[tex-live] last weekend changes

Hans Hagen pragma at wxs.nl
Sat Mar 27 23:18:23 CET 2004

At 19:09 27/03/2004, Karl Berry wrote:

>\pdfoutput is one of those control sequences (the most important one, in
>fact) to be undefined.  Setting it to 0 is not sufficient.  That's what
>the whole discussion is about.

you're right, some (probbaly plain tex based) approaches may test for 
\pdfannot instead; for instance, if a specific package only implement 


same for \pdfinfo and document info, or, \pdfmapfile{...}.

For this we shouldn't think context (which has always had a dual dvi/pdf 
concept for pdftex), nor latex (where people have specific formats for 
specific backends, which may be one of the reasons why users were/are 
unaware of pdftex having two backends - the familiar tex world confusions: 
latex is a program and pdflatex another one and pdftex is not tex and ...); 
in this case we should think 'plain tex' or 'small macro packages' or 
whatever has been written over the years; even macro code in users 
documents may test for \pdfwhatever instead of pdfoutput.

So, either undefining, or some switch or \pdfstate is needed.


ps. In context, i have \ifx\pdfoutput\undefined \chardef\pdfoutput=0 \fi 
and since all driver specific code is in driver files i don't expect any 
problems for context users (who all use pdf output anyway -).    

More information about the tex-live mailing list