fix some more edge cases

pike.c

@@ -437,7 +437,7 @@ int re_comp(rcode *prog, const char *re, int nsubs)
 if (freesub) \
 	{ s1 = freesub; freesub = (rsub*)s1->sub[0]; copy } \
 else \
-	{ s1 = (rsub*)&nsubs[rsubsize * subidx++]; init } \
+	{ s1 = (rsub*)&nsubs[suboff+=rsubsize]; init } \
 
 #define decref(csub) \
 if (--csub->ref == 0) { \
@@ -537,14 +537,14 @@ for (j = 0; j < nsubp; j++) s1->sub[j] = nsub->sub[j]; \
 
 int re_pikevm(rcode *prog, const char *s, const char **subp, int nsubp)
 {
-	int i, j, c, gen, subidx = 1, *npc;
 	int rsubsize = sizeof(rsub)+(sizeof(char*)*nsubp);
-	int clistidx = 0, nlistidx = 0, pclistidx;
+	int i, j, c, gen, suboff = rsubsize, *npc;
+	int clistidx = 0, nlistidx = 0;
 	const char *sp = s, *_sp = s;
 	int *insts = prog->insts;
 	int *pcs[prog->splits];
 	rsub *subs[prog->splits];
-	char nsubs[rsubsize * 512];
+	char nsubs[500000];
 	rsub *nsub, *s1, *matched = NULL, *freesub = NULL;
 	rthread _clist[prog->len], _nlist[prog->len];
 	rthread *clist = _clist, *nlist = _nlist, *tmp;
@@ -571,7 +571,7 @@ int re_pikevm(rcode *prog, const char *s, const char **subp, int nsubp)
 			case MATCH:
 				if (matched) {
 					decref(matched)
-					subidx = 0;
+					suboff = 0;
 				}
 				matched = nsub;
 				goto break_for;
@@ -586,19 +586,16 @@ int re_pikevm(rcode *prog, const char *s, const char **subp, int nsubp)
 		nlist = tmp;
 		clistidx = nlistidx;
 		nlistidx = 0;
-		if (clistidx != 1 && !matched) {
-			if (!clistidx && pclistidx)
-				_sp = sp;
+		if (!matched) {
 			jmp_start:
-			pclistidx = nlistidx;
 			newsub(for (i = 1; i < nsubp; i++) s1->sub[i] = NULL;, /*nop*/)
 			s1->ref = 1;
 			s1->sub[0] = _sp;
-			npc = insts; nsub = s1;
+			nsub = s1;
+			npc = insts;
 			addthread(1, clist, clistidx)
 		} else if (!clistidx)
 			break;
-		pclistidx = clistidx;
 	}
 	if (matched) {
 		for (i = 0, j = i; i < nsubp; i+=2, j++) {

test.sh

@@ -115,6 +115,12 @@ qwerty.*$
 a(?:b|c|d){4,5}(.)
 (?:c|d)(?:)(?:a(?:)(?:b)(?:b(?:))(?:b(?:)(?:b)))
 \\\\<abc
+a\\\\([COM]+\\\\)|([A-Za-z_%.]+):
+xyz\\\\([COM]+\\\\)|(abcd[A-Za-z_%.]+):cd
+aaaaa(aa)aa(aa(a)a)?aa
+^\\\\.+(((/)\\\\.\\\\.+)+)?
+^(http|https|ftp):[/]{2}([a-zA-Z0-9\\\\\\\\.]+\\\\.[a-zA-Z]{2,4})(:[0-9]+)?/?([a-zA-Z0-9\\\\\\\\._?,'/+&amp;%$#=~]*)
+(\\\\$\\\\([a-zA-Z0-9_]+\\\\))|(([A-Za-z_%.]+):)
 "
 input="\
 abcdef
@@ -231,6 +237,12 @@ abcd
 acdbcdbe
 cabbbb
 dabc
+a(COM:=.o):
+xyz(abcdCOM:cd):
+aaaaaaaaaaaaaaaaaaaaaaaaa
+../../..
+https://kyryl.tk/404
+OBJ = \$(SRC:.c=.o)
 "
 expect="\
 (0,3)
@@ -347,6 +359,12 @@ expect="\
 (0,7)(6,7)
 (0,6)
 -nomatch-
+(2,6)(2,5)
+(4,14)(4,11)
+(0,15)(5,7)(9,13)(11,12)
+(0,8)(2,8)(5,8)(5,6)
+(0,20)(0,5)(8,16)(?,?)(17,20)
+(8,12)(?,?)(8,12)(8,11)
 (0,0)
 "