[cups.general] %%PageBoundingBox in DSC-conforming PostScript

[Helge Blischke]

Till Kamppeter wrote:
> Ghostscript 8.64 produces page headers like this when converting PDF to 
> PostScript:
> 
> %%Page: 2 2
> %%BeginPageSetup
> %%PageBoundingBox: 0 0 1191 842
> GS_pswrite_3_0_1001 begin
> 1191 842 null setpagesize
> /pagesave save store 197 dict begin
> 0.1 0.1 scale
> %%EndPageSetup
> 
> Note that the "%%PageBoundingBox" line is in the "%%BeginPageSetup ... 
> %%EndPageSetup" part. The pstops CUPS filter does not search the 
> "%%PageBoundingBox" line there and adds its own "%%PageBoundingBox" line 
> (with the values from the "PageSize" option) between "%%Page" and 
> "%%BeginPageSetup":
> 
> %%Page: 2 2
> %%PageBoundingBox: 0 0 595 842
> %%BeginPageSetup
> %%PageBoundingBox: 0 0 1191 842
> GS_pswrite_3_0_1001 begin
> 1191 842 null setpagesize
> /pagesave save store 197 dict begin
> 0.1 0.1 scale
> %%EndPageSetup
> 
> It also takes the values of its own "%%PageBoundingBox" line to 
> calculate the document bounding box in the end of the file, which leads 
> to the document bounding box also having the values according to the 
> "PageSize" option and not to the document.
> 
> This breaks the printing of documents with multiple page sizes.
> 
> Is this now a bug of Ghostscript misplacing the "%%PageBoundingBox" line 
> or of the pstops filter not continuing to look for page DSC comments 
> inside the %%BeginPageSetup ... %%EndPageSetup" section?
> 
>    Till
> 

I think it is a Ghostscript bug. AFAIK, the "%%PageBoundingBox" belongs to the
page header comments, whereas the stuff between the "%%BeginPageSetup" and
"%%EndPageSetup" comment lines is thought to contain PS code that defines the
deviations ov the currend page from the general setup as defined between
"%%BeginSetup" and "%%EndSetup".

I'd suggest to file a Ghostscript bug.

By the way, how does gs 8.63 behave here?

Helge


-- 
Helge Blischke
Softwareentwicklung

H.Blischke at acm.org