`load' is a macro (!) that resolves paths relative to source file dir

[bpt/guile.git] / doc / ref / scheme-scripts.texi

@c -*-texinfo-*-
@c This is part of the GNU Guile Reference Manual.
@c Copyright (C)  1996, 1997, 2000, 2001, 2002, 2003, 2004, 2005, 2010
@c   Free Software Foundation, Inc.
@c See the file guile.texi for copying conditions.

@node Guile Scripting
@section Guile Scripting

Like AWK, Perl, or any shell, Guile can interpret script files.  A Guile
script is simply a file of Scheme code with some extra information at
the beginning which tells the operating system how to invoke Guile, and
then tells Guile how to handle the Scheme code.

@menu
* The Top of a Script File::    How to start a Guile script.
* Invoking Guile::              Command line options understood by Guile.
* The Meta Switch::             Passing complex argument lists to Guile
                                from shell scripts.
* Command Line Handling::       Accessing the command line from a script.
* Scripting Examples::
@end menu


@node The Top of a Script File
@subsection The Top of a Script File

The first line of a Guile script must tell the operating system to use
Guile to evaluate the script, and then tell Guile how to go about doing
that.  Here is the simplest case:

@itemize @bullet

@item
The first two characters of the file must be @samp{#!}.

The operating system interprets this to mean that the rest of the line
is the name of an executable that can interpret the script.  Guile,
however, interprets these characters as the beginning of a multi-line
comment, terminated by the characters @samp{!#} on a line by themselves.
(This is an extension to the syntax described in R5RS, added to support
shell scripts.)

@item
Immediately after those two characters must come the full pathname to
the Guile interpreter.  On most systems, this would be
@samp{/usr/local/bin/guile}.

@item
Then must come a space, followed by a command-line argument to pass to
Guile; this should be @samp{-s}.  This switch tells Guile to run a
script, instead of soliciting the user for input from the terminal.
There are more elaborate things one can do here; see @ref{The Meta
Switch}.

@item
Follow this with a newline.

@item
The second line of the script should contain only the characters
@samp{!#} --- just like the top of the file, but reversed.  The
operating system never reads this far, but Guile treats this as the end
of the comment begun on the first line by the @samp{#!} characters.

@item
If this source code file is not ASCII or ISO-8859-1 encoded, a coding
declaration such as @code{coding: utf-8} should appear in a comment
somewhere in the first five lines of the file: see @ref{Character
Encoding of Source Files}.

@item
The rest of the file should be a Scheme program.

@end itemize

Guile reads the program, evaluating expressions in the order that they
appear.  Upon reaching the end of the file, Guile exits.


@node Invoking Guile
@subsection Invoking Guile
@cindex invocation

Here we describe Guile's command-line processing in detail.  Guile
processes its arguments from left to right, recognizing the switches
described below.  For examples, see @ref{Scripting Examples}.

@table @code

@item -s @var{script} @var{arg...}
Read and evaluate Scheme source code from the file @var{script}, as the
@code{load} function would.  After loading @var{script}, exit.  Any
command-line arguments @var{arg...} following @var{script} become the
script's arguments; the @code{command-line} function returns a list of
strings of the form @code{(@var{script} @var{arg...})}.

@item -c @var{expr} @var{arg...}
Evaluate @var{expr} as Scheme code, and then exit.  Any command-line
arguments @var{arg...} following @var{expr} become command-line arguments; the
@code{command-line} function returns a list of strings of the form
@code{(@var{guile} @var{arg...})}, where @var{guile} is the path of the
Guile executable.

@item -- @var{arg...}
Run interactively, prompting the user for expressions and evaluating
them.  Any command-line arguments @var{arg...} following the @code{--}
become command-line arguments for the interactive session; the
@code{command-line} function returns a list of strings of the form
@code{(@var{guile} @var{arg...})}, where @var{guile} is the path of the
Guile executable.

@item -L @var{directory}
Add @var{directory} to the front of Guile's module load path.  The
given directories are searched in the order given on the command line
and before any directories in the GUILE_LOAD_PATH environment
variable.  Paths added here are @emph{not} in effect during execution
of the user's @file{.guile} file.

@item -x @var{extension}
Add @var{extension} to the front of Guile's load extension list
(@pxref{Loading, @code{%load-extensions}}).  The specified extensions
are tried in the order given on the command line, and before the default
load extensions.  Extensions added here are @emph{not} in effect during
execution of the user's @file{.guile} file.

@item -l @var{file}
Load Scheme source code from @var{file}, and continue processing the
command line.

@item -e @var{function}
Make @var{function} the @dfn{entry point} of the script.  After loading
the script file (with @code{-s}) or evaluating the expression (with
@code{-c}), apply @var{function} to a list containing the program name
and the command-line arguments --- the list provided by the
@code{command-line} function.

A @code{-e} switch can appear anywhere in the argument list, but Guile
always invokes the @var{function} as the @emph{last} action it performs.
This is weird, but because of the way script invocation works under
POSIX, the @code{-s} option must always come last in the list.

The @var{function} is most often a simple symbol that names a function
that is defined in the script.  It can also be of the form @code{(@@
@var{module-name} @var{symbol})} and in that case, the symbol is
looked up in the module named @var{module-name}.

For compatibility with some versions of Guile 1.4, you can also use the
form @code{(symbol ...)} (that is, a list of only symbols that doesn't
start with @code{@@}), which is equivalent to @code{(@@ (symbol ...)
main)}, or @code{(symbol ...)  symbol} (that is, a list of only symbols
followed by a symbol), which is equivalent to @code{(@@ (symbol ...)
symbol)}.  We recommend to use the equivalent forms directly since they
correspond to the @code{(@@ ...)}  read syntax that can be used in
normal code, @xref{Using Guile Modules}.

@xref{Scripting Examples}.

@item -ds
Treat a final @code{-s} option as if it occurred at this point in the
command line; load the script here.

This switch is necessary because, although the POSIX script invocation
mechanism effectively requires the @code{-s} option to appear last, the
programmer may well want to run the script before other actions
requested on the command line.  For examples, see @ref{Scripting
Examples}.

@item \
Read more command-line arguments, starting from the second line of the
script file.  @xref{The Meta Switch}.

@item --use-srfi=@var{list}
The option @code{--use-srfi} expects a comma-separated list of numbers,
each representing a SRFI number to be loaded into the interpreter
before starting evaluating a script file or the REPL.  Additionally,
the feature identifier for the loaded SRFIs is recognized by
`cond-expand' when using this option.

@example
guile --use-srfi=8,13
@end example

@item --debug
Start with the debugging virtual machine engine.  Using the debugging VM
will enable support for VM hooks, which are needed for tracing,
breakpoints, and accurate call counts when profiling.  The debugging VM
is slower than the regular VM, though, by about 10 percent.  @xref{VM
Hooks}, for more information.

By default, the debugging VM engine is only used when entering an
interactive session.  When executing a script with @code{-s} or
@code{-c}, the normal, faster VM is used by default.

@vnew{1.8}
@item --no-debug
Do not use the debugging VM engine, even when entering an interactive
session.

@item --listen[=@var{p}]
While this program runs, listen on a local port or a path for REPL
clients.  If @var{p} starts with a number, it is assumed to be a local
port on which to listen.  If it starts with a forward slash, it is
assumed to be a path to a UNIX domain socket on which to listen.

If @var{p} is not given, the default is local port 37146.  If you look
at it upside down, it almost spells ``Guile''.  If you have netcat
installed, you should be able to @kbd{nc localhost 37146} and get a
Guile prompt.  Alternately you can fire up Emacs and connect to the
process; see @ref{Using Guile in Emacs} for more details.

Note that opening a port allows anyone who can connect to that port---in
the TCP case, any local user---to do anything Guile can do, as the user
that the Guile process is running as.  Don't use @option{--listen} on
multi-user machines.  Of course, if you don't pass @option{--listen} to
Guile, no port will be opened.

That said, @code{--listen} is great for interactive debugging and
development.

@vnew{2.0}

@item --autocompile
Compile source files automatically (default behavior).

@vnew{2.0}

@item --no-autocompile
Disable automatic source file compilation.

@vnew{2.0}

@item -h@r{, }--help
Display help on invoking Guile, and then exit.

@item -v@r{, }--version
Display the current version of Guile, and then exit.

@end table


@node The Meta Switch
@subsection The Meta Switch

Guile's command-line switches allow the programmer to describe
reasonably complicated actions in scripts.  Unfortunately, the POSIX
script invocation mechanism only allows one argument to appear on the
@samp{#!} line after the path to the Guile executable, and imposes
arbitrary limits on that argument's length.  Suppose you wrote a script
starting like this:
@example
#!/usr/local/bin/guile -e main -s
!#
(define (main args)
  (map (lambda (arg) (display arg) (display " "))
       (cdr args))
  (newline))
@end example
The intended meaning is clear: load the file, and then call @code{main}
on the command-line arguments.  However, the system will treat
everything after the Guile path as a single argument --- the string
@code{"-e main -s"} --- which is not what we want.

As a workaround, the meta switch @code{\} allows the Guile programmer to
specify an arbitrary number of options without patching the kernel.  If
the first argument to Guile is @code{\}, Guile will open the script file
whose name follows the @code{\}, parse arguments starting from the
file's second line (according to rules described below), and substitute
them for the @code{\} switch.

Working in concert with the meta switch, Guile treats the characters
@samp{#!} as the beginning of a comment which extends through the next
line containing only the characters @samp{!#}.  This sort of comment may
appear anywhere in a Guile program, but it is most useful at the top of
a file, meshing magically with the POSIX script invocation mechanism.

Thus, consider a script named @file{/u/jimb/ekko} which starts like this:
@example
#!/usr/local/bin/guile \
-e main -s
!#
(define (main args)
        (map (lambda (arg) (display arg) (display " "))
             (cdr args))
        (newline))
@end example

Suppose a user invokes this script as follows:
@example
$ /u/jimb/ekko a b c
@end example

Here's what happens:
@itemize @bullet

@item
the operating system recognizes the @samp{#!} token at the top of the
file, and rewrites the command line to:
@example
/usr/local/bin/guile \ /u/jimb/ekko a b c
@end example
This is the usual behavior, prescribed by POSIX.

@item
When Guile sees the first two arguments, @code{\ /u/jimb/ekko}, it opens
@file{/u/jimb/ekko}, parses the three arguments @code{-e}, @code{main},
and @code{-s} from it, and substitutes them for the @code{\} switch.
Thus, Guile's command line now reads:
@example
/usr/local/bin/guile -e main -s /u/jimb/ekko a b c
@end example

@item
Guile then processes these switches: it loads @file{/u/jimb/ekko} as a
file of Scheme code (treating the first three lines as a comment), and
then performs the application @code{(main "/u/jimb/ekko" "a" "b" "c")}.

@end itemize


When Guile sees the meta switch @code{\}, it parses command-line
argument from the script file according to the following rules:
@itemize @bullet

@item
Each space character terminates an argument.  This means that two
spaces in a row introduce an argument @code{""}.

@item
The tab character is not permitted (unless you quote it with the
backslash character, as described below), to avoid confusion.

@item
The newline character terminates the sequence of arguments, and will
also terminate a final non-empty argument.  (However, a newline
following a space will not introduce a final empty-string argument;
it only terminates the argument list.)

@item
The backslash character is the escape character.  It escapes backslash,
space, tab, and newline.  The ANSI C escape sequences like @code{\n} and
@code{\t} are also supported.  These produce argument constituents; the
two-character combination @code{\n} doesn't act like a terminating
newline.  The escape sequence @code{\@var{NNN}} for exactly three octal
digits reads as the character whose ASCII code is @var{NNN}.  As above,
characters produced this way are argument constituents.  Backslash
followed by other characters is not allowed.

@end itemize


@node Command Line Handling
@subsection Command Line Handling

@c This section was written and contributed by Martin Grabmueller.

The ability to accept and handle command line arguments is very
important when writing Guile scripts to solve particular problems, such
as extracting information from text files or interfacing with existing
command line applications.  This chapter describes how Guile makes
command line arguments available to a Guile script, and the utilities
that Guile provides to help with the processing of command line
arguments.

When a Guile script is invoked, Guile makes the command line arguments
accessible via the procedure @code{command-line}, which returns the
arguments as a list of strings.

For example, if the script

@example
#! /usr/local/bin/guile -s
!#
(write (command-line))
(newline)
@end example

@noindent
is saved in a file @file{cmdline-test.scm} and invoked using the command
line @code{./cmdline-test.scm bar.txt -o foo -frumple grob}, the output
is

@example
("./cmdline-test.scm" "bar.txt" "-o" "foo" "-frumple" "grob")
@end example

If the script invocation includes a @code{-e} option, specifying a
procedure to call after loading the script, Guile will call that
procedure with @code{(command-line)} as its argument.  So a script that
uses @code{-e} doesn't need to refer explicitly to @code{command-line}
in its code.  For example, the script above would have identical
behaviour if it was written instead like this:

@example
#! /usr/local/bin/guile \
-e main -s
!#
(define (main args)
  (write args)
  (newline))
@end example

(Note the use of the meta switch @code{\} so that the script invocation
can include more than one Guile option: @xref{The Meta Switch}.)

These scripts use the @code{#!} POSIX convention so that they can be
executed using their own file names directly, as in the example command
line @code{./cmdline-test.scm bar.txt -o foo -frumple grob}.  But they
can also be executed by typing out the implied Guile command line in
full, as in:

@example
$ guile -s ./cmdline-test.scm bar.txt -o foo -frumple grob
@end example

@noindent
or

@example
$ guile -e main -s ./cmdline-test2.scm bar.txt -o foo -frumple grob
@end example

Even when a script is invoked using this longer form, the arguments that
the script receives are the same as if it had been invoked using the
short form.  Guile ensures that the @code{(command-line)} or @code{-e}
arguments are independent of how the script is invoked, by stripping off
the arguments that Guile itself processes.

A script is free to parse and handle its command line arguments in any
way that it chooses.  Where the set of possible options and arguments is
complex, however, it can get tricky to extract all the options, check
the validity of given arguments, and so on.  This task can be greatly
simplified by taking advantage of the module @code{(ice-9 getopt-long)},
which is distributed with Guile, @xref{getopt-long}.


@node Scripting Examples
@subsection Scripting Examples

To start with, here are some examples of invoking Guile directly:

@table @code

@item guile -- a b c
Run Guile interactively; @code{(command-line)} will return @*
@code{("/usr/local/bin/guile" "a" "b" "c")}.

@item guile -s /u/jimb/ex2 a b c
Load the file @file{/u/jimb/ex2}; @code{(command-line)} will return @*
@code{("/u/jimb/ex2" "a" "b" "c")}.

@item guile -c '(write %load-path) (newline)'
Write the value of the variable @code{%load-path}, print a newline,
and exit.

@item guile -e main -s /u/jimb/ex4 foo
Load the file @file{/u/jimb/ex4}, and then call the function
@code{main}, passing it the list @code{("/u/jimb/ex4" "foo")}.

@item guile -l first -ds -l last -s script
Load the files @file{first}, @file{script}, and @file{last}, in that
order.  The @code{-ds} switch says when to process the @code{-s}
switch.  For a more motivated example, see the scripts below.

@end table


Here is a very simple Guile script:
@example
#!/usr/local/bin/guile -s
!#
(display "Hello, world!")
(newline)
@end example
The first line marks the file as a Guile script.  When the user invokes
it, the system runs @file{/usr/local/bin/guile} to interpret the script,
passing @code{-s}, the script's filename, and any arguments given to the
script as command-line arguments.  When Guile sees @code{-s
@var{script}}, it loads @var{script}.  Thus, running this program
produces the output:
@example
Hello, world!
@end example

Here is a script which prints the factorial of its argument:
@example
#!/usr/local/bin/guile -s
!#
(define (fact n)
  (if (zero? n) 1
    (* n (fact (- n 1)))))

(display (fact (string->number (cadr (command-line)))))
(newline)
@end example
In action:
@example
$ ./fact 5
120
$
@end example

However, suppose we want to use the definition of @code{fact} in this
file from another script.  We can't simply @code{load} the script file,
and then use @code{fact}'s definition, because the script will try to
compute and display a factorial when we load it.  To avoid this problem,
we might write the script this way:

@example
#!/usr/local/bin/guile \
-e main -s
!#
(define (fact n)
  (if (zero? n) 1
    (* n (fact (- n 1)))))

(define (main args)
  (display (fact (string->number (cadr args))))
  (newline))
@end example
This version packages the actions the script should perform in a
function, @code{main}.  This allows us to load the file purely for its
definitions, without any extraneous computation taking place.  Then we
used the meta switch @code{\} and the entry point switch @code{-e} to
tell Guile to call @code{main} after loading the script.
@example
$ ./fact 50
30414093201713378043612608166064768844377641568960512000000000000
@end example

Suppose that we now want to write a script which computes the
@code{choose} function: given a set of @var{m} distinct objects,
@code{(choose @var{n} @var{m})} is the number of distinct subsets
containing @var{n} objects each.  It's easy to write @code{choose} given
@code{fact}, so we might write the script this way:
@example
#!/usr/local/bin/guile \
-l fact -e main -s
!#
(define (choose n m)
  (/ (fact m) (* (fact (- m n)) (fact n))))

(define (main args)
  (let ((n (string->number (cadr args)))
        (m (string->number (caddr args))))
    (display (choose n m))
    (newline)))
@end example

The command-line arguments here tell Guile to first load the file
@file{fact}, and then run the script, with @code{main} as the entry
point.  In other words, the @code{choose} script can use definitions
made in the @code{fact} script.  Here are some sample runs:
@example
$ ./choose 0 4
1
$ ./choose 1 4
4
$ ./choose 2 4
6
$ ./choose 3 4
4
$ ./choose 4 4
1
$ ./choose 50 100
100891344545564193334812497256
@end example


@c Local Variables:
@c TeX-master: "guile.texi"
@c End: