[Qt-interest] Qt 4.5.0 not 'valgrind proof'

[wim.delvaux at adaptiveplanet.com]

Thank you all !!

Indeed, my args was NOT valid after a while !

Thx

W

On Monday 16 March 2009 11:10:27 Denis Dzyubenko wrote:
> Hi,
>
> Christoph Duelli wrote:
> >> I get segfault errors in valgrind even before my app starts.
> >>
> >> Here as sample backtrace
> >>
> >> ==25795== Process terminating with default action of signal 11
> >> (SIGSEGV): dumping core
> >> ==25795==  Access not within mapped region at address 0x9D
> >> ==25795==    at 0x6A027BD: QString::fromLocal8Bit(char const*, int)
> >> (qstring.cpp:3652)
> >> ==25795==    by 0x6AB2032: QCoreApplication::arguments()
> >> (qcoreapplication.cpp:1921)
> >
> > Guessing without code is probably not too helpful, but still:
> > I once made the trivial mistake of forgetting the "&" (in int &argc) in
> > my QApplication-subclass's constructor. This had "interesting" effects...
> > Just an idea...
>
> indeed. According to the QApplication documentation[1], arguments
> provided to the QApplication constructor must be valid objects during
> the lifetime of the qapplication object.
>
> [1] http://doc.trolltech.com/4.5/qapplication.html#QApplication