Why do I show up as Linux with browser reports?

A

ahhyes

Hi All,

Apologies if this topic has been done to death, its late here and I am tired.

I have tried with both Firefox and Chromium on FreeBSD and sites report me as using linux........ Various sites that allow me to test my browser report my 'navigator platform' as Linux, same deal with both Chromium and Firefox...

Whats going on here?

[me@bsd-desktop ~]$ file /usr/local/lib/firefox/firefox
/usr/local/lib/firefox/firefox: ELF 64-bit LSB pie executable, x86-64, version 1 (FreeBSD), dynamically linked, interpreter /libexec/ld-elf.so.1, for FreeBSD 14.1, FreeBSD-style, BuildID[sha
1]=0f499288269b1fd057129ca854b77f0f5302e311, with debug_info, not stripped
Click to expand...

Apparently my navigator platform is Linux...

Whoever decided this was a bright idea are not doing the platform any favors as when webmasters look at site stats, it seems FreeBSD visitors will just show up as linux.
 
cracauer@ said:
Yes, the browsers in ports are adjusted to send a Linux id. Otherwise too many web pages would break.
Click to expand...
I'm a nobody round these parts, but my life is in enterprise web apps and this doesn't make much sense to me.

I.e. for some years now web compatibility has really been about Chrome (+derivatives/variants) vs Safari, with a sprinkling of Firefox. The underlying OS doesn't really come into it (Safari OSX/iOS notwithstanding). I can't fathom why linux would need special casing - even if somebody (for perhaps legacy reasons) wanted to special-case windows or OSX.

Also, the idiomatic way would be to include something like "like Linux" - just as Chrome includes "like Gecko" for legacy compatibility reasons - so "FreeBSD, like Linux" should have the same effect, whilst still allowing browser stats detection to see that its really FreeBSD.

All that said, I have the strong perception that FreeBSD folk generally do know what they're doing, and don't suffer fools (one of the many reasons I want to come back to FreeBSD!) - so I'd be really interested to find out more about why this is done. I can start grepping the ports source, but if you know of any writeup/mailinglist posts let me know. Thanks :)

(and fwiw, in my day job and hobbies its all the same code for everyone, except the the odd Safari special-casing. nothing else)
 
cracauer@ said:
Yes, the browsers in ports are adjusted to send a Linux id. Otherwise too many web pages would break.
Click to expand...
Personally I find that unlikely in 2024, most sites are using generic things like HTML, Javascript and CSS those are all pretty universal and not platform specific - has anyone actually tried leaving it as the default in recent times and let it report FreeBSD to see what the experience is like?

How do I undo this change and let it report the truth?
 
ahhyes said:
Personally I find that unlikely in 2024, most sites are using generic things like HTML, Javascript and CSS those are all pretty universal and not platform specific - has anyone actually tried leaving it as the default in recent times and let it report FreeBSD to see what the experience is like?

How do I undo this change and let it report the truth?
Click to expand...

Both FF and Chrome have extensions to switch out the agent string.

The first thing to break will be banking sites.
 
ralphbsz said:
You worked at CERN? Even then, it's a year short.
Click to expand...
I was building & hosting web stuff for fun in 1999 at university, which - terrifyingly - is 25 years ago :/

Back when PHP 3 was new and the name PHP still kinda stood for "Personal Home Page", CGI was all the rage (but nothing to do with graphics), and the choice for (web) backend languages was basically PHP, Perl or C. Fun times.

Interestingly my university had had a project a few years before that where their flavour of hypertext design involved separate databases of links (separate to the documents) - quite unlike the embedded html links we use today where the source document has the link and destination (URL) embedded in it. They were still pretty salty that their way had lost. I mean, theirs was probably technically more powerful, but far too painful to author content in.

Edit: edited to clarify not professionally, and make it a bit clearer this was public internet stuff, not special university projects.
Edit #2: realised you weren't replying to me, haha, oops. I'll leave this here anyway :)
 
ralphbsz said:
You worked at CERN? Even then, it's a year short.
Click to expand...
[OT]
Sorry to chime in but you're totally wrong with this. I started working for what was the biggest Italian web portal at that time ( www.virgilio.it ) in late 1999 and there were more than 100 people working there already... Many of them were web developers of course (I was a sysadmin). The web was a reality well before then, it didn't start with Google.
[/OT]
 
cracauer@ said:
The first thing to break will be banking sites.
Click to expand...
Why would they? All those sites are doing is feeding your web browser html, javascript and css stylesheets. Typically the server side application (which probably just makes API calls to some server internally in the bank would not need to care about the visitors user agent string). All those sites really need at the end of the day is a browser that can interpret html, js, css and render correctly (that just about sums up any modern web browser).
 
I feel a better approach here would be to let the browser report the real operating system and then leave it to the end user if they want to fudge the user agent string with a plugin. In my travels, any site that scrutinizes the user agent string is usually doing so to see if its a supported web browser (ie: is it firefox, chromium, opera etc), and doesnt pay attention to the OS that the browser itself is running on.
 
ahhyes said:
Why would they? All those sites are doing is feeding your web browser html, javascript and css stylesheets. Typically the server side application (which probably just makes API calls to some server internally in the bank would not need to care about the visitors user agent string). All those sites really need at the end of the day is a browser that can interpret html, js, css and render correctly (that just about sums up any modern web browser).
Click to expand...

They have a catalog of "safe" browser IDs and don't let you in if you are "insecure". That's why this change originally came to be.
 
cracauer@ said:
That's why this change originally came to be
Click to expand...
Is there a way to unwind it? I'd like to conduct my own testing with the financial institution I am with. I am happy to build it from ports - is there some patch there I can back-out etc?
 
ahhyes said:
Is there a way to unwind it? I'd like to conduct my own testing with the financial institution I am with. I am happy to build it from ports - is there some patch there I can back-out etc?
Click to expand...

As I said, just load one of the agent switcher extensions.
 
The problem with the user agent switcher apps is that they are only get an upgrade once in a year. Unfortunately that makes the
user even stands out more out of the crowed. Some of them are working well but would need frequent upgrades and I could not
find any with frequent upgrades.
 
You must log in or register to reply here.
Back
Top