so wysota, do you think this should be taken up to the trolls? i mean it makes sense that if an image is corrupted, it should be reported..or is it like they know the issue and they r just playing ostrich on it?
so wysota, do you think this should be taken up to the trolls? i mean it makes sense that if an image is corrupted, it should be reported..or is it like they know the issue and they r just playing ostrich on it?
You can report it of course it would be worth to check first if libjpeg has any means of notifying about such corruption. If not, Trolls won't be able to do anything about it other than patching libjpeg which is doubtful they would do.
Thanks for your help!
I reported the problem in the Qt tasktracker.
libjpeg provides very detailed status codes, see http://libjpeg.cvs.sourceforge.net/v....h?view=markup. However, qjpeghandler.cpp does not seem to check them. The corresponding function is
Qt Code:
static bool read_jpeg_image(QIODevice *device, QImage *outImage, const QByteArray ¶meters, QSize scaledSize, int inQuality )To copy to clipboard, switch view to plain text mode
Also, it seems that the interface provided by QImageIOHandler would need to be extended to be able to forward additional "status messages" since currently the read() and canRead() methods only return boolean values.
i always wanted to find a proper bug in Qt..it always seemed so perfect and foolproof..and here we are..good work miky![]()
The problem discussed can now be found in the tasktracker:
http://www.qtsoftware.com/developer/...ntry&id=251514
I'm facing the same issue, and no solutions has been reported on the Bug Tracker.
How did you solve it? Have you found a way to detect the corrupted images?
Thanks
Bookmarks