src/include/regex/regguts.h

   1 /*
   2  * Internal interface definitions, etc., for the reg package
   3  *
   4  * Copyright (c) 1998, 1999 Henry Spencer.      All rights reserved.
   5  *
   6  * Development of this software was funded, in part, by Cray Research Inc.,
   7  * UUNET Communications Services Inc., Sun Microsystems Inc., and Scriptics
   8  * Corporation, none of whom are responsible for the results.  The author
   9  * thanks all of them.
  10  *
  11  * Redistribution and use in source and binary forms -- with or without
  12  * modification -- are permitted for any purpose, provided that
  13  * redistributions in source form retain this entire copyright notice and
  14  * indicate the origin and nature of any modifications.
  15  *
  16  * I'd appreciate being given credit for this package in the documentation
  17  * of software which uses it, but that is not a requirement.
  18  *
  19  * THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES,
  20  * INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY
  21  * AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL
  22  * HENRY SPENCER BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
  23  * EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
  24  * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS;
  25  * OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
  26  * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
  27  * OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
  28  * ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
  29  *
  30  * $PostgreSQL$
  31  */
  32
  33
  34
  35 /*
  36  * Environmental customization.  It should not (I hope) be necessary to
  37  * alter the file you are now reading -- regcustom.h should handle it all,
  38  * given care here and elsewhere.
  39  */
  40 #include "regcustom.h"
  41
  42
  43
  44 /*
  45  * Things that regcustom.h might override.
  46  */
  47
  48 /* assertions */
  49 #ifndef assert
  50 #ifndef REG_DEBUG
  51 #define  NDEBUG                                 /* no assertions */
  52 #endif
  53 #include <assert.h>
  54 #endif
  55
  56 /* voids */
  57 #ifndef DISCARD
  58 #define DISCARD void                    /* for throwing values away */
  59 #endif
  60 #ifndef VS
  61 #define VS(x)   ((void *)(x))   /* cast something to generic ptr */
  62 #endif
  63
  64 /* function-pointer declarator */
  65 #ifndef FUNCPTR
  66 #define FUNCPTR(name, args) (*(name)) args
  67 #endif
  68
  69 /* memory allocation */
  70 #ifndef MALLOC
  71 #define MALLOC(n)       malloc(n)
  72 #endif
  73 #ifndef REALLOC
  74 #define REALLOC(p, n)   realloc(VS(p), n)
  75 #endif
  76 #ifndef FREE
  77 #define FREE(p)         free(VS(p))
  78 #endif
  79
  80 /* want size of a char in bits, and max value in bounded quantifiers */
  81 #ifndef CHAR_BIT
  82 #include <limits.h>
  83 #endif
  84 #ifndef _POSIX2_RE_DUP_MAX
  85 #define _POSIX2_RE_DUP_MAX      255 /* normally from <limits.h> */
  86 #endif
  87
  88
  89
  90 /*
  91  * misc
  92  */
  93
  94 #define NOTREACHED      0
  95 #define xxx             1
  96
  97 #define DUPMAX  _POSIX2_RE_DUP_MAX
  98 #define INFINITY        (DUPMAX+1)
  99
 100 #define REMAGIC 0xfed7                  /* magic number for main struct */
 101
 102
 103
 104 /*
 105  * debugging facilities
 106  */
 107 #ifdef REG_DEBUG
 108 /* FDEBUG does finite-state tracing */
 109 #define FDEBUG(arglist) { if (v->eflags&REG_FTRACE) printf arglist; }
 110 /* MDEBUG does higher-level tracing */
 111 #define MDEBUG(arglist) { if (v->eflags&REG_MTRACE) printf arglist; }
 112 #else
 113 #define FDEBUG(arglist) {}
 114 #define MDEBUG(arglist) {}
 115 #endif
 116
 117
 118
 119 /*
 120  * bitmap manipulation
 121  */
 122 #define UBITS   (CHAR_BIT * sizeof(unsigned))
 123 #define BSET(uv, sn)    ((uv)[(sn)/UBITS] |= (unsigned)1 << ((sn)%UBITS))
 124 #define ISBSET(uv, sn)  ((uv)[(sn)/UBITS] & ((unsigned)1 << ((sn)%UBITS)))
 125
 126
 127
 128 /*
 129  * We dissect a chr into byts for colormap table indexing.      Here we define
 130  * a byt, which will be the same as a byte on most machines...  The exact
 131  * size of a byt is not critical, but about 8 bits is good, and extraction
 132  * of 8-bit chunks is sometimes especially fast.
 133  */
 134 #ifndef BYTBITS
 135 #define BYTBITS 8                               /* bits in a byt */
 136 #endif
 137 #define BYTTAB  (1<<BYTBITS)    /* size of table with one entry per byt value */
 138 #define BYTMASK (BYTTAB-1)              /* bit mask for byt */
 139 #define NBYTS   ((CHRBITS+BYTBITS-1)/BYTBITS)
 140 /* the definition of GETCOLOR(), below, assumes NBYTS <= 4 */
 141
 142
 143
 144 /*
 145  * As soon as possible, we map chrs into equivalence classes -- "colors" --
 146  * which are of much more manageable number.
 147  */
 148 typedef short color;                    /* colors of characters */
 149 typedef int pcolor;                             /* what color promotes to */
 150
 151 #define COLORLESS       (-1)            /* impossible color */
 152 #define WHITE           0                       /* default color, parent of all others */
 153
 154
 155
 156 /*
 157  * A colormap is a tree -- more precisely, a DAG -- indexed at each level
 158  * by a byt of the chr, to map the chr to a color efficiently.  Because
 159  * lower sections of the tree can be shared, it can exploit the usual
 160  * sparseness of such a mapping table.  The tree is always NBYTS levels
 161  * deep (in the past it was shallower during construction but was "filled"
 162  * to full depth at the end of that); areas that are unaltered as yet point
 163  * to "fill blocks" which are entirely WHITE in color.
 164  */
 165
 166 /* the tree itself */
 167 struct colors
 168 {
 169         color           ccolor[BYTTAB];
 170 };
 171 struct ptrs
 172 {
 173         union tree *pptr[BYTTAB];
 174 };
 175 union tree
 176 {
 177         struct colors colors;
 178         struct ptrs ptrs;
 179 };
 180
 181 #define tcolor  colors.ccolor
 182 #define tptr    ptrs.pptr
 183
 184 /* internal per-color descriptor structure for the color machinery */
 185 struct colordesc
 186 {
 187         uchr            nchrs;                  /* number of chars of this color */
 188         color           sub;                    /* open subcolor (if any); free chain ptr */
 189 #define  NOSUB   COLORLESS
 190         struct arc *arcs;                       /* color chain */
 191         int                     flags;
 192 #define  FREECOL 01                             /* currently free */
 193 #define  PSEUDO  02                             /* pseudocolor, no real chars */
 194 #define  UNUSEDCOLOR(cd) ((cd)->flags&FREECOL)
 195         union tree *block;                      /* block of solid color, if any */
 196 };
 197
 198 /* the color map itself */
 199 struct colormap
 200 {
 201         int                     magic;
 202 #define  CMMAGIC 0x876
 203         struct vars *v;                         /* for compile error reporting */
 204         size_t          ncds;                   /* number of colordescs */
 205         size_t          max;                    /* highest in use */
 206         color           free;                   /* beginning of free chain (if non-0) */
 207         struct colordesc *cd;
 208 #define  CDEND(cm)       (&(cm)->cd[(cm)->max + 1])
 209 #define  NINLINECDS  ((size_t)10)
 210         struct colordesc cdspace[NINLINECDS];
 211         union tree      tree[NBYTS];    /* tree top, plus fill blocks */
 212 };
 213
 214 /* optimization magic to do fast chr->color mapping */
 215 #define B0(c)   ((c) & BYTMASK)
 216 #define B1(c)   (((c)>>BYTBITS) & BYTMASK)
 217 #define B2(c)   (((c)>>(2*BYTBITS)) & BYTMASK)
 218 #define B3(c)   (((c)>>(3*BYTBITS)) & BYTMASK)
 219 #if NBYTS == 1
 220 #define GETCOLOR(cm, c) ((cm)->tree->tcolor[B0(c)])
 221 #endif
 222 /* beware, for NBYTS>1, GETCOLOR() is unsafe -- 2nd arg used repeatedly */
 223 #if NBYTS == 2
 224 #define GETCOLOR(cm, c) ((cm)->tree->tptr[B1(c)]->tcolor[B0(c)])
 225 #endif
 226 #if NBYTS == 4
 227 #define GETCOLOR(cm, c) ((cm)->tree->tptr[B3(c)]->tptr[B2(c)]->tptr[B1(c)]->tcolor[B0(c)])
 228 #endif
 229
 230
 231 /*
 232  * Interface definitions for locale-interface functions in locale.c.
 233  */
 234
 235 /* Representation of a set of characters. */
 236 struct cvec
 237 {
 238         int                     nchrs;                  /* number of chrs */
 239         int                     chrspace;               /* number of chrs possible */
 240         chr                *chrs;                       /* pointer to vector of chrs */
 241         int                     nranges;                /* number of ranges (chr pairs) */
 242         int                     rangespace;             /* number of chrs possible */
 243         chr                *ranges;                     /* pointer to vector of chr pairs */
 244         /* both batches of chrs are on the end */
 245 };
 246
 247
 248 /*
 249  * definitions for NFA internal representation
 250  *
 251  * Having a "from" pointer within each arc may seem redundant, but it
 252  * saves a lot of hassle.
 253  */
 254 struct state;
 255
 256 struct arc
 257 {
 258         int                     type;
 259 #define  ARCFREE '\0'
 260         color           co;
 261         struct state *from;                     /* where it's from (and contained within) */
 262         struct state *to;                       /* where it's to */
 263         struct arc *outchain;           /* *from's outs chain or free chain */
 264 #define  freechain       outchain
 265         struct arc *inchain;            /* *to's ins chain */
 266         struct arc *colorchain;         /* color's arc chain */
 267         struct arc *colorchainRev;      /* back-link in color's arc chain */
 268 };
 269
 270 struct arcbatch
 271 {                                                               /* for bulk allocation of arcs */
 272         struct arcbatch *next;
 273 #define  ABSIZE  10
 274         struct arc      a[ABSIZE];
 275 };
 276
 277 struct state
 278 {
 279         int                     no;
 280 #define  FREESTATE       (-1)
 281         char            flag;                   /* marks special states */
 282         int                     nins;                   /* number of inarcs */
 283         struct arc *ins;                        /* chain of inarcs */
 284         int                     nouts;                  /* number of outarcs */
 285         struct arc *outs;                       /* chain of outarcs */
 286         struct arc *free;                       /* chain of free arcs */
 287         struct state *tmp;                      /* temporary for traversal algorithms */
 288         struct state *next;                     /* chain for traversing all */
 289         struct state *prev;                     /* back chain */
 290         struct arcbatch oas;            /* first arcbatch, avoid malloc in easy case */
 291         int                     noas;                   /* number of arcs used in first arcbatch */
 292 };
 293
 294 struct nfa
 295 {
 296         struct state *pre;                      /* pre-initial state */
 297         struct state *init;                     /* initial state */
 298         struct state *final;            /* final state */
 299         struct state *post;                     /* post-final state */
 300         int                     nstates;                /* for numbering states */
 301         struct state *states;           /* state-chain header */
 302         struct state *slast;            /* tail of the chain */
 303         struct state *free;                     /* free list */
 304         struct colormap *cm;            /* the color map */
 305         color           bos[2];                 /* colors, if any, assigned to BOS and BOL */
 306         color           eos[2];                 /* colors, if any, assigned to EOS and EOL */
 307         size_t          size;                   /* Current NFA size; differs from nstates as
 308                                                                  * it also counts the number of states created
 309                                                                  * by children of this state. */
 310         struct vars *v;                         /* simplifies compile error reporting */
 311         struct nfa *parent;                     /* parent NFA, if any */
 312 };
 313
 314
 315
 316 /*
 317  * definitions for compacted NFA
 318  */
 319 struct carc
 320 {
 321         color           co;                             /* COLORLESS is list terminator */
 322         int                     to;                             /* state number */
 323 };
 324
 325 struct cnfa
 326 {
 327         int                     nstates;                /* number of states */
 328         int                     ncolors;                /* number of colors */
 329         int                     flags;
 330 #define  HASLACONS       01                     /* uses lookahead constraints */
 331         int                     pre;                    /* setup state number */
 332         int                     post;                   /* teardown state number */
 333         color           bos[2];                 /* colors, if any, assigned to BOS and BOL */
 334         color           eos[2];                 /* colors, if any, assigned to EOS and EOL */
 335         struct carc **states;           /* vector of pointers to outarc lists */
 336         struct carc *arcs;                      /* the area for the lists */
 337 };
 338
 339 #define ZAPCNFA(cnfa)   ((cnfa).nstates = 0)
 340 #define NULLCNFA(cnfa)  ((cnfa).nstates == 0)
 341
 342 /*
 343  * Used to limit the maximum NFA size to something sane. [Tcl Bug 1810264]
 344  */
 345 #ifndef REG_MAX_STATES
 346 #define REG_MAX_STATES  100000
 347 #endif
 348
 349 /*
 350  * subexpression tree
 351  */
 352 struct subre
 353 {
 354         char            op;                             /* '|', '.' (concat), 'b' (backref), '(', '=' */
 355         char            flags;
 356 #define  LONGER  01                             /* prefers longer match */
 357 #define  SHORTER 02                             /* prefers shorter match */
 358 #define  MIXED   04                             /* mixed preference below */
 359 #define  CAP 010                                /* capturing parens below */
 360 #define  BACKR   020                    /* back reference below */
 361 #define  INUSE   0100                   /* in use in final tree */
 362 #define  LOCAL   03                             /* bits which may not propagate up */
 363 #define  LMIX(f) ((f)<<2)               /* LONGER -> MIXED */
 364 #define  SMIX(f) ((f)<<1)               /* SHORTER -> MIXED */
 365 #define  UP(f)   (((f)&~LOCAL) | (LMIX(f) & SMIX(f) & MIXED))
 366 #define  MESSY(f)        ((f)&(MIXED|CAP|BACKR))
 367 #define  PREF(f) ((f)&LOCAL)
 368 #define  PREF2(f1, f2)   ((PREF(f1) != 0) ? PREF(f1) : PREF(f2))
 369 #define  COMBINE(f1, f2) (UP((f1)|(f2)) | PREF2(f1, f2))
 370         short           retry;                  /* index into retry memory */
 371         int                     subno;                  /* subexpression number (for 'b' and '(') */
 372         short           min;                    /* min repetitions, for backref only */
 373         short           max;                    /* max repetitions, for backref only */
 374         struct subre *left;                     /* left child, if any (also freelist chain) */
 375         struct subre *right;            /* right child, if any */
 376         struct state *begin;            /* outarcs from here... */
 377         struct state *end;                      /* ...ending in inarcs here */
 378         struct cnfa cnfa;                       /* compacted NFA, if any */
 379         struct subre *chain;            /* for bookkeeping and error cleanup */
 380 };
 381
 382
 383
 384 /*
 385  * table of function pointers for generic manipulation functions
 386  * A regex_t's re_fns points to one of these.
 387  */
 388 struct fns
 389 {
 390         void            FUNCPTR(free, (regex_t *));
 391 };
 392
 393
 394
 395 /*
 396  * the insides of a regex_t, hidden behind a void *
 397  */
 398 struct guts
 399 {
 400         int                     magic;
 401 #define  GUTSMAGIC       0xfed9
 402         int                     cflags;                 /* copy of compile flags */
 403         long            info;                   /* copy of re_info */
 404         size_t          nsub;                   /* copy of re_nsub */
 405         struct subre *tree;
 406         struct cnfa search;                     /* for fast preliminary search */
 407         int                     ntree;
 408         struct colormap cmap;
 409         int                     FUNCPTR(compare, (const chr *, const chr *, size_t));
 410         struct subre *lacons;           /* lookahead-constraint vector */
 411         int                     nlacons;                /* size of lacons */
 412 };