Ubuntu web browsers do not properly identify host

As a consequence of not identifying their host OS properly, Ubuntu browsers fail to identify it correctly in all cases but one (stock Firefox).

As a result, 1/3 or more of Ubuntu users are not correctly recognized by online statistics.

There are two possible solutions:
Properly, promptly, and consistently package the latest versions of every popular browser for Ubuntu with correct useragents, or
Get upstream browser developers to use uname to set host OS useragent information.

ok, i installed chromium, and checked the user agent reported to websites:
Mozilla/5.0 (X11; U; Linux i686; en-US) AppleWebKit/533.2 (KHTML, like Gecko) Chrome/5.0.342.9 Safari/533.2

which looks fine to me...

i cant try firefox since i already have a self-built trunk installation, which also reports:

Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.3a5pre) Gecko/20100502 Firefox/3.7a5pre GTB7.1

what exactly is it that you notice is incorrect in your browsers' user-agent?
i agree that the U is kinda cryptic, but U is for Ubuntu.

Mozilla/5.0 (X11; U; Linux i686; en-ca) AppleWebKit/531.2+ (KHTML, like Gecko) Safari/531.2+ Debian/squeeze/sid () Epiphany/2.30.2

..Blog sites identify my Epiphany as running on Debian, not Ubuntu.

Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.2.3) Gecko/20100423 Ubuntu/10.04 (lucid) Firefox/3.6.3

..and that's stock Firefox.

The U does not mean Ubuntu.

Web browsers created in the United States, such as Netscape Navigator and Internet Explorer, use the letters U, I, and N, to specify the encryption strength in the user agent string. Until the US government allowed encryption with keys longer than 40 bits to be exported, in 1996, vendors shipped various browser versions with different encryption strengths. "U" stands for "USA" (for the version with 128-bit encryption), "I" stands for "International"--the browser has 40-bit encryption and can be used anywhere in the world--and "N" stands (de facto) for "None" (no encryption).[5] Following the lifting of export restrictions, most vendors supported 256-bit encryption.

http://en.wikipedia.org/wiki/User_agent

http://images.cheezburger.com/completestore/2010/5/22/129190627850293860.png

Since Ubuntu users are early adopters, I believe this bug affects around 30-40% of us. I've used firefox prereleases and epiphany for years, and even my mom is an avid Chrome fan. I notice more geeks using Chrome these days than I notice using Firefox.

This bug, by way of stastistics-collecting sites, is now deterring development of proprietary applications for ubuntu, and thereby its end-user adoption, on an absolutely colossal scale by making the Ubuntu market appear significantly smaller than it actually is.

Based on upstream Mozilla comments, I think we're going to mark this Won't Fix as well. Here's the document linked from the upstream Mozilla Bug about Mozilla User Agents, the concerns involved, and the relevant RFCs: https://developer.mozilla.org/en/User_Agent_Strings_Reference
I think Chromium and Epiphany would have similar ideas about Security/Privacy concerns.
Please report any other issues you may find.
I'll continue to monitor this in case Epiphany and/or Chromium decide differently.

That only means it can't be fixed *properly*, so the Ubuntu packages for these browsers must all be modified. I've done a little bit of packaging before, so I'll have to try and see if I can figure that out or hire someone who can.

First of all the Firefox builds that we make already have Ubuntu in
them since they're *our* builds. As for the other browsers, we should
follow the conventions set by upstream for their User Agents. User
agents are not made for what you want to do.

On 05/24/2010 01:55 PM, ethana2 wrote:
> That only means it can't be fixed *properly*, so the Ubuntu packages for
> these browsers must all be modified. I've done a little bit of
> packaging before, so I'll have to try and see if I can figure that out
> or hire someone who can.
>

Please tell me what useragents are for.

A user agent is to tell the web server what type of browser is
requesting content so that appropriate content can be returned.

Here are the related RFCs:
http://www.ietf.org/rfc/rfc1945.txt
http://www.ietf.org/rfc/rfc2068.txt

On 05/25/2010 10:24 PM, ethana2 wrote:
> Please tell me what useragents are for.
>

Then it is important that my browser identifies my OS as Ubuntu.
To see this in proper use, see:

google.com/chrome

I would consider my seeing a .dmg, .rpm, or .msi on that page to be inappropriately served content.

It's not about the OS, but the platform. This page correctly shows me
the page for Linux.

On 05/25/2010 11:52 PM, ethana2 wrote:
> Then it is important that my browser identifies my OS as Ubuntu. To
> see this in proper use, see:
>
> google.com/chrome
>

Are you seeing that? I tried in 4 different browsers (Firefox,
Epiphany, Arora, Konqueror) and all show me the Linux page.

On 05/25/2010 11:53 PM, ethana2 wrote:
> I would consider my seeing a .dmg, .rpm, or .msi on that page to
> be inappropriately served content.
>

Because useragents are currently unreliable, they have to rely on a *second* page asking the *user* to identify their platform and choose their executable using a radio button. We can fix that.

If you think "linux" is the proper way to identify a platform, I trust you'll help me get stock Firefox and Epiphany (which identify as Ubuntu and Debian, respectively) fixed to that effect?

Again, the point of a user agent is for rendered content to be proper,
not necessarily downloadable content. Google is doing this correctly.

On 05/26/2010 12:20 AM, ethana2 wrote:
> If you think "linux" is the proper way to identify a platform, I
> trust you'll help me get stock Firefox and Epiphany (which identify
> as Ubuntu and Debian, respectively) fixed to that effect?
>

Keep in mind the most popular linux variant very soon will be Google Android.

The kernel actually means very little in terms of what the platform actually is nearly irrelevant in determining what content is appropriate to return.

If you want this fixed, you need to file an RFC for it. This is not
the purpose of the user agent as specified in the RFCs that I provided.

On 05/26/2010 12:18 AM, ethana2 wrote:
> Because useragents are currently unreliable, they have to rely on a
> *second* page asking the *user* to identify their platform and choose
> their executable using a radio button. We can fix that.
>

I don't mind paying someone to fix this myself, but would I have to file an RFC in order to get the fixed packages accepted into Ubuntu main and universe repositories, while stock Firefox *already* has useragent information that doesn't fall within the RFCs you linked here?

Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.3a5pre) Gecko/20100502 Firefox/3.7a5pre GTB7.1

this is from a firefox compiled from source without any source modifications. a *stock* firefox as you may call it. i dont see it listing ubuntu anywhere.

Not wanted in Chromium by upstream. I wrote a patch once, they rejected it.

It seems that Epiphany Browser in Ubuntu is affected by this and is the only app with an issue here. Therefore, I am assigning this task to that package. I will make a patch for this.

This particular bug has already been fixed and is a duplicate of bug 578717, so it is being marked as such. Please look at the other bug report to see if there is any missing information that you can provide, or to see if there is a workaround for the bug. Additionally, any further discussion regarding the bug should occur in the other report. Please continue to report any other bugs you may find.