Jump to:  OMake Home • Guide Home • Guide (single-page) • Contents (short) • Contents (long)
Index:  All • Variables • Functions • Objects • Targets • Options

Chapter 11  Shell commands

Shell commands (commands to be executed by the operating system) can be freely mixed with other code.

NOTE: the syntax and shell usage is identical on all platforms, including Win32. To avoid portability problems on Win32, it is recommended that you avoid the use of the native shell interpreter cmd.

    LIB = $(dir lib)
    println(The contents of the $(LIB) directory is:)
    ls $(LIB)

11.1  Simple commands

The syntax of shell commands is similar to the syntax used by the Unix shell bash. In general, a command is a pipeline. A basic command is part of a pipeline. It is specified with the name of an executable and some arguments. Here are some examples.

    ls -AF .
    echo Hello world

The command is found using the current search path in the variable PATH[], which should define an array of directories containing executables.

A command may also be prefixed by environment variable definitions.

    # Prints "Hello world"
    env X="Hello world" Y=2 printenv X
    # Pass the include path to the Visual C++
    env include="c:\Program Files\Microsoft SDK\include" cl foo.cpp

11.2  Globbing

Commands may contain wildcard patterns. A pattern specifies a set of files through a limited kind of regular expression. Patterns are expanded before the function is executed.

   # List all files with a .c suffix
   ls *.c

   # List all files with a single character prefix, and .c suffix
   ls ?.c

   # Rename the file hello.ml to foo.ml
   mv {hello,foo}.ml

A comprehensive description of OMake glob patterns is given in Section 10.4.

11.3  Background jobs

The command may also be placed in the background by placing an ampersand after the command. Control returns to the shell without waiting for the job to complete. The job continues to run in the background.

    gcc -o hugeprogram *.c &

11.4  File redirection

Input and output can be redirected to files by using the <, >, and >& directives after the command.

    # Write to the "foo" file
    echo Hello world > foo

    # Redirect input from the foo file
    cat < foo

    # Redirect standard output and errors to the foo file
    gcc -o boo *.c >& foo

11.5  Pipelines

Pipelines are sequences of commands, where the output from each command is sent to the next. Pipelines are defined with the | and |& syntax. With | the output is redirected, but errors are not. With |& both output and errors are redirected.

   # Send the output of the ls command to the printer
   ls *.c | lpr

   # Send output and errors to jyh as email
   gcc -o hugefile *.c |& mail jyh

11.6  Conditional execution

Commands may also be composed though conditional evaluation using the || and && syntax. Every command has an integer exit code, which may be zero or some other integer. A command is said to succeed if its exit code is zero. The expression command1 && command2 executes command2 only if command1 succeeds. The expression command1 || command2 executes command2 only if command1 fails.

   # Display the x/y file if possible
   cd x && cat y

   # Run foo.exe, or print an error message
   (test -x foo.exe && foo.exe) || echo "foo.exe is not executable"

11.7  Grouping

Parenthesis are used for grouping in a pipeline or conditional command. In the following expression, the test function is used to test whether the foo.exe file is executable. If it is, the foo.exe file is executed. If the file is not executable (or if the foo.exe command fails), the message "foo.exe is not executable" is printed.

   # Run foo.exe, or print an error message
   (test -x foo.exe && foo.exe) || echo "foo.exe is not executable"

11.8  What is a shell command?

Syntactially, shell commands are any line that is not one of the following:

Commands may also be builtin (aliases). See the documentation for the Shell object for more information.

11.9  Basic builtin functions

11.9.1  echo

The echo function prints a string.

$(echo <args>)
echo <args>

11.9.2  cd

The cd function changes the current directory.

       dir : Dir

The cd function also supports a 2-argument form:

    $(cd dir, e)
       dir : Dir
       e : expression

In the two-argument form, expression e is evaluated in the directory dir. The current directory is not changed otherwise.

The behavior of the cd function can be changed with the CDPATH variable, which specifies a search path for directories. This is normally useful only in the osh command interpreter.

    CDPATH : Dir Sequence

For example, the following will change directory to the first directory ./foo, ~/dir1/foo, ~/dir2/foo.

    CDPATH[] =
    cd foo

11.10  Job control builtin functions

11.10.1  jobs

The jobs function prints a list of jobs.


11.10.2  bg

The bg function places a job in the background.

bg <pid...>

11.10.3  fg

The fg function brings a job to the foreground.

fg <pid...>

11.10.4  stop

The stop function suspends a job.

stop <pid...>

11.10.5  wait

The wait function waits for a job to finish. If no process identifiers are given, the shell waits for all jobs to complete.

wait <pid...>

11.10.6  kill

The kill function signals a job.

kill [signal] <pid...>

11.11  Command history

11.11.1  history

    $(history-index) : Int
    $(history) : String Sequence
    history-file : File
    history-length : Int

The history variables manage the command-line history in osh. They have no effect in omake.

The history-index variable is the current index into the command-line history. The history variable is the current command-line history.

The history-file variable can be redefined if you want the command-line history to be saved. The default value is ~/.omake/osh_history.

The history-length variable can be redefined to specify the maximum number of lines in the history that you want saved. The default value is 100.

Jump to:  OMake Home • Guide Home • Guide (single-page) • Contents (short) • Contents (long)
Index:  All • Variables • Functions • Objects • Targets • Options