regguts.h File Reference

#include "regcustom.h"

Include dependency graph for regguts.h:

This graph shows which files directly or indirectly include this file:

Go to the source code of this file.

Data Structures
struct	colordesc
struct	colormaprange
struct	colormap
struct	cvec
struct	arc
struct	arcbatch
struct	state
struct	statebatch
struct	nfa
struct	carc
struct	cnfa
struct	subre
struct	fns
struct	guts

Macros
#define	DISCARD   void /* for throwing values away */
#define	VS(x)   ((void *)(x)) /* cast something to generic ptr */
#define	_POSIX2_RE_DUP_MAX   255 /* normally from <limits.h> */
#define	NOTREACHED   0
#define	DUPMAX   _POSIX2_RE_DUP_MAX
#define	DUPINF   (DUPMAX+1)
#define	REMAGIC   0xfed7 /* magic number for main struct */
#define	LATYPE_AHEAD_POS   03 /* positive lookahead */
#define	LATYPE_AHEAD_NEG   02 /* negative lookahead */
#define	LATYPE_BEHIND_POS   01 /* positive lookbehind */
#define	LATYPE_BEHIND_NEG   00 /* negative lookbehind */
#define	LATYPE_IS_POS(la)   ((la) & 01)
#define	LATYPE_IS_AHEAD(la)   ((la) & 02)
#define	FDEBUG(arglist)   {}
#define	MDEBUG(arglist)   {}
#define	UBITS   (CHAR_BIT * sizeof(unsigned))
#define	BSET(uv, sn)   ((uv)[(sn)/UBITS] |= (unsigned)1 << ((sn)%UBITS))
#define	ISBSET(uv, sn)   ((uv)[(sn)/UBITS] & ((unsigned)1 << ((sn)%UBITS)))
#define	NUM_CCLASSES   14
#define	MAX_COLOR   32767 /* max color (must fit in 'color' datatype) */
#define	COLORLESS   (-1) /* impossible color */
#define	RAINBOW   (-2) /* represents all colors except pseudocolors */
#define	WHITE   0 /* default color, parent of all others */
#define	NOSUB   COLORLESS /* value of "sub" when no open subcolor */
#define	FREECOL   01 /* currently free */
#define	PSEUDO   02 /* pseudocolor, no real chars */
#define	COLMARK   04 /* temporary marker used in some functions */
#define	UNUSEDCOLOR(cd)   ((cd)->flags & FREECOL)
#define	CMMAGIC   0x876
#define	CDEND(cm)   (&(cm)->cd[(cm)->max + 1])
#define	NINLINECDS   ((size_t) 10)
#define	GETCOLOR(cm, c)   ((c) <= MAX_SIMPLE_CHR ? (cm)->locolormap[(c) - CHR_MIN] : pg_reg_getcolor(cm, c))
#define	freechain   outchain /* we do not maintain "freechainRev" */
#define	ARCBATCHSIZE(n)   ((n) * sizeof(struct arc) + offsetof(struct arcbatch, a))
#define	FIRSTABSIZE   64
#define	MAXABSIZE   1024
#define	FREESTATE   (-1)
#define	STATEBATCHSIZE(n)   ((n) * sizeof(struct state) + offsetof(struct statebatch, s))
#define	FIRSTSBSIZE   32
#define	MAXSBSIZE   1024
#define	HASLACONS   01 /* uses lookaround constraints */
#define	MATCHALL   02 /* matches all strings of a range of lengths */
#define	CNFA_NOPROGRESS   01 /* flag bit for a no-progress state */
#define	ZAPCNFA(cnfa)   ((cnfa).nstates = 0)
#define	NULLCNFA(cnfa)   ((cnfa).nstates == 0)
#define	REG_MAX_COMPILE_SPACE   (500000 * (sizeof(struct state) + 4 * sizeof(struct arc)))
#define	LONGER   01 /* prefers longer match */
#define	SHORTER   02 /* prefers shorter match */
#define	MIXED   04 /* mixed preference below */
#define	CAP   010 /* capturing parens here or below */
#define	BACKR   020 /* back reference here or below */
#define	INUSE   0100 /* in use in final tree */
#define	NOPROP   03 /* bits which may not propagate up */
#define	LMIX(f)   ((f)<<2) /* LONGER -> MIXED */
#define	SMIX(f)   ((f)<<1) /* SHORTER -> MIXED */
#define	UP(f)   (((f)&~NOPROP) | (LMIX(f) & SMIX(f) & MIXED))
#define	MESSY(f)   ((f)&(MIXED|CAP|BACKR))
#define	PREF(f)   ((f)&NOPROP)
#define	PREF2(f1, f2)   ((PREF(f1) != 0) ? PREF(f1) : PREF(f2))
#define	COMBINE(f1, f2)   (UP((f1)|(f2)) | PREF2(f1, f2))
#define	CANCEL_REQUESTED(re)   ((*((struct fns *) (re)->re_fns)->cancel_requested) ())
#define	STACK_TOO_DEEP(re)   ((*((struct fns *) (re)->re_fns)->stack_too_deep) ())
#define	GUTSMAGIC   0xfed9

Typedefs
typedef short	color
typedef struct colormaprange	colormaprange

Enumerations
enum	char_classes {   CC_ALNUM, CC_ALPHA, CC_ASCII, CC_BLANK,   CC_CNTRL, CC_DIGIT, CC_GRAPH, CC_LOWER,   CC_PRINT, CC_PUNCT, CC_SPACE, CC_UPPER,   CC_XDIGIT, CC_WORD }

Functions
void	pg_set_regex_collation (Oid collation)
color	pg_reg_getcolor (struct colormap *cm, chr c)

Macro Definition Documentation

_POSIX2_RE_DUP_MAX

#define _POSIX2_RE_DUP_MAX   255 /* normally from <limits.h> */

Definition at line 82 of file regguts.h.

ARCBATCHSIZE

#define ARCBATCHSIZE

(

n

)

((n) * sizeof(struct arc) + offsetof(struct arcbatch, a))

Definition at line 312 of file regguts.h.

Referenced by allocarc(), and freenfa().

BACKR

#define BACKR   020 /* back reference here or below */

Definition at line 479 of file regguts.h.

Referenced by cdissect(), parseqatom(), and rstacktoodeep().

BSET

#define BSET	(		uv,
			sn
	)		((uv)[(sn)/UBITS] |= (unsigned)1 << ((sn)%UBITS))

Definition at line 126 of file regguts.h.

Referenced by initialize(), and miss().

CANCEL_REQUESTED

#define CANCEL_REQUESTED

(

re

)

((*((struct fns *) (re)->re_fns)->cancel_requested) ())

Definition at line 516 of file regguts.h.

Referenced by cdissect(), checkmatchall_recurse(), copyins(), copyouts(), mergeins(), miss(), moveins(), moveouts(), newarc(), newstate(), and range().

CAP

#define CAP   010 /* capturing parens here or below */

Definition at line 478 of file regguts.h.

Referenced by parseqatom(), and rstacktoodeep().

CDEND

#define CDEND

(

cm

)

(&(cm)->cd[(cm)->max + 1])

Definition at line 232 of file regguts.h.

Referenced by colorcomplement(), okcolors(), optimizebracket(), and rainbow().

CMMAGIC

#define CMMAGIC   0x876

Definition at line 226 of file regguts.h.

Referenced by initcm().

CNFA_NOPROGRESS

#define CNFA_NOPROGRESS   01 /* flag bit for a no-progress state */

Definition at line 413 of file regguts.h.

Referenced by compact(), dumpnfa(), and miss().

COLMARK

#define COLMARK   04 /* temporary marker used in some functions */

Definition at line 182 of file regguts.h.

Referenced by colorcomplement(), and optimizebracket().

COLORLESS

#define COLORLESS   (-1) /* impossible color */

Definition at line 153 of file regguts.h.

Referenced by compact(), dumpnfa(), findprefix(), makesearch(), maxcolor(), miss(), newcolor(), newnfa(), newsub(), onechr(), parseqatom(), pg_regcomp(), pseudocolor(), rainbow(), specialcolors(), subcolor(), subcolorcvec(), subcolorhi(), and traverse_lacons().

COMBINE

#define COMBINE	(		f1,
			f2
	)		(UP((f1)|(f2)) | PREF2(f1, f2))

Definition at line 488 of file regguts.h.

Referenced by parseqatom().

DISCARD

#define DISCARD   void /* for throwing values away */

Definition at line 58 of file regguts.h.

Referenced by nfatree(), and pg_regcomp().

DUPINF

#define DUPINF   (DUPMAX+1)

Definition at line 94 of file regguts.h.

Referenced by cbrdissect(), checkmatchall(), checkmatchall_recurse(), citerdissect(), creviterdissect(), dfa_backref(), longest(), parseqatom(), rstacktoodeep(), and shortest().

DUPMAX

#define DUPMAX   _POSIX2_RE_DUP_MAX

Definition at line 93 of file regguts.h.

Referenced by scannum().

FDEBUG

#define FDEBUG

(

arglist

)

{}

Definition at line 116 of file regguts.h.

Referenced by getvacant(), lacon(), longest(), matchuntil(), miss(), pickss(), and shortest().

FIRSTABSIZE

#define FIRSTABSIZE   64

Definition at line 314 of file regguts.h.

Referenced by allocarc().

FIRSTSBSIZE

#define FIRSTSBSIZE   32

Definition at line 340 of file regguts.h.

Referenced by newstate().

freechain

#define freechain   outchain /* we do not maintain "freechainRev" */

Definition at line 298 of file regguts.h.

FREECOL

#define FREECOL   01 /* currently free */

Definition at line 180 of file regguts.h.

Referenced by freecolor().

FREESTATE

#define FREESTATE   (-1)

Definition at line 320 of file regguts.h.

Referenced by delsub(), deltraverse(), and freestate().

GETCOLOR

#define GETCOLOR	(		cm,
			c
	)		((c) <= MAX_SIMPLE_CHR ? (cm)->locolormap[(c) - CHR_MIN] : pg_reg_getcolor(cm, c))

Definition at line 252 of file regguts.h.

Referenced by colorcomplement(), findprefix(), longest(), matchuntil(), and shortest().

GUTSMAGIC

#define GUTSMAGIC   0xfed9

Definition at line 529 of file regguts.h.

Referenced by pg_regcomp(), and rstacktoodeep().

HASLACONS

#define HASLACONS   01 /* uses lookaround constraints */

Definition at line 406 of file regguts.h.

Referenced by compact(), dumpnfa(), and miss().

INUSE

#define INUSE   0100 /* in use in final tree */

Definition at line 480 of file regguts.h.

Referenced by cleanst(), markst(), and rstacktoodeep().

ISBSET

#define ISBSET	(		uv,
			sn
	)		((uv)[(sn)/UBITS] & ((unsigned)1 << ((sn)%UBITS)))

Definition at line 127 of file regguts.h.

Referenced by miss().

LATYPE_AHEAD_NEG

#define LATYPE_AHEAD_NEG   02 /* negative lookahead */

Definition at line 100 of file regguts.h.

Referenced by next(), processlacon(), and rstacktoodeep().

LATYPE_AHEAD_POS

#define LATYPE_AHEAD_POS   03 /* positive lookahead */

Definition at line 99 of file regguts.h.

Referenced by next(), processlacon(), and rstacktoodeep().

LATYPE_BEHIND_NEG

#define LATYPE_BEHIND_NEG   00 /* negative lookbehind */

Definition at line 102 of file regguts.h.

Referenced by next(), processlacon(), and rstacktoodeep().

LATYPE_BEHIND_POS

#define LATYPE_BEHIND_POS   01 /* positive lookbehind */

Definition at line 101 of file regguts.h.

Referenced by next(), processlacon(), and rstacktoodeep().

LATYPE_IS_AHEAD

#define LATYPE_IS_AHEAD

(

la

)

((la) & 02)

Definition at line 104 of file regguts.h.

Referenced by lacon(), and pg_regcomp().

LATYPE_IS_POS

#define LATYPE_IS_POS

(

la

)

((la) & 01)

Definition at line 103 of file regguts.h.

Referenced by lacon().

LMIX

#define LMIX

(

f

)

((f)<<2) /* LONGER -> MIXED */

Definition at line 482 of file regguts.h.

LONGER

#define LONGER   01 /* prefers longer match */

Definition at line 475 of file regguts.h.

Referenced by parse(), parseqatom(), and rstacktoodeep().

MATCHALL

#define MATCHALL   02 /* matches all strings of a range of lengths */

Definition at line 407 of file regguts.h.

Referenced by checkmatchall(), dumpnfa(), longest(), matchuntil(), pg_regprefix(), and shortest().

MAX_COLOR

#define MAX_COLOR   32767 /* max color (must fit in 'color' datatype) */

Definition at line 152 of file regguts.h.

Referenced by newcolor().

MAXABSIZE

#define MAXABSIZE   1024

Definition at line 315 of file regguts.h.

Referenced by allocarc().

MAXSBSIZE

#define MAXSBSIZE   1024

Definition at line 341 of file regguts.h.

Referenced by newstate().

MDEBUG

#define MDEBUG

(

arglist

)

{}

Definition at line 117 of file regguts.h.

Referenced by caltdissect(), cbrdissect(), ccondissect(), cdissect(), cfindloop(), citerdissect(), crevcondissect(), creviterdissect(), find(), and subset().

MESSY

#define MESSY

(

f

)

((f)&(MIXED|CAP|BACKR))

Definition at line 485 of file regguts.h.

Referenced by parse(), and parseqatom().

MIXED

#define MIXED   04 /* mixed preference below */

Definition at line 477 of file regguts.h.

Referenced by parseqatom(), and rstacktoodeep().

NINLINECDS

#define NINLINECDS   ((size_t) 10)

Definition at line 247 of file regguts.h.

Referenced by initcm().

NOPROP

#define NOPROP   03 /* bits which may not propagate up */

Definition at line 481 of file regguts.h.

NOSUB

#define NOSUB   COLORLESS /* value of "sub" when no open subcolor */

Definition at line 176 of file regguts.h.

Referenced by freecolor(), initcm(), newcolor(), newsub(), okcolors(), and pseudocolor().

NOTREACHED

#define NOTREACHED   0

Definition at line 91 of file regguts.h.

Referenced by brenext(), combine(), copyins(), copyouts(), lexescape(), mergeins(), moveins(), moveouts(), next(), processlacon(), pull(), and push().

NULLCNFA

#define NULLCNFA

(

cnfa

)

((cnfa).nstates == 0)

Definition at line 431 of file regguts.h.

Referenced by freecnfa(), freelacons(), freesrnode(), rfree(), and rstacktoodeep().

NUM_CCLASSES

#define NUM_CCLASSES   14

Definition at line 139 of file regguts.h.

Referenced by bracket().

PREF

#define PREF

(

f

)

((f)&NOPROP)

Definition at line 486 of file regguts.h.

Referenced by parseqatom().

PREF2

#define PREF2	(		f1,
			f2
	)		((PREF(f1) != 0) ? PREF(f1) : PREF(f2))

Definition at line 487 of file regguts.h.

PSEUDO

#define PSEUDO   02 /* pseudocolor, no real chars */

Definition at line 181 of file regguts.h.

Referenced by checkmatchall(), colorcomplement(), combine(), miss(), optimizebracket(), pg_reg_getcharacters(), pg_reg_getnumcharacters(), pseudocolor(), and rainbow().

RAINBOW

#define RAINBOW   (-2) /* represents all colors except pseudocolors */

Definition at line 154 of file regguts.h.

Referenced by checkmatchall(), checkmatchall_recurse(), colorcomplement(), combine(), dumpnfa(), findprefix(), miss(), optimizebracket(), and rainbow().

REG_MAX_COMPILE_SPACE

#define REG_MAX_COMPILE_SPACE   (500000 * (sizeof(struct state) + 4 * sizeof(struct arc)))

Definition at line 444 of file regguts.h.

Referenced by allocarc(), and newstate().

REMAGIC

#define REMAGIC   0xfed7 /* magic number for main struct */

Definition at line 96 of file regguts.h.

Referenced by pg_reg_colorisbegin(), pg_reg_colorisend(), pg_reg_getcharacters(), pg_reg_getfinalstate(), pg_reg_getinitialstate(), pg_reg_getnumcharacters(), pg_reg_getnumcolors(), pg_reg_getnumoutarcs(), pg_reg_getnumstates(), pg_reg_getoutarcs(), pg_regcomp(), pg_regexec(), pg_regprefix(), rfree(), and rstacktoodeep().

SHORTER

#define SHORTER   02 /* prefers shorter match */

Definition at line 476 of file regguts.h.

Referenced by ccondissect(), cdissect(), cfindloop(), citerdissect(), crevcondissect(), creviterdissect(), find(), parseqatom(), pg_regcomp(), and rstacktoodeep().

SMIX

#define SMIX

(

f

)

((f)<<1) /* SHORTER -> MIXED */

Definition at line 483 of file regguts.h.

STACK_TOO_DEEP

#define STACK_TOO_DEEP

(

re

)

((*((struct fns *) (re)->re_fns)->stack_too_deep) ())

Definition at line 519 of file regguts.h.

Referenced by cdissect(), checkmatchall_recurse(), cleartraverse(), clonesuccessorstates(), deltraverse(), duptraverse(), emptyreachable(), findconstraintloop(), lacon(), markcanreach(), markreachable(), removetraverse(), and subre().

STATEBATCHSIZE

#define STATEBATCHSIZE

(

n

)

((n) * sizeof(struct state) + offsetof(struct statebatch, s))

Definition at line 338 of file regguts.h.

Referenced by freenfa(), and newstate().

UBITS

#define UBITS   (CHAR_BIT * sizeof(unsigned))

Definition at line 125 of file regguts.h.

Referenced by newdfa().

UNUSEDCOLOR

#define UNUSEDCOLOR

(

cd

)

((cd)->flags & FREECOL)

Definition at line 185 of file regguts.h.

Referenced by colorcomplement(), freecolor(), newcolor(), okcolors(), optimizebracket(), and rainbow().

UP

#define UP

(

f

)

(((f)&~NOPROP) | (LMIX(f) & SMIX(f) & MIXED))

Definition at line 484 of file regguts.h.

Referenced by parse(), and parseqatom().

VS

#define VS

(

x

)

((void *)(x)) /* cast something to generic ptr */

Definition at line 61 of file regguts.h.

Referenced by cmp(), moresubs(), newcolor(), pg_regcomp(), and pg_regexec().

WHITE

#define WHITE   0 /* default color, parent of all others */

Definition at line 155 of file regguts.h.

Referenced by freecolor(), initcm(), and pickss().

ZAPCNFA

#define ZAPCNFA

(

cnfa

)

((cnfa).nstates = 0)

Definition at line 429 of file regguts.h.

Referenced by freecnfa(), newlacon(), pg_regcomp(), and subre().

Typedef Documentation

color

typedef short color

Definition at line 150 of file regguts.h.

colormaprange

typedef struct colormaprange colormaprange

Enumeration Type Documentation

char_classes

enum char_classes

Enumerator
CC_ALNUM
CC_ALPHA
CC_ASCII
CC_BLANK
CC_CNTRL
CC_DIGIT
CC_GRAPH
CC_LOWER
CC_PRINT
CC_PUNCT
CC_SPACE
CC_UPPER
CC_XDIGIT
CC_WORD

Definition at line 133 of file regguts.h.

{
    CC_ALNUM, CC_ALPHA, CC_ASCII, CC_BLANK, CC_CNTRL, CC_DIGIT, CC_GRAPH,
    CC_LOWER, CC_PRINT, CC_PUNCT, CC_SPACE, CC_UPPER, CC_XDIGIT, CC_WORD
};

Function Documentation

pg_reg_getcolor()

color pg_reg_getcolor	(	struct colormap *	cm,
		chr	c
	)

Definition at line 120 of file regc_color.c.

References assert, cclass_column_index(), colormap::cmranges, colormap::hiarraycols, colormap::hicolormap, MAX_SIMPLE_CHR, colormap::numcmranges, and colormaprange::rownum.

{
    int         rownum,
                colnum,
                low,
                high;

    /* Should not be used for chrs in the locolormap */
    assert(c > MAX_SIMPLE_CHR);

    /*
     * Find which row it's in.  The colormapranges are in order, so we can use
     * binary search.
     */
    rownum = 0;                 /* if no match, use array row zero */
    low = 0;
    high = cm->numcmranges;
    while (low < high)
    {
        int         middle = low + (high - low) / 2;
        const colormaprange *cmr = &cm->cmranges[middle];

        if (c < cmr->cmin)
            high = middle;
        else if (c > cmr->cmax)
            low = middle + 1;
        else
        {
            rownum = cmr->rownum;   /* found a match */
            break;
        }
    }

    /*
     * Find which column it's in --- this is all locale-dependent.
     */
    if (cm->hiarraycols > 1)
    {
        colnum = cclass_column_index(cm, c);
        return cm->hicolormap[rownum * cm->hiarraycols + colnum];
    }
    else
    {
        /* fast path if no relevant cclasses */
        return cm->hicolormap[rownum];
    }
}

pg_set_regex_collation()

void pg_set_regex_collation

(

Oid

collation

)

Definition at line 232 of file regc_pg_locale.c.

References pg_locale_struct::deterministic, ereport, errcode(), errhint(), errmsg(), ERROR, GetDatabaseEncoding(), lc_ctype_is_c(), OidIsValid, pg_newlocale_from_collation(), pg_regex_collation, PG_REGEX_LOCALE_1BYTE, PG_REGEX_LOCALE_1BYTE_L, PG_REGEX_LOCALE_C, PG_REGEX_LOCALE_ICU, PG_REGEX_LOCALE_WIDE, PG_REGEX_LOCALE_WIDE_L, pg_regex_strategy, PG_UTF8, and pg_locale_struct::provider.

Referenced by pg_regcomp(), pg_regexec(), and pg_regprefix().

{
    if (lc_ctype_is_c(collation))
    {
        /* C/POSIX collations use this path regardless of database encoding */
        pg_regex_strategy = PG_REGEX_LOCALE_C;
        pg_regex_locale = 0;
        pg_regex_collation = C_COLLATION_OID;
    }
    else
    {
        if (collation == DEFAULT_COLLATION_OID)
            pg_regex_locale = 0;
        else if (OidIsValid(collation))
        {
            /*
             * NB: pg_newlocale_from_collation will fail if not HAVE_LOCALE_T;
             * the case of pg_regex_locale != 0 but not HAVE_LOCALE_T does not
             * have to be considered below.
             */
            pg_regex_locale = pg_newlocale_from_collation(collation);
        }
        else
        {
            /*
             * This typically means that the parser could not resolve a
             * conflict of implicit collations, so report it that way.
             */
            ereport(ERROR,
                    (errcode(ERRCODE_INDETERMINATE_COLLATION),
                     errmsg("could not determine which collation to use for regular expression"),
                     errhint("Use the COLLATE clause to set the collation explicitly.")));
        }

        if (pg_regex_locale && !pg_regex_locale->deterministic)
            ereport(ERROR,
                    (errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
                     errmsg("nondeterministic collations are not supported for regular expressions")));

#ifdef USE_ICU
        if (pg_regex_locale && pg_regex_locale->provider == COLLPROVIDER_ICU)
            pg_regex_strategy = PG_REGEX_LOCALE_ICU;
        else
#endif
        if (GetDatabaseEncoding() == PG_UTF8)
        {
            if (pg_regex_locale)
                pg_regex_strategy = PG_REGEX_LOCALE_WIDE_L;
            else
                pg_regex_strategy = PG_REGEX_LOCALE_WIDE;
        }
        else
        {
            if (pg_regex_locale)
                pg_regex_strategy = PG_REGEX_LOCALE_1BYTE_L;
            else
                pg_regex_strategy = PG_REGEX_LOCALE_1BYTE;
        }

        pg_regex_collation = collation;
    }
}