1 /****************************************************************
2 Copyright (C) Lucent Technologies 1997
3 All Rights Reserved
4 
5 Permission to use, copy, modify, and distribute this software and
6 its documentation for any purpose and without fee is hereby
7 granted, provided that the above copyright notice appear in all
8 copies and that both that the copyright notice and this
9 permission notice and warranty disclaimer appear in supporting
10 documentation, and that the name Lucent Technologies or any of
11 its entities not be used in advertising or publicity pertaining
12 to distribution of the software without specific, written prior
13 permission.
14 
15 LUCENT DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS SOFTWARE,
16 INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS.
17 IN NO EVENT SHALL LUCENT OR ANY OF ITS ENTITIES BE LIABLE FOR ANY
18 SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
19 WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER
20 IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION,
21 ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF
22 THIS SOFTWARE.
23 ****************************************************************/
24 
25 #include <assert.h>
26 #include <stdint.h>
27 #include <stdbool.h>
28 #if __STDC_VERSION__ <= 199901L
29 #define noreturn __dead
30 #else
31 #include <stdnoreturn.h>
32 #endif
33 
34 typedef double      Awkfloat;
35 
36 /* unsigned char is more trouble than it's worth */
37 
38 typedef   unsigned char uschar;
39 
40 #define   xfree(a)  { free((void *)(intptr_t)(a)); (a) = NULL; }
41 /*
42  * We sometimes cheat writing read-only pointers to NUL-terminate them
43  * and then put back the original value
44  */
45 #define setptr(ptr, a)        (*(char *)(intptr_t)(ptr)) = (a)
46 
47 #define   NN(p)     ((p) ? (p) : "(null)")        /* guaranteed non-null for DPRINTF
48 */
49 #define   DEBUG
50 #ifdef    DEBUG
51 #         define    DPRINTF(...)        if (dbg) printf(__VA_ARGS__)
52 #else
53 #         define    DPRINTF(...)
54 #endif
55 
56 extern enum compile_states {
57           RUNNING,
58           COMPILING,
59           ERROR_PRINTING
60 } compile_time;
61 
62 extern bool         safe;               /* false => unsafe, true => safe */
63 
64 #define   RECSIZE   (8 * 1024)          /* sets limit on records, fields, etc., etc. */
65 extern int          recsize;  /* size of current record, orig RECSIZE */
66 
67 extern size_t       awk_mb_cur_max;     /* max size of a multi-byte character */
68 
69 extern char         EMPTY[];  /* this avoid -Wwritable-strings issues */
70 extern char         **FS;
71 extern char         **RS;
72 extern char         **ORS;
73 extern char         **OFS;
74 extern char         **OFMT;
75 extern Awkfloat *NR;
76 extern Awkfloat *FNR;
77 extern Awkfloat *NF;
78 extern char         **FILENAME;
79 extern char         **SUBSEP;
80 extern Awkfloat *RSTART;
81 extern Awkfloat *RLENGTH;
82 
83 extern bool         CSV;                /* true for csv input */
84 
85 extern char         *record;  /* points to $0 */
86 extern int          lineno;             /* line number in awk program */
87 extern int          errorflag;          /* 1 if error has occurred */
88 extern bool         donefld;  /* true if record broken into fields */
89 extern bool         donerec;  /* true if record is valid (no fld has changed */
90 extern int          dbg;
91 
92 extern const char *patbeg;    /* beginning of pattern matched */
93 extern    int       patlen;             /* length of pattern matched.  set in b.c */
94 
95 /* Cell:  all information about a variable or constant */
96 
97 typedef struct Cell {
98           uschar    ctype;              /* OCELL, OBOOL, OJUMP, etc. */
99           uschar    csub;               /* CCON, CTEMP, CFLD, etc. */
100           char      *nval;              /* name, for variables only */
101           char      *sval;              /* string value */
102           Awkfloat fval;                /* value as number */
103           int        tval;              /* type info: STR|NUM|ARR|FCN|FLD|CON|DONTFREE|CONVC|CONVO */
104           char      *fmt;               /* CONVFMT/OFMT value used to convert from number */
105           struct Cell *cnext; /* ptr to next if chained */
106 } Cell;
107 
108 typedef struct Array {                  /* symbol table array */
109           int       nelem;              /* elements in table right now */
110           int       size;               /* size of tab */
111           Cell      **tab;              /* hash table pointers */
112 } Array;
113 
114 #define   NSYMTAB   50        /* initial size of a symbol table */
115 extern Array        *symtab;
116 
117 extern Cell         *nrloc;             /* NR */
118 extern Cell         *fnrloc;  /* FNR */
119 extern Cell         *fsloc;             /* FS */
120 extern Cell         *nfloc;             /* NF */
121 extern Cell         *ofsloc;  /* OFS */
122 extern Cell         *orsloc;  /* ORS */
123 extern Cell         *rsloc;             /* RS */
124 extern Cell         *rstartloc;         /* RSTART */
125 extern Cell         *rlengthloc;        /* RLENGTH */
126 extern Cell         *subseploc;         /* SUBSEP */
127 extern Cell         *symtabloc;         /* SYMTAB */
128 
129 /* Cell.tval values: */
130 #define   NUM       01        /* number value is valid */
131 #define   STR       02        /* string value is valid */
132 #define DONTFREE 04 /* string space is not freeable */
133 #define   CON       010       /* this is a constant */
134 #define   ARR       020       /* this is an array */
135 #define   FCN       040       /* this is a function name */
136 #define FLD         0100      /* this is a field $1, $2, ... */
137 #define   REC       0200      /* this is $0 */
138 #define CONVC       0400      /* string was converted from number via CONVFMT */
139 #define CONVO       01000     /* string was converted from number via OFMT */
140 
141 
142 /* function types */
143 #define   FLENGTH   1
144 #define   FSQRT     2
145 #define   FEXP      3
146 #define   FLOG      4
147 #define   FINT      5
148 #define   FSYSTEM   6
149 #define   FRAND     7
150 #define   FSRAND    8
151 #define   FSIN      9
152 #define   FCOS      10
153 #define   FATAN     11
154 #define   FTOUPPER 12
155 #define   FTOLOWER 13
156 #define   FFLUSH    14
157 #define FAND        15
158 #define FFOR        16
159 #define FXOR        17
160 #define FCOMPL      18
161 #define FLSHIFT     19
162 #define FRSHIFT     20
163 #define FSYSTIME    21
164 #define FSTRFTIME   22
165 #define FMKTIME     23
166 
167 /* Node:  parse tree is made of nodes, with Cell's at bottom */
168 
169 typedef struct Node {
170           int       ntype;
171           struct    Node *nnext;
172           int       lineno;
173           int       nobj;
174           struct    Node *narg[1];      /* variable: actual size set by calling malloc */
175 } Node;
176 
177 #define   NIL       ((Node *) 0)
178 
179 extern Node         *winner;
180 extern Node         *nullnode;
181 
182 /* ctypes */
183 #define OCELL       1
184 #define OBOOL       2
185 #define OJUMP       3
186 
187 /* Cell subtypes: csub */
188 #define   CFREE     7
189 #define CCOPY       6
190 #define CCON        5
191 #define CTEMP       4
192 #define CNAME       3
193 #define CVAR        2
194 #define CFLD        1
195 #define   CUNK      0
196 
197 /* bool subtypes */
198 #define BTRUE       11
199 #define BFALSE      12
200 
201 /* jump subtypes */
202 #define JEXIT       21
203 #define JNEXT       22
204 #define   JBREAK    23
205 #define   JCONT     24
206 #define   JRET      25
207 #define   JNEXTFILE 26
208 
209 /* node types */
210 #define NVALUE      1
211 #define NSTAT       2
212 #define NEXPR       3
213 
214 
215 extern    int       pairstack[], paircnt;
216 
217 #define notlegal(n) (n <= FIRSTTOKEN || n >= LASTTOKEN || proctab[n-FIRSTTOKEN] == nullproc)
218 #define isvalue(n)  ((n)->ntype == NVALUE)
219 #define isexpr(n)   ((n)->ntype == NEXPR)
220 #define isjump(n)   ((n)->ctype == OJUMP)
221 #define isexit(n)   ((n)->csub == JEXIT)
222 #define   isbreak(n)          ((n)->csub == JBREAK)
223 #define   iscont(n) ((n)->csub == JCONT)
224 #define   isnext(n) ((n)->csub == JNEXT || (n)->csub == JNEXTFILE)
225 #define   isret(n)  ((n)->csub == JRET)
226 #define isrec(n)    ((n)->tval & REC)
227 #define isfld(n)    ((n)->tval & FLD)
228 #define isstr(n)    ((n)->tval & STR)
229 #define isnum(n)    ((n)->tval & NUM)
230 #define isarr(n)    ((n)->tval & ARR)
231 #define isfcn(n)    ((n)->tval & FCN)
232 #define istrue(n)   ((n)->csub == BTRUE)
233 #define istemp(n)   ((n)->csub == CTEMP)
234 #define   isargument(n)       ((n)->nobj == ARG)
235 /* #define freeable(p)        (!((p)->tval & DONTFREE)) */
236 #define freeable(p) ( ((p)->tval & (STR|DONTFREE)) == STR )
237 
238 /* structures used by regular expression matching machinery, mostly b.c: */
239 
240 #define NCHARS      (1256+3)            /* 256 handles 8-bit chars; 128 does 7-bit */
241                                         /* BUG: some overflows (caught) if we use 256 */
242                                         /* watch out in match(), etc. */
243 #define   HAT       (NCHARS+2)          /* matches ^ in regular expr */
244 #define NSTATES     32
245 
246 typedef struct rrow {
247           long      ltype;    /* long avoids pointer warnings on 64-bit */
248           union {
249                     int i;
250                     Node *np;
251                     uschar *up;
252                     int *rp; /* rune representation of char class */
253           } lval;             /* because Al stores a pointer in it! */
254           int       *lfollow;
255 } rrow;
256 
257 typedef struct gtte { /* gototab entry */
258           unsigned int ch;
259           unsigned int state;
260 } gtte;
261 
262 typedef struct gtt {          /* gototab */
263           size_t    allocated;
264           size_t    inuse;
265           gtte      *entries;
266 } gtt;
267 
268 typedef struct fa {
269           gtt       *gototab;
270           uschar    *out;
271           uschar    *restr;
272           int       **posns;
273           int       state_count;
274           bool      anchor;
275           int       use;
276           int       initstat;
277           int       curstat;
278           int       accept;
279           struct    rrow re[1];         /* variable: actual size set by calling malloc */
280 } fa;
281 
282 
283 #include "proto.h"
284