[gs-commits] ghostpdl branch, master, updated. jbig2dec-0.14-1612-g02c252f

The ghostpdl branch, master has been updated
       via  02c252f00e7bbc9a9a23d0a0996e88ba1839f127 (commit)
      from  3554f0fc4dadebad2e816adbd7cf4a548f76de74 (commit)

----------------------------------------------------------------------
commit 02c252f00e7bbc9a9a23d0a0996e88ba1839f127
Author: Ken Sharp <[email protected]>
Date:   Tue Aug 27 11:36:35 2019 +0100

    PDF interpreter - more /Mask strangeness in Acrobat
    
    Bug #701468 " More weirdness with images containing invalid /Mask values"
    
    In order to render this file as per Acrobat, we need to ignore the /Mask
    array. However, in order to render bug 689717 correctly we need to fix
    the /Mask array.
    
    Both cases are 1 BitsPerComponent DeviceGray images, the only difference
    is the actual /Mask values. Bug 689717 has [255 255] while 701468 has
    [243 255]. We can differentiate based on whether we have DeviceGray
    1 BPC image, where the Mask values are not the same. In the case they
    are different we use the existing technique of calculating the maximum
    value and using that (for both, so there is no range). Where there is a
    range (ie not a single colour) this is clearly unusable in a 2-colour
    image, so ignore the /Mask altogether.

diff --git a/Resource/Init/pdf_draw.ps b/Resource/Init/pdf_draw.ps
index fe3fc56..a359a9c 100644
--- a/Resource/Init/pdf_draw.ps
+++ b/Resource/Init/pdf_draw.ps
@@ -2128,13 +2128,15 @@ currentdict /last-ditch-bpc-csp undef
     1 BitsPerComponent bitshift 1 sub //false 2 index {
       % stack: max_value result_bool value
       dup 0 lt exch 3 index gt or or
-    } forall exch pop {
+    } forall exch pop
+    {
       (   **** Error: Some elements of Mask array are out of range.\n)
       pdfformaterror
       (               Output may be incorrect.\n) pdfformaterror
 
       ColorSpace dup type /arraytype eq { 0 oget } if /Indexed eq
-      BitsPerComponent 1 eq and {
+      BitsPerComponent 1 eq and
+      {
         % AR9 treats Indexed 1 bpc images specially. Bug 692852 see also 697919 and 689717.
         dup 0 oget
         dup 0 lt exch 1 BitsPerComponent bitshift ge or {
@@ -2144,37 +2146,37 @@ currentdict /last-ditch-bpc-csp undef
           % Second component is invalid; AR9 replaces it with 1.
           [ exch 0 oget 1 ]
         } ifelse
-      } {
-        % AR5, AR9 do this for most cases. Bug 690786.
-        [ exch { 1 BitsPerComponent bitshift 1 sub and } forall ]
-      } ifelse
-    } {
-      ColorSpace dup type /arraytype eq { 0 oget } if /DeviceGray eq
-      BitsPerComponent 1 eq and {
-        %% For DeviceGray, try clamping the values to 0 and 1 respectively
-        dup 0 oget
-        dup 0 lt exch 1 gt or {
-          dup 0 oget
-          dup 0 lt {
-            pop dup 0 0 put
-          }{
-            dup 0 1 put
-          } ifelse
-        } if
-        dup 1 oget
-        dup 0 lt exch 1 gt or {
-          dup 1 oget
-          dup 0 lt {
-            pop dup 0 0 put
+      }
+      {
+        BitsPerComponent 1 eq {
+          dup length 2 gt {
+            % There are too many Mask entries, discard the Mask array
+            pop //null
           }{
-            dup 0 1 put
+            % Bug 701468 and 689717, the latter requires us to set the Mask values
+            % so that the value is calculated from the BPC, the other requires us
+            % to ignore the Mask. The only difference between the two files is the
+            % actual range of the Mask values. This may yet need more work.
+            % Note! Both files are technically invalid....
+
+            aload 3 1 roll eq {
+              % There's only one Mask value, but its illegal, substitute with the
+              % calculated maximum.
+              [ exch { 1 BitsPerComponent bitshift 1 sub and } forall ]
+            } {
+              % The Mask values are not the same. This is a 2 colour image, if
+              % we mask off any range then the result will be blank, so ignore the Mask
+              pop //null
+            } ifelse
           } ifelse
-        } if
-      }{
-        % AR5, AR9 do this for most cases. Bug 690786.
-        [ exch { 1 BitsPerComponent bitshift 1 sub and } forall ]
+        } {
+          % We think AR5, AR9 do this for most other cases. Originally Bug 690786,
+          % but that is actually handled in the /Indexed case above.
+          [ exch { 1 BitsPerComponent bitshift 1 sub and } forall ]
+        } ifelse
       } ifelse
-    } ifelse
+    } if
+
     dup //null ne {
       /MaskColor exch def
     } {


Summary of changes:
 Resource/Init/pdf_draw.ps | 62 ++++++++++++++++++++++++-----------------------
 1 file changed, 32 insertions(+), 30 deletions(-)