Focus stealing protection, qt_x_user_time and Runtime vs. Compile time check

Lubos Lunak l.lunak at suse.cz
Fri Jul 23 22:05:15 BST 2004


Dne pá 23. července 2004 12:55 Holger Freyther napsal(a):
> Hey guys,
>
> after my yesterdays troubles with linking kdecore against a too old Qt and
> having problems with Intern XAtoms not present in older versions of Qt. I
> really wonder why you do CompileTime checks if you've solutions for both
> older and newer Qt versions that you check for at compile time.
>
> I wonder why we can't declare qt_x_user_time as a weak symbol and then do
> runtime check to see if it is present.

 Because it is always present in Qt3.3.

> Weak symbols are at least supported 
> by GNU/Linux (gcc/glibc), and MacOS X, I think other platforms offer that
> as well.
>
> Is there a general guideline when to use intern XAtoms of Qt and when not
> to rely on them?

 If 'intern XAtoms' is actually supposed to mean Qt symbols, then one should 
avoid those not marked as public as not having Q_EXPORT.

-- 
 Lubos Lunak
 KDE Developer




More information about the kde-core-devel mailing list