Details
-
Bug
-
Resolution: Incomplete
-
Not Evaluated
-
None
-
5.1.1, 5.3.0 Alpha
-
None
Description
Today we got a nice long Mac OS X crashlog culminating in
QtGui 0x0000000104a23b41 QImageData::create(unsigned char*, int, int, int, QImage::Format, bool, void 
(void*), void*) + 161
Reason is supposedly in line 783
if (width <= 0 || height <= 0 || !data || INT_MAX/sizeof(uchar *) < uint(height) || INT_MAX/uint(depth) < uint(width) || bpl <= 0 || height <= 0 || bpl < min_bytes_per_line || INT_MAX/uint(bpl) < uint(height)) return d; // invalid parameter(s)
depth could be 0 , as qtbase/lib/QtGui.framework/Versions/5/Headers/5.1.1/QtGui/private/qimage_p.h
returns 0 when it gets fed an invalid enumerator:
inline int qt_depthForFormat(QImage::Format format) { int depth = 0; switch(format) { case QImage::Format_Invalid: case QImage::NImageFormats: Q_ASSERT(false); case QImage::Format_Mono: case QImage::Format_MonoLSB: depth = 1; break; case QImage::Format_Indexed8: depth = 8; break; case QImage::Format_RGB32: case QImage::Format_ARGB32: case QImage::Format_ARGB32_Premultiplied: depth = 32; break; case QImage::Format_RGB555: case QImage::Format_RGB16: case QImage::Format_RGB444: case QImage::Format_ARGB4444_Premultiplied: depth = 16; break; case QImage::Format_RGB666: case QImage::Format_ARGB6666_Premultiplied: case QImage::Format_ARGB8565_Premultiplied: case QImage::Format_ARGB8555_Premultiplied: case QImage::Format_RGB888: depth = 24; break; } return depth; }
The real reason must be somewhere down in the crashlog, of course.