From Wiki
Jump to navigation Jump to search

The context script has two main uses: (1) compiling ConTeXt MkIV documents, and (2) recompiling the ConTeXt core format after you change one of the source files. It is one of ConTeXt's four command-line scripts, together with Mtxrun for running general ('meta-TeX') helper scripts, First-setup for installation and updating, and texexec for compiling MkII documents.


To compile a document myfile.tex, simply pass it as an argument to context:

context myfile.tex

Remaking the format

When you change one of ConTeXt's core files – as a temporary patch, or to try out something new, or for whatever reason – the change won't take effect immediately. Instead, you must first recompile ConTeXt, a process known as '(re)making the format'. In fact, ConTeXt has several formats: the English-language interface is a separate format from the English or the Persian one.

To make all formats:

context --make

To make just the format for a single interface language:

context --make cont-en    # English
context --make cont-uk    # British English
context --make cont-de    # German
context --make cont-fr    # French
context --make cont-nl    # Dutch
context --make cont-cs    # Czech
context --make cont-it    # Italian
context --make cont-ro    # Romanian
context --make cont-pe    # Farsi (Persian)

Basic options

Argument Description
--run process (one or more) files (default action)
--make create context formats
--ctx=name use ctx file (process management specification)
--noctx ignore ctx directives and flags
--interface use specified user interface (default: en)
--autopdf close pdf file in viewer and start pdf viewer afterwards
--purge purge files either or not after a run (--pattern=...)
--purgeall purge all files either or not after a run (--pattern=...)
--usemodule=list load the given module or style, normally part of the distribution
--environment=list load the given environment file first (document styles)
--mode=list enable given the modes (conditional processing in styles)
--path=list also consult the given paths when files are looked for
--arguments=list set variables that can be consulted during a run (key/value pairs)
--randomseed=number set the randomseed
--result=name rename the resulting output to the given name
--trackers=list set tracker variables (show list with --showtrackers)
--directives=list set directive variables (show list with --showdirectives)
--silent=list disable logcatgories (show list with --showlogcategories)
--strip strip Lua code (only meant for production where no errors are expected)
--errors=list show errors at the end of a run, quit when in list (also when --silent)
--htmlerrorpage generate html error page instead (optional: =scite)
--noconsole disable logging to the console (logfile only)
--nodummy don't create a dummy file (can confuse pdf viewers that keep file in view)
--purgeresult purge result file before run
--forcexml force xml stub
--forcecld force cld (context lua document) stub
--forcelua force lua stub (like texlua)
--forcemp force mp stub
--arrange run extra imposition pass, given that the style sets up imposition
--noarrange ignore imposition specifications in the style
--pages report pagenumbers of names pages and list references (--page=... or --page=...)
--jit use luajittex with jit turned off (only use the faster virtual machine)
--jiton use luajittex with jit turned on (in most cases not faster, even slower)
--once only run once (no multipass data file is produced)
--runs process at most this many times
--forcedruns process this many times (permits for optimization trial runs)
--batchmode run without stopping and do not show messages on the console
--nonstopmode run without stopping
--nosynctex never initializes synctex (for production runs)
--synctex run with synctex enabled (better use \setupsynctex[state=start]
--nodates omit runtime dates in pdf file (optional value: a number (this 1970 offset time) or string "YYYY-MM-DD HH:MM")
--nocompression forcefully turns off compression in the backend
--trailerid alternative trailer id (or constant one)
--generate generate file database etc. (as luatools does)
--paranoid do not descend to .. and ../..
--version report installed context version
--global assume given file present elsewhere
--nofile use dummy file as jobname

Expert options

--touch                update context version number (also provide --expert, optionally provide --basepath)
--nostatistics         omit runtime statistics at the end of the run
--profile              profile job (use: mtxrun --script profile --analyze)
--timing               generate timing and statistics overview
--keeptuc              keep previous tuc files (jobname-tuc-[run].tmp)
--keeplog              keep previous log files (jobname-log-[run].tmp)
--lmtx                 force lmtx mode (when available)
--overloadmode=error|warning|0--6|255 enable csname overload checking

Special options

--pdftex               process file with texexec using pdftex
--xetex                process file with texexec using xetex
--mkii                 process file with texexec
--pipe                 do not check for file and enter scroll mode (--dummyfile=whatever.tmp)
--sandbox              process file in a limited environment
--addbinarypath        prepend the (found) binarypath to runners