[gs-commits] rev 12302 - trunk/gs/Resource/Init

[ray at ghostscript.com]

Author: ray
Date: 2011-03-19 00:32:37 +0000 (Sat, 19 Mar 2011)
New Revision: 12302

Modified:
   trunk/gs/Resource/Init/pdf_ops.ps
Log:
Fix for some strange rendering with PDF 1.7 FTS files when we have shading and transparency
and are both filling and stroking text (Text Rendering modes 2 and 6). Customer 532.

EXPECTED DIFFERENCES:

PDF_1.7_FTS/fts_15_ group, 1509, 1511, 1513, 1515, 1517, 1519 and 1521.

Examined with bmpcmp and all seem to be progressions.


Modified: trunk/gs/Resource/Init/pdf_ops.ps
===================================================================
--- trunk/gs/Resource/Init/pdf_ops.ps	2011-03-18 05:16:08 UTC (rev 12301)
+++ trunk/gs/Resource/Init/pdf_ops.ps	2011-03-19 00:32:37 UTC (rev 12302)
@@ -985,7 +985,26 @@
    currentpoint //Tmatrix currentmatrix TextSaveMatrix setmatrix .swapcolors stroke .swapcolors
    setmatrix moveto
  } bdef
-/tB { gsave tf grestore tS } bdef
+
+% Handle transparency the same as /B operator
+/tB {
+       PDFusingtransparency {
+	  % knockout trans group around the filled and stroked object
+	  mark
+	  /Isolated //true /Knockout //true
+	  .dicttomark
+	  1 .setopacityalpha 
+	  % While text will always have a currentpoint, strokepath seems to mess with it.
+	  % we get the currentpoint, then use moveto to restore it for pathbbox
+	  gsave currentpoint strokepath moveto pathbbox grestore
+	  .begintransparencygroup
+	  gsave tf grestore tS
+	  .endtransparencygroup
+	} {
+	  gsave tf grestore tS
+	} ifelse
+} bdef
+
 % This does the wrong thing if there have been multiple text operations
 % within a single BT/ET pair, but it's a start.
 /tW { } bdef