On Thu, Jan 24, 2008 at 10:44:43AM +0100, Marcus Meissner wrote:
> On Thu, Jan 24, 2008 at 09:00:32AM +0000, Louis. Lenders wrote:
> > 
> > Hi, basically this is a resend of a patch i sent very long time ago ( 
> > http://www.winehq.org/pipermail/wine-patches/2007-April/037733.html )
> > 
> > It's a fix for bug http://bugs.winehq.org/show_bug.cgi?id=7380#c29
> > 
> > As explained in comment http://bugs.winehq.org/show_bug.cgi?id=7380#c29
> > the cause of these crashes is some buggy mfc-function
> > , and as windows is able to work around this, wine should do the same.
> > 
> > I've put the exception handler only around X11DRV_DIB_SetImageBits_24 , as 
> > that's where all the applications crash; if we want to mimic windows 
> > behaviour it should be put around all cases i guess, but i'm not sure we 
> > want to that and maybe hide other bugs.
> > 
> > If this patch is rejected could you please comment why, to help fix the bug 
> > in a better way
> 
> Remembering a different bug with the same issue I think you should not wrap 
> DIB_SetImageBits_24 but just check descr->bits in X11DRV_DIB_SetImageBits()
> 
> Something along the lines of:
> 
>       unsigned long off;
>       unsigned char volatile x;
> 
>       __TRY {
>               for (off=0;off<descr->sizeImage;off+=PAGE_SIZE) {
>                       x += ((unsigned char*)descr->bits)[off]
>               }
>       } ...

Or just IsBadReadPtr(descr->bits,descr->sizeImage) of course. :/

(Although Alexandre likely has different opinions ;)

Ciao, Marcus


Reply via email to