[dev-context] minor problem with utf-8 inside \startintertext

Aditya Mahajan adityam at umich.edu
Fri Mar 9 18:56:15 CET 2007


On Fri, 9 Mar 2007, Aditya Mahajan wrote:

> On Fri, 9 Mar 2007, Mojca Miklavec wrote:
>
>> Hello,
>>
>> I have something for (most probably) Hans/Taco/Aditya to resolve:
>>
>> \enableregime[utf-8]
>> \starttext
>> \startformula\startalign
>> 	\NC a \NC = b \NR
>> \startintertext
>> Še četrta možnost
>> %
>> For those with problems with encoding, \ccaron should be enough to
>> reproduce the bug.
>> \stopintertext
>> 	\NC c \NC = d \NR
>> \stopalign\stopformula
>> \stoptext
>
> Hmm... strange. Basically, this boils down to
>
>  \enableregime[utf-8]
>  \starttext
>  \startformula\startalign
>         \NC a \NC = b \NR
>   \noalign{\vbox{\noindent \normal %Try with and without normal
>     Še četrta možnost \par}}
>   \NC c \NC = d \NR
>  \stopalign\stopformula
>  \stoptext
>
> So, for the time being add \normal inside \startintertext. I do not
> understand the cause, so I do not know if normal should be part of the
> definition of \dointertext, or if there is a bigger problem.
>

I noticed that in TeXBook, Knuth does not use \vbox for intertext 
(example on page 193). I changed the definition of \dointertext and 
removed the vbox, and everything works fine.

\unexpanded\def\dointertext#1%
   {\penalty\postdisplaypenalty
    \afterdisplayspace
    {\forgetall\noindent#1\par}%
    \penalty\predisplaypenalty
    \beforedisplayspace}

This works correctly. THe present definition is a sanitized version of 
\intertext defined in amsmath, so the question is why do they use 
\vbox in the definition of \inertext, and is it really needed? That is 
beyond my understanding of TeX.

Aditya


More information about the dev-context mailing list