priority: P1 → --

Sorry, there was a problem with the detection of inactive users. I'm reverting the change.

Assignee: nobody → karl+moz
Status: NEW → ASSIGNED

Actually Karl isn't with Mozilla anymore, so it might as well be left unassigned for now.

Assignee: karl+moz → nobody
Status: ASSIGNED → NEW
Severity: normal → S3

The severity field for this bug is relatively low, S3. However, the bug has 11 duplicates, 55 CCs and 101 See Also bugs.
:denschub, could you consider increasing the bug severity?

For more information, please visit auto_nag documentation.

Flags: needinfo?(dschubert)

This bug is more of a "tracking" issue. We know what's happening here, and there isn't much we can do either way - so setting a higher priority won't change anything.

Flags: needinfo?(dschubert)

Dennis, the severity is a measure of the effects of the bug on the user base, rather than a measure of priority (even if, of course, priority depends on severity). This seems like a bug with high impact, so S2 would seem more fitting.

You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: