[pdftex] "Linearized hint data" problem

[Jennifer Howlett]

Hans Hagen wrote:
> 
> At 09:48 16/08/2003 -0400, you wrote:
> 
> > > > > Does anyone know the actual cause of the problem? Is there
> > > > > something wrong with Ghostscript's optimizer? Would the problem
> > > > > persist if my colleague could be persuaded to use pdftex
> > > > > originally, and only use ghostscript at the optimizing stage?
> > > > >
> > > > >> Any help would be appreciated.
> > > > >
> > > > > Bob Howlett
> >
> >What is Ghostscript's role in all this? I am not familiar with "optimizing"
> >using Ghostscript.  Does theproblem occur if the pdflatex output is
> >viewed in Acrobat Reader?
> 
> you can use ghostscript to linearize the pdf file (bigger file, but faster
> web viewing); when i tried that last time it worked ok; pdftex output in
> itself is rather straightforward; problems normally originate in included
> graphics or faulty applications of \pdfliteral/\pdfobj cum suis.
> 
> Hans


After reading Ivo Welch's message saying that pdfopt is unreliable, I
tried
it on two of my files, documents of about 130 pages & 230 pages
respectively.
It worked fine on one, but on the other I got an error, something about
"/rangecheck in --get--". I do use \pdfliteral in the 2nd file, just to
change the color of text corresponding to links. I wouldn't have thought
my use of it was faulty -- I don't believe I have anything complicated
like color changes continuing across page breaks. I'm not using pdflatex
or context, just plain with a few added macros (for equation numbering
etc.).
Is it perhaps true that pdfopt is likely to have trouble whenever
\pdfliteral's
are used? (If so, I guess hyperref would upset it.)

Bob Howlett