Comment/#ifdef out the multi-char collating element support code.

We never used or supported it. [Bug 1831425]

5 files changed, 98 insertions, 39 deletions

diff --git a/ChangeLog b/ChangeLog
index 4baa164..f853fd4 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,9 @@
+2007-11-14  Donal K. Fellows  <dkf@users.sf.net>
+
+	* generic/regc*.c: #ifdef/comment out the code that deals with
+	multi-character collating elements, which have never been supported.
+	Cuts the memory consumption of the RE compiler. [Bug 1831425]
+
 2007-11-13  Donal K. Fellows  <dkf@users.sf.net>
 
 	* generic/tclCompCmds.c (TclCompileSwitchCmd, TclCompileRegexpCmd):
diff --git a/generic/regc_cvec.c b/generic/regc_cvec.c
index a0a14c2..afb2f48 100644
--- a/generic/regc_cvec.c
+++ b/generic/regc_cvec.c
@@ -41,12 +41,15 @@
 static struct cvec *
 newcvec(
     int nchrs,			/* to hold this many chrs... */
-    int nranges,		/* ... and this many ranges... */
+    int nranges)		/* ... and this many ranges... */
+#ifdef REGEXP_MCCE_ENABLED
     int nmcces)			/* ... and this many MCCEs */
+#endif
 {
     size_t n, nc;
     struct cvec *cv;
 
+#ifdef REGEXP_MCCE_ENABLED
     nc = (size_t)nchrs + (size_t)nmcces*(MAXMCCE+1) + (size_t)nranges*2;
     n = sizeof(struct cvec) + (size_t)(nmcces-1)*sizeof(chr *)
 	    + nc*sizeof(chr);
@@ -60,6 +63,19 @@ newcvec(
     cv->ranges = cv->chrs + nchrs + nmcces*(MAXMCCE+1);
     cv->rangespace = nranges;
     return clearcvec(cv);
+#else
+    nc = (size_t)nchrs + (size_t)nranges*2;
+    n = sizeof(struct cvec) + nc*sizeof(chr);
+    cv = (struct cvec *) MALLOC(n);
+    if (cv == NULL) {
+	return NULL;
+    }
+    cv->chrspace = nchrs;
+    cv->chrs = (chr *)(((char *)cv)+sizeof(struct cvec));
+    cv->ranges = cv->chrs + nchrs;
+    cv->rangespace = nranges;
+    return clearcvec(cv);
+#endif /*REGEXP_MCCE_ENABLED*/
 }
 
 /*
@@ -71,18 +87,21 @@ static struct cvec *
 clearcvec(
     struct cvec *cv)		/* character vector */
 {
+#ifdef REGEXP_MCCE_ENABLED
     int i;
+#endif
 
     assert(cv != NULL);
     cv->nchrs = 0;
+    cv->nranges = 0;
+#ifdef REGEXP_MCCE_ENABLED
     assert(cv->chrs == (chr *)&cv->mcces[cv->mccespace]);
     cv->nmcces = 0;
     cv->nmccechrs = 0;
-    cv->nranges = 0;
     for (i = 0; i < cv->mccespace; i++) {
 	cv->mcces[i] = NULL;
     }
-
+#endif
     return cv;
 }
 
@@ -158,6 +177,7 @@ addmcce(
  - haschr - does a cvec contain this chr?
  ^ static int haschr(struct cvec *, pchr);
  */
+#ifdef REGEXP_MCCE_ENABLED
 static int			/* predicate */
 haschr(
     struct cvec *cv,		/* character vector */
@@ -178,6 +198,7 @@ haschr(
     }
     return 0;
 }
+#endif
 
 /*
  - getcvec - get a cvec, remembering it as v->cv
@@ -187,18 +208,23 @@ static struct cvec *
 getcvec(
     struct vars *v,		/* context */
     int nchrs,			/* to hold this many chrs... */
-    int nranges,		/* ... and this many ranges... */
+    int nranges)		/* ... and this many ranges... */
+#ifdef REGEXP_MCCE_ENABLED
     int nmcces)			/* ... and this many MCCEs */
+#endif
 {
     if ((v->cv != NULL) && (nchrs <= v->cv->chrspace) &&
-	    (nranges <= v->cv->rangespace) && (nmcces <= v->cv->mccespace)) {
+#ifdef REGEXP_MCCE_ENABLED
+	    (nmcces <= v->cv->mccespace) &&
+#endif
+	    (nranges <= v->cv->rangespace)) {
 	return clearcvec(v->cv);
     }
 
     if (v->cv != NULL) {
 	freecvec(v->cv);
     }
-    v->cv = newcvec(nchrs, nranges, nmcces);
+    v->cv = newcvec(nchrs, nranges/*, nmcces*/);
     if (v->cv == NULL) {
 	ERR(REG_ESPACE);
     }
diff --git a/generic/regc_locale.c b/generic/regc_locale.c
index b08c300..438e821 100644
--- a/generic/regc_locale.c
+++ b/generic/regc_locale.c
@@ -9,7 +9,7 @@
  * See the file "license.terms" for information on usage and redistribution of
  * this file, and for a DISCLAIMER OF ALL WARRANTIES.
  *
- * RCS: @(#) $Id: regc_locale.c,v 1.17 2007/04/19 09:00:55 dkf Exp $
+ * RCS: @(#) $Id: regc_locale.c,v 1.18 2007/11/14 00:07:58 dkf Exp $
  */
 
 /* ASCII character-name table */
@@ -614,6 +614,7 @@ static const chr printCharTable[] = {
  - nmcces - how many distinct MCCEs are there?
  ^ static int nmcces(struct vars *);
  */
+#ifdef REGEXP_MCCE_ENABLED
 static int
 nmcces(
     struct vars *v)		/* context */
@@ -622,23 +623,27 @@ nmcces(
      * No multi-character collating elements defined at the moment.
      */
     return 0;
-}
-
+}
+#endif
+
 /*
  - nleaders - how many chrs can be first chrs of MCCEs?
  ^ static int nleaders(struct vars *);
  */
+#ifdef REGEXP_MCCE_ENABLED
 static int
 nleaders(
     struct vars *v)		/* context */
 {
     return 0;
 }
+#endif
 
 /*
  - allmcces - return a cvec with all the MCCEs of the locale
  ^ static struct cvec *allmcces(struct vars *, struct cvec *);
  */
+#ifdef REGEXP_MCCE_ENABLED
 static struct cvec *
 allmcces(
     struct vars *v,		/* context */
@@ -646,6 +651,7 @@ allmcces(
 {
     return clearcvec(cv);
 }
+#endif
 
 /*
  - element - map collating-element name to celt
@@ -719,7 +725,7 @@ range(
     }
 
     if (!cases) {			/* easy version */
-	cv = getcvec(v, 0, 1, 0);
+	cv = getcvec(v, 0, 1/*, 0*/);
 	NOERRN();
 	addrange(cv, a, b);
 	return cv;
@@ -733,7 +739,7 @@ range(
 
     nchrs = (b - a + 1)*2 + 4;
 
-    cv = getcvec(v, nchrs, 0, 0);
+    cv = getcvec(v, nchrs, 0/*, 0*/);
     NOERRN();
 
     for (c=a; c<=b; c++) {
@@ -792,7 +798,7 @@ eclass(
      */
 
     if ((v->cflags&REG_FAKE) && c == 'x') {
-	cv = getcvec(v, 4, 0, 0);
+	cv = getcvec(v, 4, 0/*, 0*/);
 	addchr(cv, (chr)'x');
 	addchr(cv, (chr)'y');
 	if (cases) {
@@ -809,7 +815,7 @@ eclass(
     if (cases) {
 	return allcases(v, c);
     }
-    cv = getcvec(v, 1, 0, 0);
+    cv = getcvec(v, 1, 0/*, 0*/);
     assert(cv != NULL);
     addchr(cv, (chr)c);
     return cv;
@@ -889,7 +895,7 @@ cclass(
 
     switch((enum classes) index) {
     case CC_PRINT:
-	cv = getcvec(v, NUM_PRINT_CHAR, NUM_PRINT_RANGE, 0);
+	cv = getcvec(v, NUM_PRINT_CHAR, NUM_PRINT_RANGE/*, 0*/);
 	if (cv) {
 	    for (i=0 ; (size_t)i<NUM_PRINT_CHAR ; i++) {
 		addchr(cv, printCharTable[i]);
@@ -901,7 +907,7 @@ cclass(
 	}
 	break;
     case CC_ALNUM:
-	cv = getcvec(v, NUM_ALPHA_CHAR, NUM_DIGIT_RANGE + NUM_ALPHA_RANGE, 0);
+	cv = getcvec(v, NUM_ALPHA_CHAR, NUM_DIGIT_RANGE + NUM_ALPHA_RANGE/*, 0*/);
 	if (cv) {
 	    for (i=0 ; (size_t)i<NUM_ALPHA_CHAR ; i++) {
 		addchr(cv, alphaCharTable[i]);
@@ -917,7 +923,7 @@ cclass(
 	}
 	break;
     case CC_ALPHA:
-	cv = getcvec(v, NUM_ALPHA_CHAR, NUM_ALPHA_RANGE, 0);
+	cv = getcvec(v, NUM_ALPHA_CHAR, NUM_ALPHA_RANGE/*, 0*/);
 	if (cv) {
 	    for (i=0 ; (size_t)i<NUM_ALPHA_RANGE ; i++) {
 		addrange(cv, alphaRangeTable[i].start,
@@ -929,23 +935,23 @@ cclass(
 	}
 	break;
     case CC_ASCII:
-	cv = getcvec(v, 0, 1, 0);
+	cv = getcvec(v, 0, 1/*, 0*/);
 	if (cv) {
 	    addrange(cv, 0, 0x7f);
 	}
 	break;
     case CC_BLANK:
-	cv = getcvec(v, 2, 0, 0);
+	cv = getcvec(v, 2, 0/*, 0*/);
 	addchr(cv, '\t');
 	addchr(cv, ' ');
 	break;
     case CC_CNTRL:
-	cv = getcvec(v, 0, 2, 0);
+	cv = getcvec(v, 0, 2/*, 0*/);
 	addrange(cv, 0x0, 0x1f);
 	addrange(cv, 0x7f, 0x9f);
 	break;
     case CC_DIGIT:
-	cv = getcvec(v, 0, NUM_DIGIT_RANGE, 0);
+	cv = getcvec(v, 0, NUM_DIGIT_RANGE/*, 0*/);
 	if (cv) {
 	    for (i=0 ; (size_t)i<NUM_DIGIT_RANGE ; i++) {
 		addrange(cv, digitRangeTable[i].start,
@@ -954,7 +960,7 @@ cclass(
 	}
 	break;
     case CC_PUNCT:
-	cv = getcvec(v, NUM_PUNCT_CHAR, NUM_PUNCT_RANGE, 0);
+	cv = getcvec(v, NUM_PUNCT_CHAR, NUM_PUNCT_RANGE/*, 0*/);
 	if (cv) {
 	    for (i=0 ; (size_t)i<NUM_PUNCT_RANGE ; i++) {
 		addrange(cv, punctRangeTable[i].start,
@@ -975,7 +981,7 @@ cclass(
 	 * someone comes up with a better arrangement!)
 	 */
 
-	cv = getcvec(v, 0, 3, 0);
+	cv = getcvec(v, 0, 3/*, 0*/);
 	if (cv) {
 	    addrange(cv, '0', '9');
 	    addrange(cv, 'a', 'f');
@@ -983,7 +989,7 @@ cclass(
 	}
 	break;
     case CC_SPACE:
-	cv = getcvec(v, NUM_SPACE_CHAR, NUM_SPACE_RANGE, 0);
+	cv = getcvec(v, NUM_SPACE_CHAR, NUM_SPACE_RANGE/*, 0*/);
 	if (cv) {
 	    for (i=0 ; (size_t)i<NUM_SPACE_RANGE ; i++) {
 		addrange(cv, spaceRangeTable[i].start,
@@ -995,7 +1001,7 @@ cclass(
 	}
 	break;
     case CC_LOWER:
-	cv  = getcvec(v, NUM_LOWER_CHAR, NUM_LOWER_RANGE, 0);
+	cv  = getcvec(v, NUM_LOWER_CHAR, NUM_LOWER_RANGE/*, 0*/);
 	if (cv) {
 	    for (i=0 ; (size_t)i<NUM_LOWER_RANGE ; i++) {
 		addrange(cv, lowerRangeTable[i].start,
@@ -1007,7 +1013,7 @@ cclass(
 	}
 	break;
     case CC_UPPER:
-	cv  = getcvec(v, NUM_UPPER_CHAR, NUM_UPPER_RANGE, 0);
+	cv  = getcvec(v, NUM_UPPER_CHAR, NUM_UPPER_RANGE/*, 0*/);
 	if (cv) {
 	    for (i=0 ; (size_t)i<NUM_UPPER_RANGE ; i++) {
 		addrange(cv, upperRangeTable[i].start,
@@ -1019,7 +1025,7 @@ cclass(
 	}
 	break;
     case CC_GRAPH:
-	cv  = getcvec(v, NUM_GRAPH_CHAR, NUM_GRAPH_RANGE, 0);
+	cv  = getcvec(v, NUM_GRAPH_CHAR, NUM_GRAPH_RANGE/*, 0*/);
 	if (cv) {
 	    for (i=0 ; (size_t)i<NUM_GRAPH_RANGE ; i++) {
 		addrange(cv, graphRangeTable[i].start,
@@ -1057,10 +1063,10 @@ allcases(
     tc = Tcl_UniCharToTitle((chr)c);
 
     if (tc != uc) {
-	cv = getcvec(v, 3, 0, 0);
+	cv = getcvec(v, 3, 0/*, 0*/);
 	addchr(cv, tc);
     } else {
-	cv = getcvec(v, 2, 0, 0);
+	cv = getcvec(v, 2, 0/*, 0*/);
     }
     addchr(cv, lc);
     if (lc != uc) {
diff --git a/generic/regcomp.c b/generic/regcomp.c
index 7faf793..8a43240 100644
--- a/generic/regcomp.c
+++ b/generic/regcomp.c
@@ -53,10 +53,10 @@ static void bracket(struct vars *, struct state *, struct state *);
 static void cbracket(struct vars *, struct state *, struct state *);
 static void brackpart(struct vars *, struct state *, struct state *);
 static const chr *scanplain(struct vars *);
-static void leaders(struct vars *, struct cvec *);
 static void onechr(struct vars *, pchr, struct state *, struct state *);
 static void dovec(struct vars *, struct cvec *, struct state *, struct state *);
 #ifdef REGEXP_MCCE_ENABLED
+static void leaders(struct vars *, struct cvec *);
 static celt nextleader(struct vars *, pchr, pchr);
 #endif
 static void wordchrs(struct vars *);
@@ -171,20 +171,25 @@ static void dumpcnfa(struct cnfa *, FILE *);
 static void dumpcstate(int, struct carc *, struct cnfa *, FILE *);
 #endif
 /* === regc_cvec.c === */
-static struct cvec *newcvec(int, int, int);
 static struct cvec *clearcvec(struct cvec *);
 static void addchr(struct cvec *, pchr);
 static void addrange(struct cvec *, pchr, pchr);
 #ifdef REGEXP_MCCE_ENABLED
+static struct cvec *newcvec(int, int, int);
 static void addmcce(struct cvec *, const chr *, const chr *);
-#endif
-static int haschr(struct cvec *, pchr);
 static struct cvec *getcvec(struct vars *, int, int, int);
+static int haschr(struct cvec *, pchr);
+#else
+static struct cvec *newcvec(int, int);
+static struct cvec *getcvec(struct vars *, int, int);
+#endif
 static void freecvec(struct cvec *);
 /* === regc_locale.c === */
-static int nmcces(struct vars *);
+#ifdef REGEXP_MCCE_ENABLED
 static int nleaders(struct vars *);
+static int nmcces(struct vars *);
 static struct cvec *allmcces(struct vars *, struct cvec *);
+#endif
 static celt element(struct vars *, const chr *, const chr *);
 static struct cvec *range(struct vars *, celt, celt, int);
 static int before(celt, celt);
@@ -223,10 +228,12 @@ struct vars {
     int ntree;			/* number of tree nodes */
     struct cvec *cv;		/* interface cvec */
     struct cvec *cv2;		/* utility cvec */
+#ifdef REGEXP_MCCE_ENABLED
     struct cvec *mcces;		/* collating-element information */
 #define	ISCELEADER(v,c)	(v->mcces != NULL && haschr(v->mcces, (c)))
     struct state *mccepbegin;	/* in nfa, start of MCCE prototypes */
     struct state *mccepend;	/* in nfa, end of MCCE prototypes */
+#endif
     struct subre *lacons;	/* lookahead-constraint vector */
     int nlacons;		/* size of lacons */
 };
@@ -336,7 +343,9 @@ compile(
     v->treefree = NULL;
     v->cv = NULL;
     v->cv2 = NULL;
+#ifdef REGEXP_MCCE_ENABLED
     v->mcces = NULL;
+#endif
     v->lacons = NULL;
     v->nlacons = 0;
     re->re_magic = REMAGIC;
@@ -362,22 +371,22 @@ compile(
     ZAPCNFA(g->search);
     v->nfa = newnfa(v, v->cm, NULL);
     CNOERR();
-    v->cv = newcvec(100, 20, 10);
+    v->cv = newcvec(100, 20);
     if (v->cv == NULL) {
 	return freev(v, REG_ESPACE);
     }
+#ifdef REGEXP_MCCE_ENABLED
     i = nmcces(v);
     if (i > 0) {
-	v->mcces = newcvec(nleaders(v), 0, i);
+	v->mcces = newcvec(nleaders(v), 0);
 	CNOERR();
 	v->mcces = allmcces(v, v->mcces);
 	leaders(v, v->mcces);
-#ifdef REGEXP_MCCE_ENABLED
 	/* Function does nothing with NULL pointers */
 	addmcce(v->mcces, NULL, NULL); /* dummy */
-#endif
     }
     CNOERR();
+#endif
 
     /*
      * Parsing.
@@ -550,9 +559,11 @@ freev(
     if (v->cv2 != NULL) {
 	freecvec(v->cv2);
     }
+#ifdef REGEXP_MCCE_ENABLED
     if (v->mcces != NULL) {
 	freecvec(v->mcces);
     }
+#endif
     if (v->lacons != NULL) {
 	freelacons(v->lacons, v->nlacons);
     }
@@ -1467,6 +1478,7 @@ cbracket(
 {
     struct state *left = newstate(v->nfa);
     struct state *right = newstate(v->nfa);
+#ifdef REGEXP_MCCE_ENABLED
     struct state *s;
     struct arc *a;		/* arc from lp */
     struct arc *ba;		/* arc from left, from bracket() */
@@ -1474,6 +1486,7 @@ cbracket(
     color co;
     const chr *p;
     int i;
+#endif
 
     NOERR();
     bracket(v, left, right);
@@ -1490,13 +1503,14 @@ cbracket(
 
     colorcomplement(v->nfa, v->cm, PLAIN, left, lp, rp);
     NOERR();
-    if (v->mcces == NULL) {	/* no MCCEs -- we're done */
+    if (1 /*v->mcces == NULL*/) {	/* no MCCEs -- we're done */
 	dropstate(v->nfa, left);
 	assert(right->nins == 0);
 	freestate(v->nfa, right);
 	return;
     }
 
+#ifdef REGEXP_MCCE_ENABLED
     /*
      * But complementing gets messy in the presence of MCCEs...
      */
@@ -1546,6 +1560,7 @@ cbracket(
     freestate(v->nfa, left);
     assert(right->nins == 0);
     freestate(v->nfa, right);
+#endif
 }
 
 /*
@@ -1580,7 +1595,7 @@ brackpart(
 	 * Shortcut for ordinary chr (not range, not MCCE leader).
 	 */
 
-	if (!SEE(RANGE) && !ISCELEADER(v, c[0])) {
+	if (!SEE(RANGE) /*&& !ISCELEADER(v, c[0])*/) {
 	    onechr(v, c[0], lp, rp);
 	    return;
 	}
@@ -1696,6 +1711,7 @@ scanplain(
  * certainly necessary, and sets up little disconnected subNFA.
  ^ static void leaders(struct vars *, struct cvec *);
  */
+#ifdef REGEXP_MCCE_ENABLED
 static void
 leaders(
     struct vars *v,
@@ -1731,6 +1747,7 @@ leaders(
 	okcolors(v->nfa, v->cm);
     }
 }
+#endif
 
 /*
  - onechr - fill in arcs for a plain character, and possible case complements
diff --git a/generic/regguts.h b/generic/regguts.h
index 991979e..bc1d7a2 100644
--- a/generic/regguts.h
+++ b/generic/regguts.h
@@ -245,15 +245,19 @@ struct cvec {
     int nranges;		/* number of ranges (chr pairs) */
     int rangespace;		/* number of chrs possible */
     chr *ranges;		/* pointer to vector of chr pairs */
+#ifdef REGEXP_MCCE_ENABLED
     int nmcces;			/* number of MCCEs */
     int mccespace;		/* number of MCCEs possible */
     int nmccechrs;		/* number of chrs used for MCCEs */
     chr *mcces[1];		/* pointers to 0-terminated MCCEs */
 				/* and both batches of chrs are on the end */
+#endif
 };
 
+#ifdef REGEXP_MCCE_ENABLED
 /* caution:  this value cannot be changed easily */
 #define	MAXMCCE	2		/* length of longest MCCE */
+#endif
 
 /*
  * definitions for NFA internal representation