Fix for setting of PageSpotColors Bug 704660

plmain.c should not be setting the PageSpotColors
value.  This should be set by the individual
interpreters on a per page basis.  This change
fixes a crash (Bug 704660) reduces the number of
closing and opening times of the device with pdfi
and pdfi renders more than 4 spots to the separation
devices just fine. Also for gpdl with HEAD, the subsequent
execution of a PS file following a PCL file
did not properly reset the PageSpotColors to -1
as is required for PS files, resulting in improper
output when going to a separation device.  This commit
includes a fix for that.  Tested with command line

gpdl -sDEVICE=psdcmyk -r72  -o ./myoutputs/pdl_input_%d.psd
./myinputs/owl.pcl ./myinputs/Ad_InDesign.ps
./myinputs/spots_multi_page.pdf  ./myinputs/input.xps
./myinputs/deviceNImage.eps ./myinputs/DeviceN_20Colors.pdf
./myinputs/page_spots/Ad_InDesign.ps

to verify that the number of spots was getting properly
set with each interpreter and page (when applicable) change.

1 files changed, 12 insertions, 1 deletions

diff --git a/xps/xpspage.c b/xps/xpspage.c
index dafd76ed5..6237b7a33 100644
--- a/xps/xpspage.c
+++ b/xps/xpspage.c
@@ -125,6 +125,7 @@ xps_parse_fixed_page(xps_context_t *ctx, xps_part_t *part)
     char base_uri[1024];
     char *s;
     int code, code1, code2;
+    int page_spot_colors = 0;
 
     if_debug1m('|', ctx->memory, "doc: parsing page %s\n", part->name);
 
@@ -205,9 +206,19 @@ xps_parse_fixed_page(xps_context_t *ctx, xps_part_t *part)
             }
         }
 
+        /* At some point we may want to add the pre-parse for named colors and n-channel
+           colors here. The XPS spec makes it optional to put the colorant names in the
+           ICC profile. So we would need some sort of fall back and we would need to know
+           if a name color that we encounter is one that we already encountered, which would get
+           very messy in terms of comparing ICC profiles. Especially for example, if
+           the same spot color was used individually AND in an n-channel color profile.
+           Since XPS usage is rare, and the demand for support of real spot color separation
+           non-existent, we will set the PageSpotColors to 0 at this point. */
+
+        code  = param_write_int((gs_param_list *)&list, "PageSpotColors", &(page_spot_colors));
         code1 = param_write_bool((gs_param_list *)&list, "PageUsesTransparency", &(ctx->has_transparency));
         code2 = param_write_float_array((gs_param_list *)&list, ".MediaSize", &fa);
-        if ( code1 >= 0 || code2 >= 0)
+        if ( code >= 0 || code1 >= 0 || code2 >= 0)
         {
             gs_c_param_list_read(&list);
             code = gs_putdeviceparams(dev, (gs_param_list *)&list);