[NTG-context] Movable code, was Re: starttexdefinition error in standalone, works in TL2013

12 Mar 2014

      On 2014-03-01 07:32, Wolfgang Schuster wrote:
...
\unexpanded\def\startTranslation
   {\begingroup
    \dosingleempty\dostartTranslation}
\def\dostartTranslation[#1]%
   {\iffirstargument
      \getrawparameters[Translation][setups=,language=en,#1]%
    \fi
    \grabbufferdata[Translation][startTranslation][stopTranslation]}
\def\stopTranslation
   {\language[\Translationlanguage]%
    \Translationsetups
    (\,\getbufferdata[Translation]\removeunwantedspaces\,)%
    \endgroup}
\starttext
\startTranslation[language=nl]
It betrays a slow-witted mentality to pursue the streams, but not to see the sources of things.
\stopTranslation
\stoptext
Wolfgang
Wolfgang (and list),

The above code works fine with simple texts, but is not movable. Adding 
a footnote, as in the following, to the text section shows that:

    \starttext
    \startTranslation[language=en]
       It betrays a slow-witted mentality to pursue the streams, but not to
       see the sources of things.%
       \footnote{%
         \startTranslation[language=la]
           Tardi ingeni est rivulos consectari, fontes rerum non videre.
         \stopTranslation%
       }
    \stopTranslation
    \stoptext

or even by:

    \starttext
       It betrays a slow-witted mentality to pursue the streams, but not to
       see the sources of things.%
       \footnote{%
         \startTranslation[language=la]
           Tardi ingeni est rivulos consectari, fontes rerum non videre.
         \stopTranslation%
       }
    \stoptext

In LaTeX, this might be dealt with by a combination robust command 
definitions and \protected. When I wrote an endnotes routine in LaTeX, I 
used those and wrote the notes to a file which was then read back in at 
the appropriate time, but that does not help with pagenotes.

I imagine that at the time the footnote is processed the 
Translationsetups and Translationlanguage macros are no longer defined, 
and I have no idea how one would reference the buffer contents, although 
it does look like they have stable names in the log file.

Can this be made movable?

Also, am I correct that the \iffirstargument test in the definition of 
dostarttranslation is redundant? My understanding is that \dosingleempty 
assures the presence of [#1].

-- 
Rik