[XeTeX] wrong paper size (geometry and truedimen)

[Jonathan Kew]

On 24 Jun 2006, at 7:16 pm, Pablo Rodríguez wrote:

> Using xdvipdfmx code from today afternoon (and xetex code from June  
> 21st
> at 22:55) I don't get the right papersize with geometry.
>
> Here is the file:
>
> \documentclass[a4paper,12pt,twoside,openright]{book}
>
> \usepackage[dvipdfm,compat2,paperheight=24.75truecm,
> paperwidth=17.5truecm,truedimen,margin=2truecm,mag=1050,twosideshift=1 
> 0truept,headsep=7pt,headheight=14.5pt]{geometry}
>
> \begin{document}
>
> This is a sample text.
>
> \end{document}
>
> I guess this is due to the use of the mag option in geometry, that  
> makes
> xetex or xdvipdfmx understand truecm as cm and calculates the “true
> centimeters” of the real papersize. In the example, 496.063 x 701.575
> pts, which are with mag=1050 472,44 x 668,17 “truepts".
>
> I wonder whether xetex (or xdvipdfmx) handle “true units” right  
> with and
>  without geometry (under different circumstances than the ones  
> explained
> by me).

I don't know.... sounds like something to investigate further.

As far as I know, 'true' units should be handled correctly within  
xetex; but it's possible that in the process of passing the values on  
to the output driver, something doesn't work right.

Do you get the correct paper size if you don't use the mag option?

Actually, xetex should be happiest with the [pdftex] driver option to  
geometry, as it supports the \pdfpagewidth and \pdfpageheight  
commands. But the [dvipdfm] option ought to work too, so I don't  
think that's the real problem. (Though maybe mag would be handled  
differently by the different driver options?)

> And another issue is that neither -papersize in xelatex nor -p in
> xdvipdfmx seem to change paper size of this file.

That makes sense; I would expect the paper size specified through the  
geometry package to override command-line options. These options just  
change the default that will be used when the document itself doesn't  
specify size.

JK