Bug report and patch: photometric interpretation of TIFF files

[Paul Williams]

This is a multi-part message in MIME format.
--------------010501080009090100000303
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit

tumble 0.33 currently treats TIFFs as having photometric interpretation 
min-is-white, resulting in negative images in the PDF if the photometric 
interpretation was min-is-black.

It appears that a simple hack to tumble_tiff.c to read and use the 
photometric interpretation does the trick, so that's what I've applied 
and tested locally. I've attached a patch file. I've never submitted one 
of these before, so I'd better say that I renamed tumble_tiff.c to 
tumble_tiff.c.orig, made my change and tested it, and then "diff -u 
tumble_tiff.c.orig tumble_tiff.c".

Regards,
Paul

--------------010501080009090100000303
Content-Type: text/plain;
 name="tumble_tiff.c.patch"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
 filename="tumble_tiff.c.patch"

--- tumble_tiff.c.orig	2003-12-09 16:45:43.000000000 +0000
+++ tumble_tiff.c	2003-12-20 20:50:01.000000000 +0000
@@ -274,6 +274,8 @@
   Bitmap *bitmap = NULL;
 
   int row;
+  uint16_t photometric_interp;
+  bool blackis1;
 
   rect.min.x = 0;
   rect.min.y = 0;
@@ -289,6 +291,11 @@
       rect.max.y = image_info->height_samples;
     }
 
+  if (1 != TIFFGetField (tiff_in, TIFFTAG_PHOTOMETRIC, & photometric_interp))
+    blackis1 = 0;
+  else
+    blackis1 = PHOTOMETRIC_MINISBLACK == photometric_interp;
+
   bitmap = create_bitmap (& rect);
 
   if (! bitmap)
@@ -333,7 +340,7 @@
 			  bitmap,
 			  0, /* ImageMask */
 			  0, 0, 0,  /* r, g, b */
-			  0); /* BlackIs1 */
+			  blackis1); /* BlackIs1 */
 #endif
 
   result = 1;

--------------010501080009090100000303--