[Webkit-unassigned] [Bug 27158] Add WebKit version API to Qt.

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=27158





--- Comment #13 from robert <robert at roberthogan.net>  2009-07-15 11:10:17 PDT ---
> 
> I have a few small comments below. Otherwise I have only one suggestion:
> Instead of generating the entire header file from within the .pro file simply
> include the contents of the generated header file in your patch itself.

My intention was to remove the need for manual intervention.  I don't
understand why keeping version updates manual is necessary, but happy to do
what is required. I'll provide an updated patch on this basis.

> 
> Afterwards all we need is
> 
> 1) The other ports including the header file and using it (very simple patch)
> 

No problem. I guess the other ports can chime in at that point.

> and
> 
> 2) Whenever WebCore/Configurations/Version.xcconfig is touched someone needs to
> re-run your perl script to update the header file (simple).
> 
Still not sure why this can't be automated.

> 
> Out of curiousity, why did you decide to append the Safari bit at the end
> instead of just keeping the existing
> initual ua.append() line and use arg() twice?
> 

To bring Qt's user agent string into line with the format adopted by other
ports:

http://www.mattcutts.com/blog/google-chrome-user-agent/

http://209.85.229.132/search?q=cache:fd3ykESxOycJ:www.useragentstring.com/pages/Safari/+safari+user+agent&cd=3&hl=en&ct=clnk

http://gitorious.org/qtwebkit/qtwebkit/blobs/1a9740a524a3cc5958c996f945a99ddad144f3f4/WebKit/gtk/webkit/webkitwebsettings.cpp
(line 182)

> Doesn't this patch also change the parentheses of the user agent?
> 
> I think if we change the structure of the user agent string, then we have to
> adjust the API documentation of the method, too.

Yes, missed that one.

> 
> > +void tst_QWebView::getWebKitVersion()
> > +{
> > +    QVERIFY(qWebKitVersion().toDouble());
> > +}
> > +
> 
> I'm not sure if this test is very useful :)
> 
> Given the simplicity of the function I think it's okay to skip this "test"
> alltogether if you want.

It's intended to catch situations where the auto-generated header file is so
broken the defines are not numeric. It should really be toDouble() > 0 in that
case.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.