Marco Patzer schrieb am 27.11.18 um 21:56:

On Tue, 27 Nov 2018 21:41:52 +0100 Wolfgang Schuster wrote:

...

...

A nudge in the right direction appreciated. \appendtoks   \mainlanguage[\documentvariable{language}]% %\doifdocumentvariable{language}{\mainlanguage[\documentvariable{language}]}% \to \everysetupdocument That's basically what I've done in my project file (te-proj.tex). At least the code is supposed to do exactly that.

But even your solution doesn't work here. Did you try it in the example project file?

\startenvironment [*] \startsetups [itemgroup:en]   \setupitemgroup [itemize] [i] \stopsetups \startsetups [itemgroup:sv]   \setupitemgroup [itemize] [a] \stopsetups \setuplanguage   [en]   [setups=itemgroup:en] \setuplanguage   [sv]   [setups=itemgroup:sv] \appendtoks \doifdocumentvariable{language}{\mainlanguage[\documentvariable{language}]} \to \everysetupdocument \stopenvironment Wolfgang

Marco Patzer schrieb am 27.11.18 um 22:23:

On Tue, 27 Nov 2018 22:05:27 +0100 Wolfgang Schuster wrote:

...

...

But even your solution doesn't work here. Did you try it in the example project file? \startenvironment [*]

\startsetups [itemgroup:en]   \setupitemgroup [itemize] [i] \stopsetups

\startsetups [itemgroup:sv]   \setupitemgroup [itemize] [a] \stopsetups

\setuplanguage   [en]   [setups=itemgroup:en]

\setuplanguage   [sv]   [setups=itemgroup:sv]

\appendtoks \doifdocumentvariable{language}{\mainlanguage[\documentvariable{language}]} \to \everysetupdocument

\stopenvironment That works, thanks.

One more question: I used to use modes, since I could just

\startmode[*en] … \stopmode

any number of times. There are many language dependent settings in the environment files.

With named setups I need to specify a name for each occurrence and gather them at the end. This is not as clean. Is it possible e.g. to add something to an already existing setup instead of creating a new one every time?

No, this is only possible with a token list. \startenvironment [*] \newtoks\EnglishSettings \newtoks\SwedishSettings \appendtoks   \setupitemgroup [itemize] [i] \to \EnglishSettings \appendtoks   \setupitemgroup [itemize] [a] \to \SwedishSettings % the order of the following settings is important because you have % to set the language before you can make the language dependant % mode setting \appendtoks \doifdocumentvariable{language}{\mainlanguage[\documentvariable{language}]} \to \everysetupdocument \appendtoks   \startmodeset     [**en] {\flushtoks\EnglishSettings}     [**sv] {\flushtoks\SwedishSettings}   \stopmodeset \to \everysetupdocument \stopenvironment Wolfgang

Marco Patzer schrieb am 28.11.18 um 08:38:

On Tue, 27 Nov 2018 22:41:25 +0100 Wolfgang Schuster wrote:

...

...

With named setups I need to specify a name for each occurrence and gather them at the end. This is not as clean. Is it possible e.g. to add something to an already existing setup instead of creating a new one every time? No, this is only possible with a token list. Works perfectly.

Your first solution is technically better since it works even if the language is changed mid document (something not used in this project). However, the second solution is cleaner. I'll make up my mind.

The solution with the token list would be nicer if there where a way to store the content in a similar fashion as buffers (which allow adding content but this requires Lua) or setups. \unprotect \installnamespace {tokenlist} \unexpanded\def\definetokenlist[#1]%   {\expandafter\newtoks\csname\????tokenlist#1\endcsname} \unexpanded\def\starttokenlist[#1]#2\stoptokenlist   {\ifcsname\????tokenlist#1\endcsname \else      \expandafter\newtoks\csname\????tokenlist#1\endcsname    \fi    \expandafter\toksapp\csname\????tokenlist#1\endcsname{#2}} \def\gettokenlist[#1]%   {\ifcsname\????tokenlist#1\endcsname      \expandafter\the\csname\????tokenlist#1\endcsname    \fi} % \unexpanded\def\definetokenlist[#1]% %   {\expandafter\let\csname\????tokenlist#1\endcsname\empty} % % \unexpanded\def\starttokenlist[#1]#2\stoptokenlist %   {\ifcsname\????tokenlist#1\endcsname \else %      \expandafter\let\csname\????tokenlist#1\endcsname\empty %    \fi %    \expandafter\edef\csname\????tokenlist#1\endcsname % {\normalunexpanded\expandafter\expandafter\expandafter{\csname\????tokenlist#1\endcsname}% %       \normalunexpanded\expandafter{#2}}} % % \def\gettokenlist[#1]% %   {\ifcsname\????tokenlist#1\endcsname %      \csname\????tokenlist#1\endcsname %    \fi} \protect \starttext % \definetokenlist[test] \starttokenlist[test] a \stoptokenlist b \starttokenlist[test] c \stoptokenlist \gettokenlist[test] \stoptext Wolfgang

On Thu, 29 Nov 2018 20:30:53 +0100 Wolfgang Schuster wrote:

Below is a improved version of the token method which works with the new beta.

\startenvironment [*]

\definetokenlist [english] % optional because \starttokenlist \definetokenlist [swedish] % checks if the token lists exists

\starttokenlist [english]   \setupitemgroup [itemize] [i] \stoptokenlist

\starttokenlist [swedish]   \setupitemgroup [itemize] [a] \stoptokenlist

\startsetups [document:start]

\doifdocumentvariable{language}{\mainlanguage[\documentvariable{language}]}

  \startmodeset     [*en] {\gettokenlist[english]}     [*sv] {\gettokenlist[swedish]}   \stopmodeset

\stopsetups

\stopenvironment

So that's basically a glorified \appendtoks … \to\… using the usual context syntax? Very neat. Marco

On Wed, 28 Nov 2018 09:20:49 +0100 Hans Hagen wrote:

An alternative is to use system modes:

\mainlanguage[nl]

% \startmode[*en] % \setupitemize[n] % \stopmode

% \startmode[*nl] % \setupitemize[a] % \stopmode

% \startmode[*de] % \setupitemize[i] % \stopmode

\startmodeset [*en] { \setupitemize[n] } [*nl] { \setupitemize[a] } [*de] { \setupitemize[i] } \stopmodeset

I do use modes and modesets in this project, but mainly in the components. In the environment files it's a bit more tricky because the language is unknown at this point and the system modes trigger the wrong language. But Wolfgang already provided a solution, which works fine. Marco