Current HTTP headers dump: WinNT4.0 GET / HTTP/1.1 Host: bubblegum:4321 User-Agent: Mozilla/5.0 (Windows; U; WinNT4.0; en-US; m18) Gecko/20010208 Netscape6/x.x Accept: */* Accept-Language: en Accept-Encoding: gzip,deflate,compress,identity Keep-Alive: 300 Connection: keep-alive Linux rh6 GET / HTTP/1.1 Host: bubblegum:4321 User-Agent: Mozilla/5.0 (X11; U; Linux 2.2.5-22 i686; en-US; m18) Gecko/20010208 Netscape6/x.x Accept: */* Accept-Language: en Accept-Encoding: gzip,deflate,compress,identity Keep-Alive: 300 Connection: keep-alive Mac os9 GET / HTTP/1.1 If-Modified-Since: Wed Nov 01 17:01:08 GMT 2000 Host: bubblegum:4321 User-Agent: Mozilla/5.0 (Macintosh; N; PPC; en-US; m18) Gecko/20010208 Netscape6/x.x Accept: */* Accept-Language: en Accept-Encoding: gzip,deflate,compress,identity Keep-Alive: 300 Connection: keep-alive In a news article, I questioned why Mac and the others were different with respect to the 'N' and 'U' part of the string. I'm filing this bug to track this issue. Darin said: > That element of the UA string indicates the security level supported by > the browser. It comes from the pref: general.useragent.security, where > 'N' is the default value (#define'd as UA_APPSECURITY_FALLBACK in > nsHTTPHandler.cpp). Sfraser said: Does that mean that the Mac build used to generate these user-agent strings did not have PSM installed, and the others did? Or that there is a bug detecting PSM on the Mac when generating the user-agent? Darin said: Hmm... probably a bug... not sure if PSM sets general.useragent.security when loaded. The HTTP handler should probably register a callback for that pref, so it can update the UA string whenever it changes. Who might know more about this pref?

Would someone at least please LOOK at this bug? It might cause all sorts of issues with Mozilla on secure sites, and we'd be none the wiser.

The lastest User-Agent headers show that Mac is now picking up PSM: User-Agent: Mozilla/5.0 (Macintosh; U; PPC; en-US; 0.8.1) Netscape6/xxxb0

Verified.