navigator.appVersion doesn't contains "linux"




Networking: HTTP
18 years ago
17 years ago


(Reporter: Thibaut Colar, Assigned: Darin Fisher)



Firefox Tracking Flags

(Not tracked)




18 years ago
In Netscape 4.x or ie
calling navigator.appVersion would return for example
	4.72 (Windows; en-US)
	4.72 (Linux X11; en-US)
	4.72 (Solaris; en-US)
	4.72 (Macintoch; en-US)
Using Netscape 6:
It returns something similar on windows and Mac (btw why 5 ?? ):
	5.0 (Windows; en-US)
	5.0 (Macintoch; en-US)
	5.0 (X11; en-US)   -> no more Linux
probably same on soalris (not tested)
It was usefull to test the platform and we have (as other companies) scripts
using that.

Also we can't modify them easily, since they are on clients web sites (over 30000)
Thank you.

Comment 1

18 years ago
according to ekrock in bug 53260, it is long since decided to return 5.0
The comment refers to a decision in the newsgroups.

X11 for X system environments was also mentioned as a problem in 53260.
ekrock states moz should return X11 for X Windows systems. The bug was then set

However, ekrock also refers to
where it quite contrary reads:

"Values for Unix systems shall be the output of the command uname -srm (also
accessible as the sysname, release, and machine fields of the utsname structure.)"

A uname -srm on my system (linux RH6.2) show: Linux 2.2.16-3 i686

I remember seeing another bug about the full linux kernel version displaying,
however: it was thought to represent a potential security risk. I can't find
that bug again, so don't know the outcome. Perhaps someone remembers it.

Resolving this one as new, since appversion spec on web states an equivalent to
"uname -srm" should be included in appversion.
Ever confirmed: true

Comment 2

18 years ago
Browser, not engine. Reassigning to DOM Level 0 component. 

cc'ing ekrock -
Assignee: rogerl → jst
Component: Javascript Engine → DOM Level 0
QA Contact: pschwartau → desale

Comment 3

18 years ago
I think the "uname -srm" is the good solution since it was what was returning 
by previous browsers, also it used to return the system name on other platforms 
( ie "Windows"  "Macintosh") and "X" is not a system name but a graphical 
environment name (in this case windows should return "explorer" and 
mac "Finder")!
Thank you.

Comment 4

18 years ago
R.K.Aa: see bug 57555

Comment 5

18 years ago
That's the one. Thanks.
Worth consideration. But the OS as such should absolutely display.
For many sites it will be of statistical interest whether Linux, IRIX, Solaris,
BSD etc. is used.

Comment 6

18 years ago
I agrre that the full kernel version is risky and almost totally not usefull in 
a website point of view. (except maybe some very sharp stat) but not big deal.
In another hand the system name (Linux) is very usefull, for statistics and for 
platform specific code.
The DOM code gets this value from the HTTP protocol handler so I'm reassigning
to the network team.
Assignee: jst → gagan

Comment 8

18 years ago
This value corresponds to the platform attribute on the nsIHTTPProtocolHandler
interface.  Right now, it is hard coded as X11 for XP_UNIX.  The oscpu attribute
gives you the equivalent of uname -srm.  I'm not really sure why X11 is reported
as the platform... perhaps just to indicate that the user agent has a graphical
display... although I'm not sure why this would need to be indicated.  At any
rate I generally agree with the suggestion to change the value of the platform
attribute to include the OS name.  Any objections?

Comment 9

18 years ago
Not for me ;-)
It seems that I'm not the only one to think it makes sense.
Thank you.

Comment 10

18 years ago
added keyword 4xp.
Keywords: 4xp

Comment 11

18 years ago
Assignee: gagan → darin
Component: DOM Level 0 → Networking: HTTP
QA Contact: desale → tever

Comment 12

18 years ago
What's up with the bug, is it fxed, will it be soon ?
Thank you.

Comment 13

18 years ago
The issue of whether or not navigator.appVersion should be X11 was resolved
by bug 53260... the conclusion there is that X11 stands.

*** This bug has been marked as a duplicate of 53260 ***
Last Resolved: 18 years ago
Resolution: --- → DUPLICATE

Comment 14

18 years ago
you said
I generally agree with the suggestion to change the value of the platform
attribute to include the OS name.  
and now you see that it's fine like it is.
I don't get it.
Whatever !

Comment 15

18 years ago
I'm not saying that it shouldn't be changed.  I am simply saying that
this issue has apparently been discussed and resolved.  But, that does
not mean that the issue is dead (only this particular bug report)...
feel free to reopen bug 53260 if you feel strongly that X11 is not
the correct value for navigator.appVersion.

When I said that I generally agree with you, I said that without knowing
why others have called this INVALID.  You should find out why this was

Comment 16

17 years ago
verified dup
You need to log in before you can comment on or make changes to this bug.