summaryrefslogtreecommitdiff
path: root/doc/ref/expect.texi
diff options
context:
space:
mode:
Diffstat (limited to 'doc/ref/expect.texi')
-rw-r--r--doc/ref/expect.texi147
1 files changed, 147 insertions, 0 deletions
diff --git a/doc/ref/expect.texi b/doc/ref/expect.texi
new file mode 100644
index 000000000..140fd477e
--- /dev/null
+++ b/doc/ref/expect.texi
@@ -0,0 +1,147 @@
+@c -*-texinfo-*-
+@c This is part of the GNU Guile Reference Manual.
+@c Copyright (C) 1996, 1997, 2000, 2001, 2002, 2003, 2004
+@c Free Software Foundation, Inc.
+@c See the file guile.texi for copying conditions.
+
+@node Expect
+@section Expect
+
+The macros in this section are made available with:
+
+@lisp
+(use-modules (ice-9 expect))
+@end lisp
+
+@code{expect} is a macro for selecting actions based on the output from
+a port. The name comes from a tool of similar functionality by Don Libes.
+Actions can be taken when a particular string is matched, when a timeout
+occurs, or when end-of-file is seen on the port. The @code{expect} macro
+is described below; @code{expect-strings} is a front-end to @code{expect}
+based on regexec (see the regular expression documentation).
+
+@defmac expect-strings clause @dots{}
+By default, @code{expect-strings} will read from the current input port.
+The first term in each clause consists of an expression evaluating to
+a string pattern (regular expression). As characters
+are read one-by-one from the port, they are accumulated in a buffer string
+which is matched against each of the patterns. When a
+pattern matches, the remaining expression(s) in
+the clause are evaluated and the value of the last is returned. For example:
+
+@lisp
+(with-input-from-file "/etc/passwd"
+ (lambda ()
+ (expect-strings
+ ("^nobody" (display "Got a nobody user.\n")
+ (display "That's no problem.\n"))
+ ("^daemon" (display "Got a daemon user.\n")))))
+@end lisp
+
+The regular expression is compiled with the @code{REG_NEWLINE} flag, so
+that the ^ and $ anchors will match at any newline, not just at the start
+and end of the string.
+
+There are two other ways to write a clause:
+
+The expression(s) to evaluate
+can be omitted, in which case the result of the regular expression match
+(converted to strings, as obtained from regexec with match-pick set to "")
+will be returned if the pattern matches.
+
+The symbol @code{=>} can be used to indicate that the expression is a
+procedure which will accept the result of a successful regular expression
+match. E.g.,
+
+@lisp
+("^daemon" => write)
+("^d(aemon)" => (lambda args (for-each write args)))
+("^da(em)on" => (lambda (all sub)
+ (write all) (newline)
+ (write sub) (newline)))
+@end lisp
+
+The order of the substrings corresponds to the order in which the
+opening brackets occur.
+
+A number of variables can be used to control the behaviour
+of @code{expect} (and @code{expect-strings}).
+Most have default top-level bindings to the value @code{#f},
+which produces the default behaviour.
+They can be redefined at the
+top level or locally bound in a form enclosing the expect expression.
+
+@table @code
+@item expect-port
+A port to read characters from, instead of the current input port.
+@item expect-timeout
+@code{expect} will terminate after this number of
+seconds, returning @code{#f} or the value returned by expect-timeout-proc.
+@item expect-timeout-proc
+A procedure called if timeout occurs. The procedure takes a single argument:
+the accumulated string.
+@item expect-eof-proc
+A procedure called if end-of-file is detected on the input port. The
+procedure takes a single argument: the accumulated string.
+@item expect-char-proc
+A procedure to be called every time a character is read from the
+port. The procedure takes a single argument: the character which was read.
+@item expect-strings-compile-flags
+Flags to be used when compiling a regular expression, which are passed
+to @code{make-regexp} @xref{Regexp Functions}. The default value
+is @code{regexp/newline}.
+@item expect-strings-exec-flags
+Flags to be used when executing a regular expression, which are
+passed to regexp-exec @xref{Regexp Functions}.
+The default value is @code{regexp/noteol}, which prevents @code{$}
+from matching the end of the string while it is still accumulating,
+but still allows it to match after a line break or at the end of file.
+@end table
+
+Here's an example using all of the variables:
+
+@smalllisp
+(let ((expect-port (open-input-file "/etc/passwd"))
+ (expect-timeout 1)
+ (expect-timeout-proc
+ (lambda (s) (display "Times up!\n")))
+ (expect-eof-proc
+ (lambda (s) (display "Reached the end of the file!\n")))
+ (expect-char-proc display)
+ (expect-strings-compile-flags (logior regexp/newline regexp/icase))
+ (expect-strings-exec-flags 0))
+ (expect-strings
+ ("^nobody" (display "Got a nobody user\n"))))
+@end smalllisp
+@end defmac
+
+@defmac expect clause @dots{}
+@code{expect} is used in the same way as @code{expect-strings},
+but tests are specified not as patterns, but as procedures. The
+procedures are called in turn after each character is read from the
+port, with two arguments: the value of the accumulated string and
+a flag to indicate whether end-of-file has been reached. The flag
+will usually be @code{#f}, but if end-of-file is reached, the procedures
+are called an additional time with the final accumulated string and
+@code{#t}.
+
+The test is successful if the procedure returns a non-false value.
+
+If the @code{=>} syntax is used, then if the test succeeds it must return
+a list containing the arguments to be provided to the corresponding
+expression.
+
+In the following example, a string will only be matched at the beginning
+of the file:
+
+@lisp
+(let ((expect-port (open-input-file "/etc/passwd")))
+ (expect
+ ((lambda (s eof?) (string=? s "fnord!"))
+ (display "Got a nobody user!\n"))))
+@end lisp
+
+The control variables described for @code{expect-strings} also
+influence the behaviour of @code{expect}, with the exception of
+variables whose names begin with @code{expect-strings-}.
+@end defmac