Bug 9189 – [S60] Reindeer does not send any User-agent HTTP header

• Component: WebKit Misc.
• Status: CLOSED FIXED

[Franklin Davis]

Reindeer browser should send a reasonable HTTP header to allow AJAX features and other proper content formatting by sites. Suggestion:

Mozilla/5.0 (SymbianOS/9.1; U; [en]; Series60/3.0 Reindeer/2006-05-31) AppleWebKit/413 (KHTML, like Gecko) Safari/413

where Reindeer/xxxx is some unique version identifier, either the build date or some other version number.

[Krishna]

Reproducible on the S60 webkit, displays nothing.
Not reproducible on the Safari TOT. 
Safari displays the following user-agent header: 
"Mozilla/5.0 (Macintosh; U; Intel Mac OS X; en) AppleWebKit/420+ (KHTML, like Gecko) Safari/417.9.2"

Added "PlatformOnly" to keywords field.

[Franklin Davis]

Would also be good to send other HTTP headers compatible with full browser. See http://discussion.forum.nokia.com/forum/showthread.php?t=90610&highlight=HTTP+headers+browser


I suggest we send the same headers as 3.1, but for User-Agent I suggest:
Mozilla/5.0 (SymbianOS/9.2; U; Series60/3.1 NokiaReindeerEmulator/1.0; Profile/MIDP-2.0 Configuration/CLDC-1.1) AppleWebKit/413 (KHTML, like Gecko) Safari/413

[Bradley Morrison]

Adding S60 prefix to synopsis

[Bradley Morrison]

Adding S60 prefix to synopsis

[Bradley Morrison]

Created attachment 10936 [details]
Return static User agent string if __BROWSER_SDK

Please note I have updated to hardcoded UserAgentString to mention *Reindeer*, so this means if the UserAgentString is found in the central repository this value will be returned even if __BROWSER_SDK is not specified.

Not sure if this is a problem but the patch can be updated if so.

[Bradley Morrison]

Comment on attachment 10936 [details]
Return static User agent string if __BROWSER_SDK

Please also see previous comment about updated value of KHttpUserAgentString. I put *@nokia.com instead of *gmail.com in the review request

[Bradley Morrison]

(In reply to comment #5)
> Please note I have updated to hardcoded UserAgentString to mention *Reindeer*,
> so this means if the UserAgentString is found in the central repository this
> value will be returned even if __BROWSER_SDK is not specified.

s/if the UserAgentString is found in the central repository/if the UserAgentString is NOT found in the central repository/

It's late, I'm going home. ;-)
 

[alan]

Comment on attachment 10936 [details]
Return static User agent string if __BROWSER_SDK

1. please use #if _browser_sdk in webkitcontrol.cpp to separate sdk and rdn functionality.
2. do we really need to include browserversion.h?

[Bradley Morrison]

Created attachment 10948 [details]
Removed BrowserVersion.h include

[Bradley Morrison]

(In reply to comment #8)
> (From update of attachment 10936 [details] [edit])
> 1. please use #if _browser_sdk in webkitcontrol.cpp to separate sdk and rdn
> functionality.

You mean:
-            // come up with a hardcoded UA string
-            iUserAgent = KHttpUserAgentString().AllocL();
+            // return hardcoded string
+            iUserAgent = HBufC::NewL( KHttpUserAgentString().Length() );
+            iUserAgent->Des().Copy( KHttpUserAgentString );

?

Since KHttpUserAgentString is now always 8 bits, this needs to happen regardless of whether __BROWSER_SDK is set.

> 2. do we really need to include browserversion.h?

Ahh, nope, removed and patch updated.

[alan]

Comment on attachment 10948 [details]
Removed BrowserVersion.h include

wrong patch file. (the patch is about webuilogger)

[Bradley Morrison]

Created attachment 10993 [details]
Correct patch this time :)

Correct patch this time :) I will probably send this to spadma for a 2nd opinion as he is available now.

[alan]

Comment on attachment 10993 [details]
Correct patch this time :)

go back to 3.0 ua string. sending yet another (and very much different) is confusing.

[Bradley Morrison]

Created attachment 11033 [details]
Updated static UA string to 3.1

Also reworked patch a little so that static uastring not declared with _LIT8. does a conversion via HBufC8 in CHttpTransaction::AddRequestHeadersL instead.

[alan]

Comment on attachment 11033 [details]
Updated static UA string to 3.1

push uastring to the cleanupstack by calling NewLC instead.
(cleanupstack gets out of sync and crashes in this patch)

[Bradley Morrison]

Created attachment 11039 [details]
bah!

Patch calls HBufC::NewLC not HBufC::NewL now. I have to admit, good catch. But anyway ..... arrrrrghhhh! :P

[alan]

Comment on attachment 11039 [details]
bah!

what a great patch!
r=me

[alan]

r17001

[Bradley Morrison]

Sachin would prefer that the UA string is not different if not found in the cenrep.

[Bradley Morrison]

Created attachment 11080 [details]
Value of KHttpUserAgentString depends on __BROWSER_SDK

[Sachin Padma]

Comment on attachment 11080 [details]
Value of KHttpUserAgentString depends on __BROWSER_SDK

r=me

[Bradley Morrison]

The UA string is correct, I'm setting this to resolved.

[Krishna]

TSW ID:  KPOY-6UMP8F 

[Krishna]

Closed