-
-
Save cramforce/63a8f25639b0201a7c48c90f7a4e1eba to your computer and use it in GitHub Desktop.
We are seeing this Firefox bug (or a variant of it) get triggered https://bugzilla.mozilla.org/show_bug.cgi?id=548397 | |
This showed up in our error reporting and the stack trace format is clearly a Gecko stack trace. | |
Example URL https://m-wikihow-com.cdn.ampproject.org/v/m.wikihow.com/Dress-for-Skiing?amp=1&_js_v=7 embedded as an iframe on https://www.google.com | |
The UAs are sometimes: | |
Mozilla/5.0 (Linux; Android 6.0.1; SAMSUNG SM-G920F Build/MMB29K) AppleWebKit/537.36 (KHTML, like Gecko) SamsungBrowser/4.0 Chrome/44.0.2403.133 Mobile Safari/537.36" | |
And more often: | |
Mozilla/5.0 (iPhone; CPU iPhone OS 9_1 like Mac OS X) AppleWebKit/601.1.46 (KHTML, like Gecko) Version/9.0 Mobile/13B143 Safari/601.1" |
That is the point. But we know that the browser is Firefox, so we are wondering why a lot of Firefoxes use a non-Firefox UA.
Our theories:
- Firefox for Android cloaks itself when accessing www.google.com
- A bot based on Firefox
- A popular Firefox extension generates those UAs.
@cramforce I must be thick. but how do you know it's Firefox. You said:
stack trace format is clearly a Gecko stack trace
What do you see in your logs? An example? For example on the Chromium bug, it shows it had been deployed at a point that would create the same issue than Firefox.
When you say a lot, you mean more than Firefox Android UA usage? In this case I doubt it is something coming from Mozilla.
The UA override list for Firefox Android is public
https://hg.mozilla.org/mozilla-central/raw-file/tip/mobile/android/app/ua-update.json.in
- It doesn't have the two user-agents you gave
- it doesn't operate on google.com
Before b2g (aka Firefox OS) went RIP, the UA override list was
https://dxr.mozilla.org/mozilla-central/source/b2g/app/ua-update.json.in
and didn't contain google.com either.
For all UA overrides with Google and services, we directly negotiate with google on a specific mailing-list. So that would have been known if there was an issue.
All of that said, I wish I understand better your issue on how to help you.
The reason why we know it is Firefox is because the error report we receive uses Firefox's stack trace format, which is distinctly different from Chrome and WebKit. We don't really have an immediate problem. The issue is clear in terms of how we can fix it (although we might have found a new bug in Firefox), we were just wondering why we see these UAs, as that drove our debugging session into the wrong direction.
These are not Firefox UA. :)
Maybe you should check
whatwg/html#1813
And
https://bugs.chromium.org/p/chromium/issues/detail?id=650433
and
https://bugs.webkit.org/show_bug.cgi?id=107236