home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Collection of Tools & Utilities
/
Collection of Tools and Utilities.iso
/
edit
/
xvisrc.zip
/
REGEXP.C
< prev
next >
Wrap
C/C++ Source or Header
|
1992-07-28
|
32KB
|
1,428 lines
#ifndef lint
static char *sccsid = "@(#)regexp.c 2.1 7/29/92";
static char *copyright = "Copyright (c) 1986 by University of Toronto.";
#endif
/***
* program name:
xvi
* function:
PD version of UNIX "vi" editor, with extensions.
* module name:
regexp.c
* module function:
Regular expression routines.
* history:
Regular expression routines by Henry Spencer.
Modfied for use with STEVIE (ST Editor for VI Enthusiasts,
Version 3.10) by Tony Andrews.
Adapted for use with Xvi by Chris & John Downey.
Original copyright notice appears below.
Please note that this is a modified version.
***/
/*
* regcomp and regexec -- regsub and regerror are elsewhere
*
* Copyright (c) 1986 by University of Toronto.
* Written by Henry Spencer. Not derived from licensed software.
*
* Permission is granted to anyone to use this software for any
* purpose on any computer system, and to redistribute it freely,
* subject to the following restrictions:
*
* 1. The author is not responsible for the consequences of use of
* this software, no matter how awful, even if they arise
* from defects in it.
*
* 2. The origin of this software must not be misrepresented, either
* by explicit claim or by omission.
*
* 3. Altered versions must be plainly marked as such, and must not
* be misrepresented as being the original software.
*
* Beware that some of this code is subtly aware of the way operator
* precedence is structured in regular expressions. Serious changes in
* regular-expression syntax might require a total rethink.
*
* $Log: regexp.c,v $
* Revision 1.2 88/04/28 08:09:45 tony
* First modification of the regexp library. Added an external variable
* 'reg_ic' which can be set to indicate that case should be ignored.
* Added a new parameter to regexec() to indicate that the given string
* comes from the beginning of a line and is thus eligible to match
* 'beginning-of-line'.
*
* xvi, version 1.7:
*
* Pb(P_ignorecase) replaces reg_ic.
*
* BWORD (beginning of word) & EWORD (end of word) implemented.
*
* Some strings passed to regerror() are altered slightly, for
* consistency with other error messages in xvi.
*/
#include "xvi.h"
#include "regexp.h"
#include "regmagic.h"
#ifdef MEGAMAX
overlay "regexp"
#endif
/*
* Exported functions.
*/
int cstrncmp P((char *s1, char *s2, int n));
char *cstrchr P((char *s, int c));
/*
* The "internal use only" fields in regexp.h are present to pass info from
* compile to execute that permits the execute phase to run lots faster on
* simple cases. They are:
*
* regstart char that must begin a match; '\0' if none obvious
* reganch is the match anchored (at beginning-of-line only)?
* regmust string (pointer into program) that match must include, or NULL
* regmlen length of regmust string
*
* Regstart and reganch permit very fast decisions on suitable starting points
* for a match, cutting down the work a lot. Regmust permits fast rejection
* of lines that cannot possibly match. The regmust tests are costly enough
* that regcomp() supplies a regmust only if the r.e. contains something
* potentially expensive (at present, the only such thing detected is * or +
* at the start of the r.e., which can involve a lot of backup). Regmlen is
* supplied because the test in regexec() needs it and regcomp() is computing
* it anyway.
*/
/*
* Structure for regexp "program". This is essentially a linear encoding
* of a nondeterministic finite-state machine (aka syntax charts or
* "railroad normal form" in parsing technology). Each node is an opcode
* plus a "next" pointer, possibly plus an operand. "Next" pointers of
* all nodes except BRANCH implement concatenation; a "next" pointer with
* a BRANCH on both ends of it is connecting two alternatives. (Here we
* have one of the subtle syntax dependencies: an individual BRANCH (as
* opposed to a collection of them) is never concatenated with anything
* because of operator precedence.) The operand of some types of node is
* a literal string; for others, it is a node leading into a sub-FSM. In
* particular, the operand of a BRANCH node is the first node of the branch.
* (NB this is *not* a tree structure: the tail of the branch connects
* to the thing following the set of BRANCHes.) The opcodes are:
*/
/* definition number opnd? meaning */
#define END 0 /* no End of program. */
#define BOL 1 /* no Match "" at beginning of line. */
#define EOL 2 /* no Match "" at end of line. */
#define ANY 3 /* no Match any one character. */
#define ANYOF 4 /* str Match any character in this string. */
#define ANYBUT 5 /* str Match any character not in this string. */
#define BRANCH 6 /* node Match this alternative, or the next... */
#define BACK 7 /* no Match "", "next" ptr points backward. */
#define EXACTLY 8 /* str Match this string. */
#define NOTHING 9 /* no Match empty string. */
#define STAR 10 /* node Match this (simple) thing 0 or more times. */
#define PLUS 11 /* node Match this (simple) thing 1 or more times. */
#define OPEN 20 /* no Mark this point in input as start of #n. */
/* OPEN+1 is number 1, etc. */
#define CLOSE 30 /* no Analogous to OPEN. */
#define BWORD 64 /* no Beginning of word. */
#define EWORD 65 /* no End of word. */
/*
* Opcode notes:
*
* BRANCH The set of branches constituting a single choice are hooked
* together with their "next" pointers, since precedence prevents
* anything being concatenated to any individual branch. The
* "next" pointer of the last BRANCH in a choice points to the
* thing following the whole choice. This is also where the
* final "next" pointer of each individual branch points; each
* branch starts with the operand node of a BRANCH node.
*
* BACK Normal "next" pointers all implicitly point forward; BACK
* exists to make loop structures possible.
*
* STAR,PLUS '?', and complex '*' and '+', are implemented as circular
* BRANCH structures using BACK. Simple cases (one character
* per match) are implemented with STAR and PLUS for speed
* and to minimize recursive plunges.
*
* OPEN,CLOSE ...are numbered at compile time.
*/
/*
* A node is one char of opcode followed by two chars of "next" pointer.
* "Next" pointers are stored as two 8-bit pieces, high order first. The
* value is a positive offset from the opcode of the node containing it.
* An operand, if any, simply follows the node. (Note that much of the
* code generation knows about this implicit relationship.)
*
* Using two bytes for the "next" pointer is vast overkill for most things,
* but allows patterns to get big without disasters.
*/
#define OP(p) (*(p))
#define NEXT(p) (((*((p)+1)&0377)<<8) + (*((p)+2)&0377))
#define OPERAND(p) ((p) + 3)
/*
* See regmagic.h for one further detail of program structure.
*/
/*
* Utility definitions.
*/
#ifndef CHARBITS
#define UCHARAT(p) ((int)*(unsigned char *)(p))
#else
#define UCHARAT(p) ((int)*(p)&CHARBITS)
#endif
#define FAIL(m) { regerror(m); return(NULL); }
#define ISMULT(c) ((c) == '*' || (c) == '+' || (c) == '?')
#define META "^$.[()|?+*\\"
/*
* Flags to be passed up and down.
*/
#define HASWIDTH 01 /* Known never to match null string. */
#define SIMPLE 02 /* Simple enough to be STAR/PLUS operand. */
#define SPSTART 04 /* Starts with * or +. */
#define WORST 0 /* Worst case. */
/*
* mkup - convert to upper case IF we're doing caseless compares
*/
#define mkup(c) ((Pb(P_ignorecase) && is_lower(c)) ? \
to_upper(c) : (c))
/*
* Global work variables for regcomp().
*/
static char *regparse; /* Input-scan pointer. */
static int regnpar; /* () count. */
static char regdummy;
static char *regcode; /* Code-emit pointer; ®dummy = don't. */
static long regsize; /* Code size. */
/*
* Forward declarations for regcomp()'s friends.
*/
#ifndef STATIC
# define STATIC static
#endif
STATIC char *reg P((int paren, int *flagp));
STATIC char *regbranch P((int *flagp));
STATIC char *regpiece P((int *flagp));
STATIC char *regatom P((int *flagp));
STATIC char *regnode P((int op));
STATIC char *regnext P((char *p));
STATIC void regc P((int b));
STATIC void reginsert P((int op, ch