regexp(5) regexp(5)
NAME
regexp - regular expression compile and match routines
SYNOPSIS
#define INIT <declarations>
#define GETC() <getc code>
#define PEEKC() <peekc code>
#define UNGETC(c) <ungetc code>
#define RETURN (pointer) <return code>
#define ERROR (val) <error code>
#include <regexp.h>
char *compile (instring, expbuf, endbuf, eof)
char *instring, *expbuf, *endbuf;
int eof;
int step (string, exbuf)
char *string, *exbuf;
extern char *loc1, *loc2, *locs;
extern int circf, sed, nbra;
DESCRIPTION
This page describes general-purpose regular expression
matching routines in the form of ed(1), defined in
/usr/include/regexp.h. Programs such as ed(1), sed(1),
grep(1), bs(1), expr(1), etc., which perform regular
expression matching use this source file. In this way, only
this file need be changed to maintain regular expression
compatibility.
The interface to this file is unpleasantly complex.
Programs that include this file must have the following five
macros declared before the ``#include <regexp.h>''
statement. These macros are used by the compile routine.
GETC() Return the value of the next character
in the regular expression pattern.
Successive calls to GETC() should return
successive characters of the regular
expression.
PEEKC() Return the next character in the regular
expression. Successive calls to PEEKC()
should return the same character (which
should also be the next character
returned by GETC()).
UNGETC(c) Cause the argument c to be returned by
the next call to GETC() (and PEEKC()).
Page 1 (last mod. 1/14/87)
regexp(5) regexp(5)
No more that one character of pushback
is ever needed and this character is
guaranteed to be the last character read
by GETC(). The value of the macro
UNGETC(c) is always ignored.
RETURN(pointer) This macro is used on normal exit of the
compile routine. The value of the
argument pointer is a pointer to the
character after the last character of
the compiled regular expression. This
is useful to programs which have memory
allocation to manage.
ERROR(val) This is the abnormal return from the
compile routine. The argument val is an
error number (see table below for
meanings). This call should never
return.
ERROR MEANING
11 Range endpoint too large.
16 Bad number.
25 ``\digit'' out of range.
36 Illegal or missing delimiter.
41 No remembered search string.
42 \( \) imbalance.
43 Too many \(.
44 More than 2 numbers given in \{ \}.
45 } expected after \.
46 First number exceeds second in \{ \}.
49 [ ] imbalance.
50 Regular expression overflow.
The syntax of the compile routine is as follows:
compile(instring, expbuf, endbuf, eof)
The first parameter instring is never used explicitly by the
compile routine but is useful for programs that pass down
different pointers to input characters. It is sometimes
used in the INIT declaration (see below). Programs which
call functions to input characters or have characters in an
external array can pass down a value of ((char *) 0) for
this parameter.
The next parameter expbuf is a character pointer. It points
to the place where the compiled regular expression will be
placed.
The parameter endbuf is one more than the highest address
where the compiled regular expression may be placed. If the
Page 2 (last mod. 1/14/87)
regexp(5) regexp(5)
compiled expression cannot fit in (endbuf-expbuf) bytes, a
call to ERROR(50) is made.
The parameter eof is the character which marks the end of
the regular expression. For example, in ed(1), this
character is usually a /.
Each program that includes this file must have a #define
statement for INIT. This definition will be placed right
after the declaration for the function compile and the
opening curly brace ({). It is used for dependent
declarations and initializations. Most often it is used to
set a register variable to point the beginning of the
regular expression so that this register variable can be
used in the declarations for GETC(), PEEKC() and UNGETC().
Otherwise it can be used to declare external variables that
might be used by GETC(), PEEKC() and UNGETC(). See the
example below of the declarations taken from grep(1).
There are other functions in this file which perform actual
regular expression matching, one of which is the function
step. The call to step is as follows:
step(string,expbuf)
The first parameter to step is a pointer to a string of
characters to be checked for a match. This string should be
null terminated.
The second parameter expbuf is the compiled regular
expression which was obtained by a call of the function
compile.
The function step returns non-zero if the given string
matches the regular expression, and zero if the expressions
do not match. If there is a match, two external character
pointers are set as a side effect to the call to step. The
variable set in step is loc1. This is a pointer to the
first character that matched the regular expression. The
variable loc2, which is set by the function advance, points
to the character after the last character that matches the
regular expression. Thus if the regular expression matches
the entire line, loc1 will point to the first character of
string and loc2 will point to the null at the end of string.
step uses the external variable circf which is set by
compile if the regular expression begins with ^. If this is
set then step will try to match the regular expression to
the beginning of the string only. If more than one regular
expression is to be compiled before the first is executed
the value of circf should be saved for each compiled
expression and circf should be set to that saved value
Page 3 (last mod. 1/14/87)
regexp(5) regexp(5)
before each call to step.
The function advance is called from step with the same
arguments as step. The purpose of step is to step through
the string argument and call advance until advance returns
non-zero indicating a match or until the end of string is
reached. If one wants to constrain string to the beginning
of the line in all cases, step need not be called; simply
call advance.
When advance encounters a * or \{ \} sequence in the regular
expression, it will advance its pointer to the string to be
matched as far as possible and will recursively call itself
trying to match the rest of the string to the rest of the
regular expression. As long as there is no match, advance
will back up along the string until it finds a match or
reaches the point in the string that initially matched the *
or \{ \}. It is sometimes desirable to stop this backing up
before the initial point in the string is reached. If the
external character pointer locs is equal to the point in the
string at sometime during the backing up process, advance
will break out of the loop that backs up and will return
zero. This is used by ed(1) and sed(1) for substitutions
done globally (not just the first occurrence, but the whole
line) so, for example, expressions like s/y*//g do not loop
forever.
The additional external variables sed and nbra are used for
special purposes.
EXAMPLES
The following is an example of how the regular expression
macros and calls look from grep(1):
#define INIT register char *sp = instring;
#define GETC() (*sp++)
#define PEEKC() (*sp)
#define UNGETC(c) (--sp)
#define RETURN(c) return;
#define ERROR(c) regerr()
#include <regexp.h>
...
(void) compile(*argv, expbuf, &expbuf[ESIZE], '\0');
...
if (step(linebuf, expbuf))
succeed();
FILES
/usr/include/regexp.h
SEE ALSO
Page 4 (last mod. 1/14/87)
regexp(5) regexp(5)
bs(1), ed(1), expr(1), grep(1), sed(1).
BUGS
The handling of circf is kludgy.
The actual code is probably easier to understand than this
manual page.
Page 5 (last mod. 1/14/87)