Version: 0.20 (using KDE KDE 3.4.1) Installed from: Gentoo Packages Compiler: gcc (GCC) 3.3.5-20050130 (Gentoo 3.3.5.20050130-r1, ssp-3.3.5.20050130-1, pie-8.7.7.1) USE flags: +arts -debug -kdeenablefinal -kdexdeltas +xinerama CFLAGS (in make.conf): -O3 -march=pentium3 -fomit-frame-pointer -pipe OS: Linux Maybe it is the same bug as the 97799, but i get not only problem with the graphics. It is tables in the files, and just 2 of those was there, the others was invisible The file is the UNDP human developmant report 2005 in french at http://hdr.undp.org/reports/global/2005/francais/pdf/HDR05_fr_complete.pdf I have try xpdf, and only the graphics was wrong, absolutely black, but the tables was OK. A strange thing is at xpdf is working just fine when reading the file inside mozilla, it is only when i read the file with xpdf as standalone application at the graphics are black. I have installed (...) and try acrobad reader. It is working just fine with this file. It is a big file, more as 11Mb. For the graphics, see page 21 (37 on the left column), for the tables pages 231 (247 on the left) and followings.
It was the same problem with the graphics in this file with xpdf but the tables was OK. I send a bug rapport, and a few hours later, i get this mail in return: That was a bug in handling sampled functions with 32-bit samples. I'm attaching a patch file. Index: Function.cc =================================================================== RCS file: /home/derekn/src/cvs/xpdf/Function.cc,v retrieving revision 1.24 retrieving revision 1.25 diff -c -r1.24 -r1.25 *** Function.cc 14 Feb 2005 04:27:46 -0000 1.24 --- Function.cc 13 Sep 2005 23:25:33 -0000 1.25 *************** *** 190,196 **** Object obj1, obj2; Guint buf, bitMask; int bits; ! int s; int i; samples = NULL; --- 190,196 ---- Object obj1, obj2; Guint buf, bitMask; int bits; ! Guint s; int i; samples = NULL; *************** *** 239,245 **** goto err2; } sampleBits = obj1.getInt(); ! sampleMul = 1.0 / (double)((1 << sampleBits) - 1); obj1.free(); //----- Encode --- 239,245 ---- goto err2; } sampleBits = obj1.getInt(); ! sampleMul = 1.0 / (pow(2, sampleBits) - 1); obj1.free(); //----- Encode ################################################## I have not try this patch, i am too busy.
The linked document is not available anymore. Could you please confirm whether or not this works as expected with Okular?
No answer for years.