[tex-live] page size bug or problem

Dan Luecking luecking at uark.edu
Mon Feb 15 19:16:17 CET 2010

On Mon Feb 15 2010 at 17:42:57 CET Rainer J.H. Brandt 
<<mailto:tex-live%40tug.org?Subject=Re%3A%20%5Btex-live%5D%20page%20size%20bug%20or%20problem&In-Reply-To=%3C19321.31121.286259.574404%40t1000.bb-c.de%3E>rjhb at bb-c.de> 

>I notice that nobody commented on what I said about DVI files produced
>by plain tex (not LaTeX).  That's good, because I was wrong.
>The DVI files _do_ reflect my page specifications, except that the page
>height is always 8.435mm larger than what I requested.

You left out the details: how do you request page size in
plain tex? And do you mean text size or paper size?

The dvi format has no provision for paper size data. Setting
\hsize and \vsize simple determines the width of paragraph
lines and vertical amount of text (ignoring \headline and
\footline) at which page breaks occur.

Paper sizes are determined by the dvi post-processors, as
  1. What paper is in the printer, if printed.
  2. How xdvi (or whatever dvi previewer) is configure, if
     viewed on screen (this is just to put a white
     background rectangle on the screen).
  3. How dvips is configured or by \special commands for
     dvips, if dvips is used to generated ps.
  4. How dvipdfm(x) is configured or by \special commands for
     dvipdfm(x), if dvipdfm(x) is used to generated pdf.
  5. How GSview (or whatever PS viewer) is configured, if ps
     output is viewed on screen (again, just for the white
  6. If pdftex is use to produce pdf, the dimension registers
     \pdfpageheight and \pdfpagewidth determine what paper size
     is specified in the output file. Defaults are stored in the
     format files.

The number 8.435mm is almost exactly 24pt, which is the extra
space taken in plain tex for the footline (page number). Note
that plain tex adds this on after the amount of space determined
by \vsize. If you are measuring from top of text to bottom of
page number, then that exactly accounts for the extra 8.435mm.
Plain tex (in all distributions) has always behaved this way.


Daniel H. Luecking
Department of Mathematical Sciences
Fayetteville, Arkansas

More information about the tex-live mailing list