Added regular expression matching script commands and operators

- The script command pcre_match and the operator ~= will return the
  number of regular expression matches in a given string (roughly
  equivalent to the php function preg_match or the perl operator =~)
- The operator ~! is the opposite of ~= (roughly equivalent to the perl
  operator !~)
- See script_commands and npc/custom/test.txt for more information.

Signed-off-by: Haru <haru@dotalux.com>

2 files changed, 97 insertions, 4 deletions

diff --git a/src/map/script.c b/src/map/script.c
index 17eba6b21..41453e366 100644
--- a/src/map/script.c
+++ b/src/map/script.c
@@ -112,6 +112,10 @@ const char* script_op2name(int op) {
 	RETURN_OP_NAME(C_SUB_POST);
 	RETURN_OP_NAME(C_ADD_PRE);
 	RETURN_OP_NAME(C_SUB_PRE);
+#ifdef PCRE_SUPPORT
+	RETURN_OP_NAME(C_RE_EQ);
+	RETURN_OP_NAME(C_RE_NE);
+#endif // PCRE_SUPPORT
 
 	default:
 		ShowDebug("script_op2name: unexpected op=%d\n", op);
@@ -1224,6 +1228,10 @@ const char* script_parse_subexpr(const char* p,int limit) {
 	|| (op=C_XOR,    opl=4, len=1,*p=='^')              // ^
 	|| (op=C_EQ,     opl=6, len=2,*p=='=' && p[1]=='=') // ==
 	|| (op=C_NE,     opl=6, len=2,*p=='!' && p[1]=='=') // !=
+#ifdef PCRE_SUPPORT
+	|| (op=C_RE_EQ,  opl=6, len=2,*p=='~' && p[1]=='=') // ~=
+	|| (op=C_RE_NE,  opl=6, len=2,*p=='~' && p[1]=='!') // ~!
+#endif // PCRE_SUPPORT
 	|| (op=C_R_SHIFT,opl=8, len=2,*p=='>' && p[1]=='>') // >>
 	|| (op=C_GE,     opl=7, len=2,*p=='>' && p[1]=='=') // >=
 	|| (op=C_GT,     opl=7, len=1,*p=='>')              // >
@@ -3405,6 +3413,8 @@ void op_3(struct script_state* st, int op)
 /// s1 GE s2 -> i
 /// s1 LT s2 -> i
 /// s1 LE s2 -> i
+/// s1 RE_EQ s2 -> i
+/// s1 RE_NE s2 -> i
 /// s1 ADD s2 -> s
 void op_2str(struct script_state* st, int op, const char* s1, const char* s2)
 {
@@ -3417,6 +3427,72 @@ void op_2str(struct script_state* st, int op, const char* s1, const char* s2)
 	case C_GE: a = (strcmp(s1,s2) >= 0); break;
 	case C_LT: a = (strcmp(s1,s2) <  0); break;
 	case C_LE: a = (strcmp(s1,s2) <= 0); break;
+#ifdef PCRE_SUPPORT
+	case C_RE_EQ:
+	case C_RE_NE:
+		{
+			int inputlen = (int)strlen(s1);
+			pcre *compiled_regex;
+			pcre_extra *extra_regex;
+			const char *pcre_error, *pcre_match;
+			int pcre_erroroffset, offsetcount, i;
+			int offsets[256*3]; // (max_capturing_groups+1)*3
+
+			compiled_regex = libpcre->compile(s2, 0, &pcre_error, &pcre_erroroffset, NULL);
+
+			if( compiled_regex == NULL ) {
+				ShowError("script:op2_str: Invalid regex '%s'.\n", s2);
+				script->reportsrc(st);
+				script_pushnil(st);
+				st->state = END;
+				return;
+			}
+
+			extra_regex = libpcre->study(compiled_regex, 0, &pcre_error);
+
+			if( pcre_error != NULL ) {
+				libpcre->free(compiled_regex);
+				ShowError("script:op2_str: Unable to optimize the regex '%s': %s\n", s2, pcre_error);
+				script->reportsrc(st);
+				script_pushnil(st);
+				st->state = END;
+				return;
+			}
+
+			offsetcount = libpcre->exec(compiled_regex, extra_regex, s1, inputlen, 0, 0, offsets, 256*3);
+
+			if( offsetcount == 0 ) {
+				offsetcount = 256;
+			} else if( offsetcount == PCRE_ERROR_NOMATCH ) {
+				offsetcount = 0;
+			} else if( offsetcount < 0 ) {
+				libpcre->free(compiled_regex);
+				if( extra_regex != NULL )
+					libpcre->free(extra_regex);
+				ShowWarning("script:op2_str: Unable to process the regex '%s'.\n", s2);
+				script->reportsrc(st);
+				script_pushnil(st);
+				st->state = END;
+				return;
+			}
+
+			if( op == C_RE_EQ ) {
+				for( i = 0; i < offsetcount; i++ ) {
+					libpcre->get_substring(s1, offsets, offsetcount, i, &pcre_match);
+					mapreg->setregstr(reference_uid(script->add_str("$@regexmatch$"), i), pcre_match);
+					libpcre->free_substring(pcre_match);
+				}
+				mapreg->setreg(script->add_str("$@regexmatchcount"), i);
+				a = offsetcount;
+			} else { // C_RE_NE
+				a = (offsetcount == 0);
+			}
+			libpcre->free(compiled_regex);
+			if( extra_regex != NULL )
+				libpcre->free(extra_regex);
+		}
+		break;
+#endif // PCRE_SUPPORT
 	case C_ADD:
 		{
 			char* buf = (char *)aMalloc((strlen(s1)+strlen(s2)+1)*sizeof(char));
@@ -3988,6 +4064,10 @@ void run_script_main(struct script_state *st) {
 			case C_LOR:
 			case C_R_SHIFT:
 			case C_L_SHIFT:
+#ifdef PCRE_SUPPORT
+			case C_RE_EQ:
+			case C_RE_NE:
+#endif // PCRE_SUPPORT
 				script->op_2(st, c);
 				break;
 
@@ -18500,10 +18580,18 @@ BUILDIN(shopcount) {
 
 // declarations that were supposed to be exported from npc_chat.c
 #ifdef PCRE_SUPPORT
-	BUILDIN(defpattern);
-	BUILDIN(activatepset);
-	BUILDIN(deactivatepset);
-	BUILDIN(deletepset);
+BUILDIN(defpattern);
+BUILDIN(activatepset);
+BUILDIN(deactivatepset);
+BUILDIN(deletepset);
+
+BUILDIN(pcre_match) {
+	const char *input = script_getstr(st, 2);
+	const char *regex = script_getstr(st, 3);
+
+	script->op_2str(st, C_RE_EQ, input, regex);
+	return true;
+}
 #endif
 
 /**
@@ -18867,6 +18955,7 @@ void script_parse_builtin(void) {
 		BUILDIN_DEF(activatepset,"i"), // Activate a pattern set [MouseJstr]
 		BUILDIN_DEF(deactivatepset,"i"), // Deactive a pattern set [MouseJstr]
 		BUILDIN_DEF(deletepset,"i"), // Delete a pattern set [MouseJstr]
+		BUILDIN_DEF(pcre_match,"ss"),
 #endif
 		BUILDIN_DEF(dispbottom,"s"), //added from jA [Lupus]
 		BUILDIN_DEF(getusersname,""),
diff --git a/src/map/script.h b/src/map/script.h
index cf7f22aa9..eed0dbf1d 100644
--- a/src/map/script.h
+++ b/src/map/script.h
@@ -213,6 +213,10 @@ typedef enum c_op {
 	C_SUB_POST, // a--
 	C_ADD_PRE, // ++a
 	C_SUB_PRE, // --a
+#ifdef PCRE_SUPPORT
+	C_RE_EQ, // ~=
+	C_RE_NE, // ~!
+#endif // PCRE_SUPPORT
 } c_op;
 
 enum hQueueOpt {