On Tue, Jun 11, 2002 at 05:17:46PM +0200, Hans Hagen wrote:
> At 04:23 PM 6/10/2002 +0200, Simon Pepping wrote:
> >Hi,
> >
> >This file causes an error which I think should not occur. The error
> >does not occur when I remove \NC from \XMLDBshowElementTH. It seems to
> >me that either \csname next1\endcsname or \p!doprocessaction has the
> >wrong value.
> >
> >\tracingmacros=2\tracingcommands=2
> >
> >\def\XMLDBshowElementTH#1{\NC\csname XMLDB#1\endcsname}
> >
> >\starttext
> >\def\XMLDBElementList{revnumber,date,authorinitials,revremark,revdescription}
> >\starttabulate[|l|l|l|l|l|]
> > \processcommacommand[\XMLDBElementList]\XMLDBshowElementTH \NC\NR\ML
> >\stoptabulate
> >
> >\stoptext
>
>
> Ah, you mean the \par message? This is related to the non existing i.e.
> only living inside tex \endtemplate primitive and the message is one of
> those originating deep inside tex when tex is not really able to determine
> the problem (similar things happen with missing { or } messages when
> v/halignments are used in the wrong v/h mode).
>
> The problem is that commalist processing can be nested and has no grouping,
> which means that the internal counters get confused when used inside
> implicitly grouped situations like alignment cells. So, one way out is a
> rather dirty trick:
>
> \expanded{\globalprocesscommalist[\XMLDBElementList]}
>
> i.e. use a special version of comma list processing
Well, I get a different error message, but this does solve the problem.
> [BTW, i now have a version of bTABLE cum suis running which breaks across
> pages and has auto-repeated headers, handy for xml too]
Good. I encountered the above problem in an effort to map to a
tabulate. I have now worked out a version using bTABLE, which seems
easier to automate.
Regards, Simon
--
Simon Pepping
email: [EMAIL PROTECTED]
