xref: /freebsd/contrib/one-true-awk/awk.h (revision f39dd6a9784467f0db5886012b3f4b13899be6b8)
12a55deb1SDavid E. O'Brien /****************************************************************
22a55deb1SDavid E. O'Brien Copyright (C) Lucent Technologies 1997
32a55deb1SDavid E. O'Brien All Rights Reserved
42a55deb1SDavid E. O'Brien 
52a55deb1SDavid E. O'Brien Permission to use, copy, modify, and distribute this software and
62a55deb1SDavid E. O'Brien its documentation for any purpose and without fee is hereby
72a55deb1SDavid E. O'Brien granted, provided that the above copyright notice appear in all
82a55deb1SDavid E. O'Brien copies and that both that the copyright notice and this
92a55deb1SDavid E. O'Brien permission notice and warranty disclaimer appear in supporting
102a55deb1SDavid E. O'Brien documentation, and that the name Lucent Technologies or any of
112a55deb1SDavid E. O'Brien its entities not be used in advertising or publicity pertaining
122a55deb1SDavid E. O'Brien to distribution of the software without specific, written prior
132a55deb1SDavid E. O'Brien permission.
142a55deb1SDavid E. O'Brien 
152a55deb1SDavid E. O'Brien LUCENT DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS SOFTWARE,
162a55deb1SDavid E. O'Brien INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS.
172a55deb1SDavid E. O'Brien IN NO EVENT SHALL LUCENT OR ANY OF ITS ENTITIES BE LIABLE FOR ANY
182a55deb1SDavid E. O'Brien SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
192a55deb1SDavid E. O'Brien WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER
202a55deb1SDavid E. O'Brien IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION,
212a55deb1SDavid E. O'Brien ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF
222a55deb1SDavid E. O'Brien THIS SOFTWARE.
232a55deb1SDavid E. O'Brien ****************************************************************/
242a55deb1SDavid E. O'Brien 
25c263f9bfSRuslan Ermilov #include <assert.h>
26*f39dd6a9SWarner Losh #include <stdint.h>
27*f39dd6a9SWarner Losh #include <stdbool.h>
28*f39dd6a9SWarner Losh #if __STDC_VERSION__ <= 199901L
29*f39dd6a9SWarner Losh #define noreturn
30*f39dd6a9SWarner Losh #else
31*f39dd6a9SWarner Losh #include <stdnoreturn.h>
32*f39dd6a9SWarner Losh #endif
33c263f9bfSRuslan Ermilov 
342a55deb1SDavid E. O'Brien typedef double	Awkfloat;
352a55deb1SDavid E. O'Brien 
362a55deb1SDavid E. O'Brien /* unsigned char is more trouble than it's worth */
372a55deb1SDavid E. O'Brien 
382a55deb1SDavid E. O'Brien typedef	unsigned char uschar;
392a55deb1SDavid E. O'Brien 
40*f39dd6a9SWarner Losh #define	xfree(a)	{ if ((a) != NULL) { free((void *)(intptr_t)(a)); (a) = NULL; } }
41*f39dd6a9SWarner Losh /*
42*f39dd6a9SWarner Losh  * We sometimes cheat writing read-only pointers to NUL-terminate them
43*f39dd6a9SWarner Losh  * and then put back the original value
44*f39dd6a9SWarner Losh  */
45*f39dd6a9SWarner Losh #define setptr(ptr, a)	(*(char *)(intptr_t)(ptr)) = (a)
462a55deb1SDavid E. O'Brien 
47*f39dd6a9SWarner Losh #define	NN(p)	((p) ? (p) : "(null)")	/* guaranteed non-null for DPRINTF
48813da98dSDavid E. O'Brien */
492a55deb1SDavid E. O'Brien #define	DEBUG
502a55deb1SDavid E. O'Brien #ifdef	DEBUG
51*f39dd6a9SWarner Losh #	define	DPRINTF(...)	if (dbg) printf(__VA_ARGS__)
522a55deb1SDavid E. O'Brien #else
53*f39dd6a9SWarner Losh #	define	DPRINTF(...)
542a55deb1SDavid E. O'Brien #endif
552a55deb1SDavid E. O'Brien 
56*f39dd6a9SWarner Losh extern enum compile_states {
57*f39dd6a9SWarner Losh 	RUNNING,
58*f39dd6a9SWarner Losh 	COMPILING,
59*f39dd6a9SWarner Losh 	ERROR_PRINTING
60*f39dd6a9SWarner Losh } compile_time;
61*f39dd6a9SWarner Losh 
62*f39dd6a9SWarner Losh extern bool	safe;		/* false => unsafe, true => safe */
632a55deb1SDavid E. O'Brien 
642a55deb1SDavid E. O'Brien #define	RECSIZE	(8 * 1024)	/* sets limit on records, fields, etc., etc. */
652a55deb1SDavid E. O'Brien extern int	recsize;	/* size of current record, orig RECSIZE */
662a55deb1SDavid E. O'Brien 
67*f39dd6a9SWarner Losh extern char	EMPTY[];	/* this avoid -Wwritable-strings issues */
682a55deb1SDavid E. O'Brien extern char	**FS;
692a55deb1SDavid E. O'Brien extern char	**RS;
702a55deb1SDavid E. O'Brien extern char	**ORS;
712a55deb1SDavid E. O'Brien extern char	**OFS;
722a55deb1SDavid E. O'Brien extern char	**OFMT;
732a55deb1SDavid E. O'Brien extern Awkfloat *NR;
742a55deb1SDavid E. O'Brien extern Awkfloat *FNR;
752a55deb1SDavid E. O'Brien extern Awkfloat *NF;
762a55deb1SDavid E. O'Brien extern char	**FILENAME;
772a55deb1SDavid E. O'Brien extern char	**SUBSEP;
782a55deb1SDavid E. O'Brien extern Awkfloat *RSTART;
792a55deb1SDavid E. O'Brien extern Awkfloat *RLENGTH;
802a55deb1SDavid E. O'Brien 
812a55deb1SDavid E. O'Brien extern char	*record;	/* points to $0 */
822a55deb1SDavid E. O'Brien extern int	lineno;		/* line number in awk program */
832a55deb1SDavid E. O'Brien extern int	errorflag;	/* 1 if error has occurred */
84*f39dd6a9SWarner Losh extern bool	donefld;	/* true if record broken into fields */
85*f39dd6a9SWarner Losh extern bool	donerec;	/* true if record is valid (no fld has changed */
862a55deb1SDavid E. O'Brien extern int	dbg;
872a55deb1SDavid E. O'Brien 
88*f39dd6a9SWarner Losh extern const char *patbeg;	/* beginning of pattern matched */
892a55deb1SDavid E. O'Brien extern	int	patlen;		/* length of pattern matched.  set in b.c */
902a55deb1SDavid E. O'Brien 
912a55deb1SDavid E. O'Brien /* Cell:  all information about a variable or constant */
922a55deb1SDavid E. O'Brien 
932a55deb1SDavid E. O'Brien typedef struct Cell {
942a55deb1SDavid E. O'Brien 	uschar	ctype;		/* OCELL, OBOOL, OJUMP, etc. */
952a55deb1SDavid E. O'Brien 	uschar	csub;		/* CCON, CTEMP, CFLD, etc. */
962a55deb1SDavid E. O'Brien 	char	*nval;		/* name, for variables only */
972a55deb1SDavid E. O'Brien 	char	*sval;		/* string value */
982a55deb1SDavid E. O'Brien 	Awkfloat fval;		/* value as number */
99b5253557SWarner Losh 	int	 tval;		/* type info: STR|NUM|ARR|FCN|FLD|CON|DONTFREE|CONVC|CONVO */
100b5253557SWarner Losh 	char	*fmt;		/* CONVFMT/OFMT value used to convert from number */
1012a55deb1SDavid E. O'Brien 	struct Cell *cnext;	/* ptr to next if chained */
1022a55deb1SDavid E. O'Brien } Cell;
1032a55deb1SDavid E. O'Brien 
1042a55deb1SDavid E. O'Brien typedef struct Array {		/* symbol table array */
1052a55deb1SDavid E. O'Brien 	int	nelem;		/* elements in table right now */
1062a55deb1SDavid E. O'Brien 	int	size;		/* size of tab */
1072a55deb1SDavid E. O'Brien 	Cell	**tab;		/* hash table pointers */
1082a55deb1SDavid E. O'Brien } Array;
1092a55deb1SDavid E. O'Brien 
1102a55deb1SDavid E. O'Brien #define	NSYMTAB	50	/* initial size of a symbol table */
1112a55deb1SDavid E. O'Brien extern Array	*symtab;
1122a55deb1SDavid E. O'Brien 
1132a55deb1SDavid E. O'Brien extern Cell	*nrloc;		/* NR */
1142a55deb1SDavid E. O'Brien extern Cell	*fnrloc;	/* FNR */
115b5253557SWarner Losh extern Cell	*fsloc;		/* FS */
1162a55deb1SDavid E. O'Brien extern Cell	*nfloc;		/* NF */
117b5253557SWarner Losh extern Cell	*ofsloc;	/* OFS */
118b5253557SWarner Losh extern Cell	*orsloc;	/* ORS */
119b5253557SWarner Losh extern Cell	*rsloc;		/* RS */
1202a55deb1SDavid E. O'Brien extern Cell	*rstartloc;	/* RSTART */
1212a55deb1SDavid E. O'Brien extern Cell	*rlengthloc;	/* RLENGTH */
122b5253557SWarner Losh extern Cell	*subseploc;	/* SUBSEP */
123*f39dd6a9SWarner Losh extern Cell	*symtabloc;	/* SYMTAB */
1242a55deb1SDavid E. O'Brien 
1252a55deb1SDavid E. O'Brien /* Cell.tval values: */
1262a55deb1SDavid E. O'Brien #define	NUM	01	/* number value is valid */
1272a55deb1SDavid E. O'Brien #define	STR	02	/* string value is valid */
1282a55deb1SDavid E. O'Brien #define DONTFREE 04	/* string space is not freeable */
1292a55deb1SDavid E. O'Brien #define	CON	010	/* this is a constant */
1302a55deb1SDavid E. O'Brien #define	ARR	020	/* this is an array */
1312a55deb1SDavid E. O'Brien #define	FCN	040	/* this is a function name */
1322a55deb1SDavid E. O'Brien #define FLD	0100	/* this is a field $1, $2, ... */
1332a55deb1SDavid E. O'Brien #define	REC	0200	/* this is $0 */
134b5253557SWarner Losh #define CONVC	0400	/* string was converted from number via CONVFMT */
135b5253557SWarner Losh #define CONVO	01000	/* string was converted from number via OFMT */
1362a55deb1SDavid E. O'Brien 
1372a55deb1SDavid E. O'Brien 
1382a55deb1SDavid E. O'Brien /* function types */
1392a55deb1SDavid E. O'Brien #define	FLENGTH	1
1402a55deb1SDavid E. O'Brien #define	FSQRT	2
1412a55deb1SDavid E. O'Brien #define	FEXP	3
1422a55deb1SDavid E. O'Brien #define	FLOG	4
1432a55deb1SDavid E. O'Brien #define	FINT	5
1442a55deb1SDavid E. O'Brien #define	FSYSTEM	6
1452a55deb1SDavid E. O'Brien #define	FRAND	7
1462a55deb1SDavid E. O'Brien #define	FSRAND	8
1472a55deb1SDavid E. O'Brien #define	FSIN	9
1482a55deb1SDavid E. O'Brien #define	FCOS	10
1492a55deb1SDavid E. O'Brien #define	FATAN	11
1502a55deb1SDavid E. O'Brien #define	FTOUPPER 12
1512a55deb1SDavid E. O'Brien #define	FTOLOWER 13
1522a55deb1SDavid E. O'Brien #define	FFLUSH	14
15369679fc1SWarner Losh #define FAND	15
15469679fc1SWarner Losh #define FFOR	16
15569679fc1SWarner Losh #define FXOR	17
15669679fc1SWarner Losh #define FCOMPL	18
15769679fc1SWarner Losh #define FLSHIFT	19
15869679fc1SWarner Losh #define FRSHIFT	20
159*f39dd6a9SWarner Losh #define FSYSTIME	21
160*f39dd6a9SWarner Losh #define FSTRFTIME	22
1612a55deb1SDavid E. O'Brien 
1622a55deb1SDavid E. O'Brien /* Node:  parse tree is made of nodes, with Cell's at bottom */
1632a55deb1SDavid E. O'Brien 
1642a55deb1SDavid E. O'Brien typedef struct Node {
1652a55deb1SDavid E. O'Brien 	int	ntype;
1662a55deb1SDavid E. O'Brien 	struct	Node *nnext;
1672a55deb1SDavid E. O'Brien 	int	lineno;
1682a55deb1SDavid E. O'Brien 	int	nobj;
1692a55deb1SDavid E. O'Brien 	struct	Node *narg[1];	/* variable: actual size set by calling malloc */
1702a55deb1SDavid E. O'Brien } Node;
1712a55deb1SDavid E. O'Brien 
1722a55deb1SDavid E. O'Brien #define	NIL	((Node *) 0)
1732a55deb1SDavid E. O'Brien 
1742a55deb1SDavid E. O'Brien extern Node	*winner;
1752a55deb1SDavid E. O'Brien extern Node	*nullstat;
1762a55deb1SDavid E. O'Brien extern Node	*nullnode;
1772a55deb1SDavid E. O'Brien 
1782a55deb1SDavid E. O'Brien /* ctypes */
1792a55deb1SDavid E. O'Brien #define OCELL	1
1802a55deb1SDavid E. O'Brien #define OBOOL	2
1812a55deb1SDavid E. O'Brien #define OJUMP	3
1822a55deb1SDavid E. O'Brien 
1832a55deb1SDavid E. O'Brien /* Cell subtypes: csub */
1842a55deb1SDavid E. O'Brien #define	CFREE	7
1852a55deb1SDavid E. O'Brien #define CCOPY	6
1862a55deb1SDavid E. O'Brien #define CCON	5
1872a55deb1SDavid E. O'Brien #define CTEMP	4
1882a55deb1SDavid E. O'Brien #define CNAME	3
1892a55deb1SDavid E. O'Brien #define CVAR	2
1902a55deb1SDavid E. O'Brien #define CFLD	1
1912a55deb1SDavid E. O'Brien #define	CUNK	0
1922a55deb1SDavid E. O'Brien 
1932a55deb1SDavid E. O'Brien /* bool subtypes */
1942a55deb1SDavid E. O'Brien #define BTRUE	11
1952a55deb1SDavid E. O'Brien #define BFALSE	12
1962a55deb1SDavid E. O'Brien 
1972a55deb1SDavid E. O'Brien /* jump subtypes */
1982a55deb1SDavid E. O'Brien #define JEXIT	21
1992a55deb1SDavid E. O'Brien #define JNEXT	22
2002a55deb1SDavid E. O'Brien #define	JBREAK	23
2012a55deb1SDavid E. O'Brien #define	JCONT	24
2022a55deb1SDavid E. O'Brien #define	JRET	25
2032a55deb1SDavid E. O'Brien #define	JNEXTFILE	26
2042a55deb1SDavid E. O'Brien 
2052a55deb1SDavid E. O'Brien /* node types */
2062a55deb1SDavid E. O'Brien #define NVALUE	1
2072a55deb1SDavid E. O'Brien #define NSTAT	2
2082a55deb1SDavid E. O'Brien #define NEXPR	3
2092a55deb1SDavid E. O'Brien 
2102a55deb1SDavid E. O'Brien 
2112a55deb1SDavid E. O'Brien extern	int	pairstack[], paircnt;
2122a55deb1SDavid E. O'Brien 
2132a55deb1SDavid E. O'Brien #define notlegal(n)	(n <= FIRSTTOKEN || n >= LASTTOKEN || proctab[n-FIRSTTOKEN] == nullproc)
2142a55deb1SDavid E. O'Brien #define isvalue(n)	((n)->ntype == NVALUE)
2152a55deb1SDavid E. O'Brien #define isexpr(n)	((n)->ntype == NEXPR)
2162a55deb1SDavid E. O'Brien #define isjump(n)	((n)->ctype == OJUMP)
2172a55deb1SDavid E. O'Brien #define isexit(n)	((n)->csub == JEXIT)
2182a55deb1SDavid E. O'Brien #define	isbreak(n)	((n)->csub == JBREAK)
2192a55deb1SDavid E. O'Brien #define	iscont(n)	((n)->csub == JCONT)
2202a55deb1SDavid E. O'Brien #define	isnext(n)	((n)->csub == JNEXT || (n)->csub == JNEXTFILE)
2212a55deb1SDavid E. O'Brien #define	isret(n)	((n)->csub == JRET)
2222a55deb1SDavid E. O'Brien #define isrec(n)	((n)->tval & REC)
2232a55deb1SDavid E. O'Brien #define isfld(n)	((n)->tval & FLD)
2242a55deb1SDavid E. O'Brien #define isstr(n)	((n)->tval & STR)
2252a55deb1SDavid E. O'Brien #define isnum(n)	((n)->tval & NUM)
2262a55deb1SDavid E. O'Brien #define isarr(n)	((n)->tval & ARR)
2272a55deb1SDavid E. O'Brien #define isfcn(n)	((n)->tval & FCN)
2282a55deb1SDavid E. O'Brien #define istrue(n)	((n)->csub == BTRUE)
2292a55deb1SDavid E. O'Brien #define istemp(n)	((n)->csub == CTEMP)
2302a55deb1SDavid E. O'Brien #define	isargument(n)	((n)->nobj == ARG)
2312a55deb1SDavid E. O'Brien /* #define freeable(p)	(!((p)->tval & DONTFREE)) */
2322a55deb1SDavid E. O'Brien #define freeable(p)	( ((p)->tval & (STR|DONTFREE)) == STR )
2332a55deb1SDavid E. O'Brien 
2342a55deb1SDavid E. O'Brien /* structures used by regular expression matching machinery, mostly b.c: */
2352a55deb1SDavid E. O'Brien 
236c263f9bfSRuslan Ermilov #define NCHARS	(256+3)		/* 256 handles 8-bit chars; 128 does 7-bit */
2372a55deb1SDavid E. O'Brien 				/* watch out in match(), etc. */
238*f39dd6a9SWarner Losh #define	HAT	(NCHARS+2)	/* matches ^ in regular expr */
2392a55deb1SDavid E. O'Brien #define NSTATES	32
240ae692c42SAlex Richardson #define	HAT	(NCHARS+2)	/* matches ^ in regular expr */
241ae692c42SAlex Richardson 				/* NCHARS is 2**n */
2422a55deb1SDavid E. O'Brien 
2432a55deb1SDavid E. O'Brien typedef struct rrow {
2442a55deb1SDavid E. O'Brien 	long	ltype;	/* long avoids pointer warnings on 64-bit */
2452a55deb1SDavid E. O'Brien 	union {
2462a55deb1SDavid E. O'Brien 		int i;
2472a55deb1SDavid E. O'Brien 		Node *np;
2482a55deb1SDavid E. O'Brien 		uschar *up;
2492a55deb1SDavid E. O'Brien 	} lval;		/* because Al stores a pointer in it! */
2502a55deb1SDavid E. O'Brien 	int	*lfollow;
2512a55deb1SDavid E. O'Brien } rrow;
2522a55deb1SDavid E. O'Brien 
2532a55deb1SDavid E. O'Brien typedef struct fa {
254*f39dd6a9SWarner Losh 	unsigned int	**gototab;
255*f39dd6a9SWarner Losh 	uschar	*out;
2562a55deb1SDavid E. O'Brien 	uschar	*restr;
257*f39dd6a9SWarner Losh 	int	**posns;
258*f39dd6a9SWarner Losh 	int	state_count;
259*f39dd6a9SWarner Losh 	bool	anchor;
2602a55deb1SDavid E. O'Brien 	int	use;
2612a55deb1SDavid E. O'Brien 	int	initstat;
2622a55deb1SDavid E. O'Brien 	int	curstat;
2632a55deb1SDavid E. O'Brien 	int	accept;
2642a55deb1SDavid E. O'Brien 	struct	rrow re[1];	/* variable: actual size set by calling malloc */
2652a55deb1SDavid E. O'Brien } fa;
2662a55deb1SDavid E. O'Brien 
2672a55deb1SDavid E. O'Brien 
2682a55deb1SDavid E. O'Brien #include "proto.h"
269