On Sat, Aug 22, 2009 at 06:47:53PM +0200, Lucas Nussbaum wrote:
> > t/t104ppm.........ok
> > t/t105gif.........ok
> > 
> > #   Failed test 'reading multiple images from tiff'
> > #   at t/t106tiff.t line 315.
> > Use of uninitialized value in subroutine entry at t/t106tiff.t line 318.
> > Use of uninitialized value in subroutine entry at t/t106tiff.t line 318.
> > im2 is not of type Imager::ImgRaw at t/t106tiff.t line 318.
> > # Looks like you planned 213 tests but only ran 67.
> > # Looks like you failed 1 test of 67 run.
> > # Looks like your test died just after 67.
> > t/t106tiff........dubious
> >     Test returned status 255 (wstat 65280, 0xff00)
> > DIED. FAILED tests 67-213
> >     Failed 147/213 tests, 30.99% okay
> > t/t107bmp.........ok

Hi,

I'm the upstream author.

This appears to be change in behaviour for tifflib and is present in
the original non-Debian sources, eg.

#include <tiffio.h>
#include <stdio.h>

int main(void) {
  TIFF *tif;
  int i, y;
  char buf[100] = "\0";

  tif = TIFFOpen("foo.tif", "w");
  if (!tif) {
    fprintf(stderr, "Cannot create file\n");
    return 1;
  }
  for (i = 0; i < 2; ++i) {
    int bytes_pp;
    TIFFSetField(tif, TIFFTAG_IMAGEWIDTH, 20);
    TIFFSetField(tif, TIFFTAG_IMAGELENGTH, 20);
    TIFFSetField(tif, TIFFTAG_SAMPLESPERPIXEL, 3);
    TIFFSetField(tif, TIFFTAG_PLANARCONFIG, PLANARCONFIG_CONTIG);
    bytes_pp = TIFFScanlineSize(tif);
    for (y = 0; y < 20; ++y) {
      TIFFWriteScanline(tif, buf, y, 0);
    }
    TIFFWriteDirectory(tif);
  }
  TIFFClose(tif);
  
  return 0;
}

gcc -otifdirtest tiffdirtest.c -ltiff
./tiffdirtest
tiffdump foo.tif | grep StripByteCounts

# Etch:
StripByteCounts (279) LONG (4) 1<160>
StripByteCounts (279) LONG (4) 1<160>

# Sid:
StripByteCounts (279) LONG (4) 1<0>
StripByteCounts (279) LONG (4) 1<0>

# original libtiff 3.9.0
(using LD_LIBRARY_PATH=/home/tony/local/tiff/lib ./tiffdirtest with tiffdirtest 
builting against an installed tifflib 3.9.0):
StripByteCounts (279) LONG (4) 1<0>
StripByteCounts (279) LONG (4) 1<0>

>From the Imager side of things, the single image function *doesn't*
produce the problem StripByteOffsets - and the only real difference
between that and the multiple image writer is the call to
TIFFWriteDirectory for each image in the multi-image code.






-- 
To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Reply via email to