See bug #92035 for details. kview is also affected, although it does not die. QMetaObject::findSignal:KViewViewer: Conflict with KImageViewer::Viewer::imageOpened(const KURL&) QMetaObject::findSignal:KViewViewer: Conflict with KImageViewer::Viewer::imageOpened(const KURL&) Corrupt JPEG data: 59 extraneous bytes before marker 0xd9
Sent report and patch upstream, so others distros get the information as well. kview notices the invalid jpeg data, but doesn't care about exif data at all.
Is this bug vendor-sec restricted or why is it a hidden one? I did not get any response yet (and "yet" may go against infinite, since I did not get a direct response the last time either), but there's no reason not to apply the necessary patch and disclose, imho.
Auditors please check wether this is exploitable or just a crash.
Created attachment 58825 [details] kdegraphics-kfile-plugins-3.4.0-exif.diff Oh, if it even gets audited, I'm attaching the patch I want to apply. I know it's quick and dirty and I would wrap the function properly, if I'd maintain the code, but I don't see any problem with it, since the function is not used elsewhere.
Auditors please confirm the issue and patch and let's get patching.
No issues other than annoyance that I can see. It may recurse to far and overflow the stack and crash. Can't see it being exploitable.
Not a security issue -> opening and reassinging to KDE. Have fun:-)
Carsten, is there a reason you haven't applied this locally? Was there a bad reaction upstream or something?
I think this can be closed.