🚀 go-pugleaf

In article <repasky.715398745@sun.bdc.ubc.ca> repasky@bdc.ubc.ca (Dick Repasky) writes:
> I've made the following attempt:
> 
>     \def\doit#1{\expandafter\newinsert\csname big#1\endcsname}
> 
> and I receive the error that \newinsert is a forbidden command sequence in
> a macro definition.  Does anyone have advice?
> 

The problem arises because all the \new... macros (\newcount, \newdimen,
..., \newinsert, etc.) are defined as \outer in plain TeX, and so are
``not allowed to appear in any place where tokens are being absorbed at
high speed.''  An \outer macro cannot appear in the replacement text of
a definition, among other places (The TeXbook, p.206).

I do not know why this decision (i.e., to have all allocations to be
global) was made; it sure is a nuisance to have to first name all the
``variables'' (i.e., give a name to the registers) that you will ever
need, instead of have them ``declared'' when you actually need them.
(And since all allocations are \global, you also cannot get away with
temporary ``declarations'' inside \begingroup and \endgroup.)


> Note that this proceedure works wonderfully if I change it to generate macro 
> definitions on macros that have different names.  The problem arises
> when any \new"register" sequence is to be defined.
> 

Figures, because \def is different from \new<whatever>.


> My goal is to be able to allocate multiple insert lists and operate on them
> from a common macro.
> 

Short of redefining \new... to the exact same thing, but without the
\outer, I do not see any way out.  But this in not very desirable,
since ``the inside story of how allocation is actually performed should
not be relevant when the allocation macros are used at a higher level;
you mustn't assume that plain.tex really does allocation in any
particular way'' (The TeXbook, p.346).

Maybe the Wizards will soon followup.

-- 
---------------------------------------------------------------------------
Navindra S Gambhir                                   gambhir@cs.cornell.edu

I would rather write programs to help me write programs than write programs