Add UPGRADING for mb_ereg changes

Also some minor code cleanup.

Author: nikic

UPGRADING

@@ -99,6 +99,12 @@ BCMath:
   . bcmul() and bcpow() now return numbers with the requested scale. Formerly,
     the returned numbers may have omitted trailing decimal zeroes.
 
+MBString:
+  . Due to added support for named captures, mb_ereg_*() patterns using named
+    captures will behave differently. In particular named captures will be part
+    of matches and mb_ereg_replace() will interpret additional syntax. See
+    "New Features" section for more information.
+
 mysqli:
   . Prepared statements now properly report the fractional seconds for DATETIME/
     TIME/TIMESTAMP columns with decimals specifier (e.g. TIMESTAMP(6) when using
@@ -192,6 +198,21 @@ MBString:
   . Mbstring now correctly supports strings larger than 2GB.
   . Performance of the mbstring extension has been significantly improved
     across the board. The largest improvements are in case conversion functions.
+  . mb_ereg_*() functions now support named captures. Matching functions like
+    mb_ereg() will now return named captures both using their group number and
+    their name, similar to PCRE:
+
+        mb_ereg('(?<word>\w+)', '国', $matches);
+        // => [0 => "国", 1 => "国", "word" => "国"];
+
+    Additionally, mb_ereg_replace() now supports the \k<> and \k'' notations
+    to reference named captures in the replacement string:
+
+        mb_ereg_replace('\s*(?<word>\w+)\s*', "_\k<word>_\k'word'_", ' foo ');
+        // => "_foo_foo_"
+
+    \k<> and \k'' can also be used for numbered references, which also works
+    with group numbers greater than 9.
 
 readline:
   . Support for the completion_append_character and completion_suppress_append

ext/mbstring/php_mbregex.c

@@ -655,7 +655,7 @@ _php_mb_regex_init_options(const char *parg, size_t narg, OnigOptionType *option
 typedef struct mb_regex_groups_iter_args {
 	zval		*groups;
 	char		*search_str;
-	int			search_len;
+	size_t		search_len;
 	OnigRegion	*region;
 } mb_regex_groups_iter_args;
 /* }}} */
@@ -665,25 +665,19 @@ static int
 mb_regex_groups_iter(const OnigUChar* name, const OnigUChar* name_end, int ngroup_num, int* group_nums, regex_t* reg, void* parg)
 {
 	mb_regex_groups_iter_args *args = (mb_regex_groups_iter_args *) parg;
-	int i, gn, ref, beg, end;
-
-	for (i = 0; i < ngroup_num; i++) {
-		gn = group_nums[i];
-		ref = onig_name_to_backref_number(reg, name, name_end, args->region);
-		if (ref != gn) {
-			/*
-			 * In case of duplicate groups, keep only the last suceeding one
-			 * to be consistent with preg_match with the PCRE_DUPNAMES option.
-			 */
-			continue;
-		}
-		beg = args->region->beg[gn];
-		end = args->region->end[gn];
-		if (beg >= 0 && beg < end && end <= args->search_len) {
-			add_assoc_stringl_ex(args->groups, (char *)name, name_end - name, &args->search_str[beg], end - beg);
-		} else {
-			add_assoc_bool_ex(args->groups, (char *)name, name_end - name, 0);
-		}
+	int gn, beg, end;
+
+	/*
+	 * In case of duplicate groups, keep only the last suceeding one
+	 * to be consistent with preg_match with the PCRE_DUPNAMES option.
+	 */
+	gn = onig_name_to_backref_number(reg, name, name_end, args->region);
+	beg = args->region->beg[gn];
+	end = args->region->end[gn];
+	if (beg >= 0 && beg < end && end <= args->search_len) {
+		add_assoc_stringl_ex(args->groups, (char *)name, name_end - name, &args->search_str[beg], end - beg);
+	} else {
+		add_assoc_bool_ex(args->groups, (char *)name, name_end - name, 0);
 	}
 
 	return 0;