[XeTeX] page size of pdf file

[Andreas Matthias]

Ross Moore wrote:

> On 19/10/2009, at 3:04 AM, Andreas Matthias wrote:
> 
>> The following example calculates the size of a pdf twice.
>> Once by using XeTeX primitives and once by using LaTeX's
>> graphics driver. Unfortunately, the size isn't the same.
>> 
>> 
>> \documentclass{article}
>> \usepackage{graphicx}
>> 
>> \begin{document}
>> \makeatletter
>> 
>> \setbox\z@=\hbox{\XeTeXpdffile "dummy.pdf"\relax}
>> \typeout{plain: \the\wd\z@, \the\ht\z@}
>> 
>> 
>> \setbox\z@=\hbox{\includegraphics{dummy.pdf}}
>> \typeout{latex: \the\wd\z@, \the\ht\z@}
>> 
>> \end{document}
>> 
>> 
>> 
>> Which yields the following:
>> 
>> plain: 597.5083pt, 845.04706pt
>> <use  "dummy.pdf" >
>> latex: 599.74895pt, 848.21597pt
>> 
> 
> I've tried this with a couple of 1-page PDFs produced
> using XeTeX, and find no such discrepancy in the sizes:

I get this discrepancy with every PDF I have tried.

> Can you supply the  dummy.pdf  that you used?

CTAN:/macros/latex/contrib/pdfpages/dummy.pdf

> Maybe my XeTeX version isn't the most recent; viz.
> 
>   This is XeTeXk, Version 3.1415926-2.2-0.999.7 (Web2C 7.5.6)
> 
> What are you using?

This is XeTeXk, Version 3.1415926-2.2-0.999.6 (Web2C 7.5.7)

> What exactly are you claiming is wrong

The size that I get by using \XeTeXpdffile directly is the
one that matches with the MediaBox of the PDF. The size I get
by using \includegraphics is too big. That's why I thought
I could be a problem with xetex.def.

> The LaTeX "graphics driver" is just an interface to the
> XeTeX primitive, so both *should* give the same result
> --- unless somehow an extra space character is getting
> inserted somewhere. If this is so, then it's not happening
> in my version of  xetex.def . Send me what you have, so that
> I can do a  diff  to check whether somehow it has become
> different after packaging and unpacking again.

I just downloaded xetex.def from CTAN again, but I still
get different sizes.

CTAN:macros/latex/contrib/xetex-def/xetex.def


Can anyone reproduce the size discrepancies that I get?


Ciao
Andreas