[BZ #544]

Update.
2004-11-18  Jakub Jelinek  <jakub@redhat.com>

	[BZ #544]
	* posix/regex.h (RE_NO_SUB): New define.
	* posix/regex_internal.h (OP_DELETED_SUBEXP): New.
	(re_dfa_t): Add subexp_map.
	* posix/regcomp.c (struct subexp_optimize): New type.
	(optimize_subexps): New routine.
	(re_compile_internal): Call it.
	(re_compile_pattern): Set preg->no_sub to 1 if RE_NO_SUB.
	(free_dfa_content): Free subexp_map.
	(calc_inveclosure, calc_eclosure): Skip OP_DELETED_SUBEXP
	nodes.
	* posix/regexec.c (re_search_internal): If subexp_map
	is not NULL, duplicate registers as needed.
	* posix/Makefile: Add rules to build and run tst-regex2.
	* posix/tst-regex2.c: New test.
	* posix/rxspencer/tests: Fix last two tests (\0 -> \1).
	Add some new tests for nested subexpressions.

ChangeLog

@@ -1,3 +1,23 @@
+2004-11-18  Jakub Jelinek  <jakub@redhat.com>
+
+	[BZ #544]
+	* posix/regex.h (RE_NO_SUB): New define.
+	* posix/regex_internal.h (OP_DELETED_SUBEXP): New.
+	(re_dfa_t): Add subexp_map.
+	* posix/regcomp.c (struct subexp_optimize): New type.
+	(optimize_subexps): New routine.
+	(re_compile_internal): Call it.
+	(re_compile_pattern): Set preg->no_sub to 1 if RE_NO_SUB.
+	(free_dfa_content): Free subexp_map.
+	(calc_inveclosure, calc_eclosure): Skip OP_DELETED_SUBEXP
+	nodes.
+	* posix/regexec.c (re_search_internal): If subexp_map
+	is not NULL, duplicate registers as needed.
+	* posix/Makefile: Add rules to build and run tst-regex2.
+	* posix/tst-regex2.c: New test.
+	* posix/rxspencer/tests: Fix last two tests (\0 -> \1).
+	Add some new tests for nested subexpressions.
+
 2004-11-18  Ulrich Drepper  <drepper@redhat.com>
 
 	* libio/libio.h (_IO_FLAGS2_FORTIFY): Renamed from

posix/Makefile

@@ -80,7 +80,7 @@ tests		:= tstgetopt testfnm runtests runptests	     \
 		   bug-regex13 bug-regex14 bug-regex15 bug-regex16 \
 		   bug-regex17 bug-regex18 bug-regex19 bug-regex20 \
 		   bug-regex21 bug-regex22 bug-regex23 bug-regex24 \
-		   tst-nice tst-nanosleep \
+		   tst-nice tst-nanosleep tst-regex2 \
 		   transbug tst-rxspencer tst-pcre tst-boost \
 		   bug-ga1 tst-vfork1 tst-vfork2 tst-waitid \
 		   tst-getaddrinfo2 bug-glob1 bug-glob2
@@ -160,6 +160,7 @@ tst-fnmatch-ENV = LOCPATH=$(common-objpfx)localedata
 tst-regexloc-ENV = LOCPATH=$(common-objpfx)localedata
 bug-regex1-ENV = LOCPATH=$(common-objpfx)localedata
 tst-regex-ENV = LOCPATH=$(common-objpfx)localedata
+tst-regex2-ENV = LOCPATH=$(common-objpfx)localedata
 bug-regex5-ENV = LOCPATH=$(common-objpfx)localedata
 bug-regex6-ENV = LOCPATH=$(common-objpfx)localedata
 bug-regex17-ENV = LOCPATH=$(common-objpfx)localedata
@@ -244,8 +245,10 @@ $(objpfx)tst-getconf.out: tst-getconf.sh $(objpfx)getconf
 
 ifeq (yes,$(build-shared))
 $(objpfx)tst-regex: $(common-objpfx)rt/librt.so
+$(objpfx)tst-regex2: $(common-objpfx)rt/librt.so
 else
 $(objpfx)tst-regex: $(common-objpfx)rt/librt.a
+$(objpfx)tst-regex2: $(common-objpfx)rt/librt.a
 endif
 
 $(objpfx)bug-ga2-mem: $(objpfx)bug-ga2.out

posix/regcomp.c

@@ -33,6 +33,14 @@ static reg_errcode_t create_initial_state (re_dfa_t *dfa);
 #ifdef RE_ENABLE_I18N
 static void optimize_utf8 (re_dfa_t *dfa);
 #endif
+struct subexp_optimize
+{
+  re_dfa_t *dfa;
+  re_token_t *nodes;
+  int no_sub, re_nsub;
+};
+static bin_tree_t *optimize_subexps (struct subexp_optimize *so,
+                                     bin_tree_t *node, int sidx, int depth);
 static reg_errcode_t analyze (re_dfa_t *dfa);
 static reg_errcode_t analyze_tree (re_dfa_t *dfa, bin_tree_t *node);
 static void calc_first (re_dfa_t *dfa, bin_tree_t *node);
@@ -238,8 +246,8 @@ re_compile_pattern (pattern, length, bufp)
 
   /* And GNU code determines whether or not to get register information
      by passing null for the REGS argument to re_match, etc., not by
-     setting no_sub.  */
-  bufp->no_sub = 0;
+     setting no_sub, unless RE_NO_SUB is set.  */
+  bufp->no_sub = !!(re_syntax_options & RE_NO_SUB);
 
   /* Match anchors at newline.  */
   bufp->newline_anchor = 1;
@@ -633,6 +641,7 @@ free_dfa_content (re_dfa_t *dfa)
   if (dfa->sb_char != utf8_sb_map)
     re_free (dfa->sb_char);
 #endif
+  re_free (dfa->subexp_map);
 #ifdef DEBUG
   re_free (dfa->re_str);
 #endif
@@ -810,6 +819,17 @@ re_compile_internal (preg, pattern, length, syntax)
     optimize_utf8 (dfa);
 #endif
 
+  if (preg->re_nsub > 0)
+    {
+      struct subexp_optimize so;
+
+      so.dfa = dfa;
+      so.nodes = dfa->nodes;
+      so.no_sub = preg->no_sub;
+      so.re_nsub = preg->re_nsub;
+      dfa->str_tree = optimize_subexps (&so, dfa->str_tree, -1, 0);
+    }
+
   /* Analyze the tree and collect information which is necessary to
      create the dfa.  */
   err = analyze (dfa);
@@ -1121,6 +1141,82 @@ optimize_utf8 (dfa)
 }
 #endif
 
+static bin_tree_t *
+optimize_subexps (so, node, sidx, depth)
+     struct subexp_optimize *so;
+     bin_tree_t *node;
+     int sidx, depth;
+{
+  int idx, new_depth, new_sidx;
+  bin_tree_t *ret;
+  if (node == NULL)
+    return NULL;
+
+  new_depth = 0;
+  new_sidx = sidx;
+  if ((depth & 1) && node->type == CONCAT
+      && node->right && node->right->type == 0
+      && so->nodes[idx = node->right->node_idx].type == OP_CLOSE_SUBEXP)
+    {
+      new_depth = depth + 1;
+      if (new_depth == 2
+          || (so->nodes[idx].opr.idx < 8 * sizeof (so->dfa->used_bkref_map)
+              && so->dfa->used_bkref_map & (1 << so->nodes[idx].opr.idx)))
+        new_sidx = so->nodes[idx].opr.idx;
+    }
+  node->left = optimize_subexps (so, node->left, new_sidx, new_depth);
+  new_depth = (depth & 1) == 0 && node->type == CONCAT
+              && node->left && node->left->type == 0
+              && so->nodes[node->left->node_idx].type == OP_OPEN_SUBEXP
+              ? depth + 1 : 0;
+  node->right = optimize_subexps (so, node->right, sidx, new_depth);
+
+  if (node->type != CONCAT)
+    return node;
+  if ((depth & 1) == 0
+      && node->left
+      && node->left->type == 0
+      && so->nodes[idx = node->left->node_idx].type == OP_OPEN_SUBEXP)
+    ret = node->right;
+  else if ((depth & 1)
+           && node->right
+           && node->right->type == 0
+           && so->nodes[idx = node->right->node_idx].type == OP_CLOSE_SUBEXP)
+    ret = node->left;
+  else
+    return node;
+
+  if (so->nodes[idx].opr.idx < 8 * sizeof (so->dfa->used_bkref_map)
+      && so->dfa->used_bkref_map & (1 << so->nodes[idx].opr.idx))
+    return node;
+
+  if (!so->no_sub)
+    {
+      int i;
+
+      if (depth < 2)
+        return node;
+
+      if (so->dfa->subexp_map == NULL)
+        {
+          so->dfa->subexp_map = re_malloc (int, so->re_nsub);
+          if (so->dfa->subexp_map == NULL)
+            return node;
+
+          for (i = 0; i < so->re_nsub; i++)
+            so->dfa->subexp_map[i] = i;
+        }
+
+      i = so->nodes[idx].opr.idx;
+      assert (sidx < i);
+      so->dfa->subexp_map[i] = sidx;
+    }
+
+  so->nodes[idx].type = OP_DELETED_SUBEXP;
+  ret->parent = node->parent;
+  return ret;
+}
+
 /* Analyze the structure tree, and calculate "first", "next", "edest",
    "eclosure", and "inveclosure".  */
 
@@ -1525,6 +1621,8 @@ calc_inveclosure (dfa)
   int src, idx, dest;
   for (src = 0; src < dfa->nodes_len; ++src)
     {
+      if (dfa->nodes[src].type == OP_DELETED_SUBEXP)
+        continue;
       for (idx = 0; idx < dfa->eclosures[src].nelem; ++idx)
 	{
 	  dest = dfa->eclosures[src].elems[idx];
@@ -1560,6 +1658,9 @@ calc_eclosure (dfa)
 #ifdef DEBUG
       assert (dfa->eclosures[node_idx].nelem != -1);
 #endif
+      if (dfa->nodes[node_idx].type == OP_DELETED_SUBEXP)
+        continue;
+
       /* If we have already calculated, skip it.  */
       if (dfa->eclosures[node_idx].nelem != 0)
 	continue;

posix/regex.h

@@ -179,6 +179,10 @@ typedef unsigned long int reg_syntax_t;
    immediately after an alternation or begin-group operator.  */
 #define RE_CONTEXT_INVALID_DUP (RE_CARET_ANCHORS_HERE << 1)
 
+/* If this bit is set, then no_sub will be set to 1 during
+   re_compile_pattern.  */
+#define RE_NO_SUB (RE_CONTEXT_INVALID_DUP << 1)
+
 /* This global variable defines the particular regexp syntax to use (for
    some interfaces).  When a regexp is compiled, the syntax used is
    stored in the pattern buffer, so changing this does not affect

posix/regex_internal.h

@@ -189,6 +189,7 @@ typedef enum
   OP_DUP_PLUS = EPSILON_BIT | 4,
   OP_DUP_QUESTION = EPSILON_BIT | 5,
   ANCHOR = EPSILON_BIT | 6,
+  OP_DELETED_SUBEXP = EPSILON_BIT | 7,
 
   /* Tree type, these are used only by tree. */
   CONCAT = 16,
@@ -644,6 +645,7 @@ struct re_dfa_t
   int mb_cur_max;
   bitset word_char;
   reg_syntax_t syntax;
+  int *subexp_map;
 #ifdef DEBUG
   char* re_str;
 #endif

posix/regexec.c

@@ -882,6 +882,18 @@ re_search_internal (preg, string, length, start, range, stop, nmatch, pmatch,
 	    pmatch[reg_idx].rm_so += match_first;
 	    pmatch[reg_idx].rm_eo += match_first;
 	  }
+
+      if (dfa->subexp_map)
+        for (reg_idx = 0;
+             reg_idx + 1 < nmatch && reg_idx < preg->re_nsub;
+             reg_idx++)
+          if (dfa->subexp_map[reg_idx] != reg_idx)
+            {
+              pmatch[reg_idx + 1].rm_so
+                = pmatch[dfa->subexp_map[reg_idx] + 1].rm_so;
+              pmatch[reg_idx + 1].rm_eo
+                = pmatch[dfa->subexp_map[reg_idx] + 1].rm_eo;
+            }
     }
 
  free_return:

posix/rxspencer/tests

@@ -508,5 +508,21 @@ a*a*a*a*a*a*a*	&	aaaaaa	aaaaaa
 (\b){0}	-	x	@x	-
 \(\b\)\{0,0\}	b	abc	@abc	-
 a(\b){0}c	-	ac	ac	-
-a(.*)b(\0){0}c	-	abc	abc	@bc,-
-a(.*)b(\0){0}c	-	axbc	axbc	x,-
+a(.*)b(\1){0}c	-	abc	abc	@bc,-
+a(.*)b(\1){0}c	-	axbc	axbc	x,-
+
+a\(\(b*\)\)c\1d	b	abbcbbd	abbcbbd	bb,bb
+a\(\([bc]\)\)\2d	b	abcdabbd	abbd	b,b
+a\(\(\(\([bc]\)\)\3\)\)*d	b	abbccd	abbccd	cc,cc,c,c
+a(b)(c)d	-	abcd	abcd	b,c
+a(((b)))c	-	abc	abc	b,b,b
+a(((b|(((c))))))d	-	abd	abd	b,b,b,-,-,-
+a(((b*|c|e)))d	-	abbd	abbd	bb,bb,bb
+a((b|c)){0,0}d	-	ad	ad	-,-
+a((b|c)){0,1}d	-	abd	abd	b,b
+a((b|c)){0,2}d	-	abcd	abcd	c,c
+a((b+|((c)*)))+d	-	abd	abd	b,b,-,-
+a((b+|((c)*)))+d	-	abcd	abcd	c,c,c,c
+(((\b))){0}	-	x	@x	-,-,-
+a(((.*)))b((\2)){0}c	-	abc	abc	@bc,@bc,@bc,-,-
+a(((.*)))b((\1)){0}c	-	axbc	axbc	x,x,x,-,-