AWARE SYSTEMS
TIFF and LibTiff Mail List Archive

Thread

2013.04.27 21:34 "[Tiff] Problem with TIFFReadEndcodeStrip()", by Steve McKinney
2013.04.28 01:22 "Re: [Tiff] Problem with TIFFReadEndcodeStrip()", by Bob Friesenhahn
2013.04.28 03:22 "Re: [Tiff] Problem with TIFFReadEndcodeStrip()", by Steve McKinney
2013.04.28 13:51 "Re: [Tiff] Problem with TIFFReadEndcodeStrip()", by Bob Friesenhahn
2013.04.28 06:37 "Re: [Tiff] Problem with TIFFReadEndcodeStrip()", by Steve McKinney

2013.04.28 13:51 "Re: [Tiff] Problem with TIFFReadEndcodeStrip()", by Bob Friesenhahn

On Sat, 27 Apr 2013, Steve McKinney wrote:

I understand the effect of setting TIFFTAG_JPEGCOLORMODE to JPEGCOLORMODE_RGB, but why is it necessary for me to do this in order to prevent an access violation when calling TIFFReadEncodedStrip()? For what it's worth, the access violation occurred when reading strip zero, and the buffer I passed was large enough to hold the entire decoded image. Under what circumstances do I need to set TIFFTAG_JPEGCOLORMODE to JPEGCOLORMODE_RGB? Always? Only when the photometric interpretation is YCbCr? Is it not possible to read a JPEG compressed TIFF in YCbCr format without converting it to RGB? If so, shouldn't that be the default behavior under the circumstances? Certainly an access violation doesn't seem desirable.

Access violations are not desireable. What versions of libtiff and libjpeg are you using? Are you able to provide your small sample TIFF file which causes the crash?

Bob
--
Bob Friesenhahn
bfriesen@simple.dallas.tx.us, http://www.simplesystems.org/users/bfriesen/
GraphicsMagick Maintainer, http://www.GraphicsMagick.org/