Hi...

I think because some are ESR (Extended Service Release) builds. Thus, it is possible to have a Firefox 10 build which is newer than a Firefox 17 build, but with an older gecko.

Unless I missed something in the xml referring to FF 20 or so with a 2010 gecko, which I would not expect...
Back to top

Some New Browser UAs with Old dates:
<useragent description="Firefox 20.0 (OS X 10.7 Intel)" useragent="Mozilla/5.0 (Macintosh; Intel Mac OS X 10.7; rv:20.0) Gecko/20100101 Firefox/20.0" appcodename="" appname="" appversion="" platform="" vendor="" vendorsub=""/>
<useragent description="Firefox 19.0 (Slackware 13 32 bit)" useragent="Mozilla/5.0 (X11; U; Linux i686; rv:19.0) Gecko/20100101 Slackware/13 Firefox/19.0" appcodename="" appname="" appversion="" platform="" vendor="" vendorsub=""/>
<useragent description="Firefox 20.0 (Ubuntu 64 bit)" useragent="Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:20.0) Gecko/20100101 Firefox/20.0" appcodename="" appname="" appversion="" platform="" vendor="" vendorsub=""/>
<useragent description="Firefox 20.0 (32 bit)" useragent="Mozilla/5.0 (X11; Linux i686; rv:20.0) Gecko/20100101 Firefox/20.0" appcodename="" appname="" appversion="" platform="" vendor="" vendorsub=""/>

Some Old browser UAs with Newer dates:
<useragent description="Firefox 3.8 (Ubuntu/9.25)" useragent="Mozilla/5.0 (X11; U; Linux i686; pl-PL; rv:1.9.0.2) Gecko/20121223 Ubuntu/9.25 (jaunty) Firefox/3.8" appcodename="" appname="" appversion="" platform="" vendor="" vendorsub=""/>
<useragent description="Firefox 11.0 (32 bit)" useragent="Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.1.16) Gecko/20120421 Gecko Firefox/11.0" appcodename="" appname="" appversion="" platform="" vendor="" vendorsub=""/>

Some Old Browsers UAs with Same 20100101 date:
<useragent description="Firefox 12.0 (32 bit)" useragent="Mozilla/5.0 (X11; Linux i686; rv:12.0) Gecko/20100101 Firefox/12.0 " appcodename="" appname="" appversion="" platform="" vendor="" vendorsub=""/>
<useragent description="Firefox 14.0.1 (Ubuntu 64 bit)" useragent="Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:14.0) Gecko/20100101 Firefox/14.0.1" appcodename="" appname="" appversion="" platform="" vendor="" vendorsub=""/>
<useragent description="Firefox 16.0 (32 bit)" useragent="Mozilla/5.0 (X11; Linux i686; rv:16.0) Gecko/20100101 Firefox/16.0" appcodename="" appname="" appversion="" platform="" vendor="" vendorsub=""/>

UAs do not look authentic, i.e, not directly from actual browser. The version number is simply incremented or constructed and date of original base UA has been carried forward.

*Note: I only wrote a few i could screen in one go, there might be more such incremented/constructed UAs
Back to top

Yes, I'd have to agree. Those don't look like ESR builds to me, which would normally be FF10 as a base with perhaps a newer Gecko.

I don't think it's even possible to build FF 20 with a Gecko dated 20100101 (I've never tried, but I know from my work with extensions that so much has changed since 2010 that it's highly unlikely).

So, your question is a good one!
Back to top

icywolfy, those mobile ua are added.

lewisr, os/2 browsers added in misc section.

Some other browsers added too. At some point I'll need to start cleaning out older mobile ua to keep the overall size manageable but it's fine for now.
Back to top

Excellent. Thanks!
Back to top

It would be nice to see a UA added for the new BlackBerry 10 devices:

Sample from BlackBerry Z10:

Back to top

Does anyone know an existing string or custom string that would allow the following:-

Recognise as Mobile Device but automatically default to PC/Desktop view (for use in Firefox PC browser), the mobile device aspect is unfortunately a must .
Back to top

"Recognise as Mobile Device but automatically default to PC/Desktop view (for use in Firefox PC browser), the mobile device aspect is unfortunately a must ."

That's not possible, you are assuming detections will know what the string intended, for me, on my php browser detection script, it will search for all possible mobile terms in the string then assign a mobile value if any is found.

Since you have no control over what browser sniffer methods are used, how could a string that has a mobile device recognition then also show as pc/desktop? If it shows the word ios, android, blackberry, etc, it is a mobile by definition, if it shows mobile, tablet, etc, in the string, it is mobile. A sniffer could be written to customize that on one website, but that's not related to a useragent string, that's programming, and would only apply to that one website.

If I were you I'd re-examine the project requirements and then point out that you cannot require two mutually contradictory things from one string, either it id's as a mobile device, or it does not. Whoever created this requirement made a mistake, a string is a string and cannot act dynamically.
Back to top

Thanks for sharing your knowledge

I was thinking more along the lines of when you are using a browser on an android phone, you can click 'desktop view' and it will go to the full site view. Perhaps the browser simply switches to a desktop type profile or alternative trickery.
Back to top

something like desktop view totally depends on where the link you clicked was located, if on the website, all it does is change the type of html/css sent to the browser.

Many websites serve mobile html/css to mobile devices, and it's a fairly trivial thing for them to add the option to serve desktop html/css.

With a useragent switcher, if the site is serving up content based on the useragent, you simply change useragents, then reload the page, though that does not always work if it's a separate html page, like mobile.bloomberg.com, then you have to reload the home page of the main site.

Sometimes the sites set cookies that tell the site it's a mobile, if that is set, you have to unset it then reload the page using useragent of desktop browser.

There's not one way to handle mobiles, so I don't actually understand what you are trying to do, usually it helps to not be vague and general and explain in explicit detail what the end result desired is.

If it's your website, or a site you are working on, that's one thing, if it's all sites on a browser, that's another thing, totally unrelated.

There are many ways for a browser to send its data to the server, and many ways for the server to decide what type of data to send the browser. There's also header links that identify, if you want, the useragent type, which the server can also use.
Back to top