The script xmlproc.sh takes an xml file as input,
and processes it to create .pdf, .html or .ps output.
It uses format.pl, a perl script to format <pre> blocks nicely,
 and add CDATA tags so writers do not have to use eg. &lt; 

The file "entities.xml" must be edited to reflect current
version, year, etc.


Usage:

  xmlproc.sh -v manual.xml
  Validates an xml file to ensure no dtd-compliance errors

  xmlproc.sh -html manual.xml
  Output: manual.html

  xmlproc.sh -pdf manual.xml
  Output: manual.pdf

  xmlproc.sh -ps manual.xml
  Output: manual.ps


Notum bene: 
- pdfxmltex barfs if given a filename with an underscore in it

- xmltex won't work yet - there's a bug in passivetex
    which we are all waiting for Sebastian to fix.
  So we are going the xml -> pdf -> ps route for the time being,
    using pdfxmltex.