1225414 - incorrect user-agent string for ESR-channel build

Status: RESOLVED WONTFIX

(Firefox :: General, defect)

Description

[Ikonta]

User Agent: Mozilla/5.0 (X11; Linux x86_64; rv:38.0) Gecko/20100101 Firefox/38.0
Build ID: 20151110093942

Steps to reproduce:

I use stable (ESR-channel) FireFox browser.
For now it is =www-client/firefox-38.4.0.


Actual results:

Looking its User-Agent string in about:support or in server's logs I see incorrect version info:
User Agent
Mozilla/5.0 (X11; Linux x86_64; rv:38.0) Gecko/20100101 Firefox/38.0




Expected results:

I expect to see at least correct version:
User Agent
Mozilla/5.0 (X11; Linux x86_64; rv:38.4) Gecko/20100101 Firefox/38.4

Maybe, extended with non-default ("Release") channel's name:
User Agent
Mozilla/5.0 (X11; Linux x86_64; rv:38.4 ESR) Gecko/20100101 Firefox/38.4

Comment 1

[Sebastian Hengst [:aryx] (needinfo me if it's about an intermittent or backout)]

This might be by design to reduce fingerprinting.

Comment 2

[Sebastian Hengst [:aryx] (needinfo me if it's about an intermittent or backout)]

From https://dxr.mozilla.org/mozilla-central/rev/a2f83cbe53ac4009afa4cb2b0b8f549289b23eeb/python/mozbuild/mozbuild/milestone.py#58

>        # Only expose the major milestone in the UA string, hide the patch
>        # level (bugs 572659 and 870868).
>        uaversion = "%s.0" % (get_milestone_major(milestone),)
See bug 572659 (X.Y.0) and bug 870868 (X.0).

Comment 3

[Ikonta]

(In reply to Sebastian H. [:aryx][:archaeopteryx] from comment #2)
> See bug 572659 (X.Y.0) and bug 870868 (X.0).

Thank you!

But subject of bug #870868 looks at least strange: why _freeze_ zero value instead of simple removing pre-defined constant?
Adding missed, but nowadays necessary channel mark.

Comment 4

[YF (Yang)]

(In reply to Ikonta from comment #3)
> (In reply to Sebastian H. [:aryx][:archaeopteryx] from comment #2)
> > See bug 572659 (X.Y.0) and bug 870868 (X.0).
> 
> Thank you!
> 
> But subject of bug #870868 looks at least strange: why _freeze_ zero value
> instead of simple removing pre-defined constant?
> Adding missed, but nowadays necessary channel mark.

You mean like "rv:38)"? It is possible to break many existing recognition mode.

Comment 5

[Ikonta]

(In reply to YF (Yang) from comment #4)
> > But subject of bug #870868 looks at least strange: why _freeze_ zero value
> > instead of simple removing pre-defined constant?
> > Adding missed, but nowadays necessary channel mark.
> 
> You mean like "rv:38)"? It is possible to break many existing recognition
> mode.

As for now, recodnition modes works correctly only for Release channel, not for ESR.
Thatis why recognition mode algorythm should be reviewed at least for FireFox.

I mean usage of two values: "rel:38)" and "esr:38)".
Ideally with some optional fiels to validate up-to-date state of ESR version.

Comment 6

[YF (Yang)]

I think the website should be not needs to distinguish whether the user is ESR, the features should be no differents with the stable version.

Comment 7

[Sebastian Hengst [:aryx] (needinfo me if it's about an intermittent or backout)]

For Firefox 45, Service Workers will be enabled in the fast track release and turned off for the ESR release: https://groups.google.com/d/msg/mozilla.dev.platform/yuNHtDhl3lY/FNYiDh2YAgAJ

Comment 8

[Justin Dolske [:Dolske]]

I believe this is working as intended.

User Agent string changes are notorious for causing web compat issues, so changing them without a significant benefit is likely to just be a net loss for users due to sites breaking. (Without looking, I'd assume that's why it's "frozen" to .0 for ESR -- ESR users are a tiny minority of Firefox users, and the web is full of dumb scripts and regexs that likely assume the version is always .0 and may break if it's not.)