[Bug #20504] Move string literal concatenation to iseq compiler

compile.c

@@ -4506,42 +4506,46 @@ static int
 compile_dstr_fragments(rb_iseq_t *iseq, LINK_ANCHOR *const ret, const NODE *const node, int *cntp)
 {
     const struct RNode_LIST *list = RNODE_DSTR(node)->nd_next;
-    VALUE lit = rb_node_dstr_string_val(node);
-    LINK_ELEMENT *first_lit = 0;
+    rb_parser_string_t *str = RNODE_DSTR(node)->string;
+    VALUE lit = 0;
+    const NODE *lit_node = node;
     int cnt = 0;
 
-    debugp_param("nd_lit", lit);
-    if (!NIL_P(lit)) {
-        cnt++;
-        if (!RB_TYPE_P(lit, T_STRING)) {
-            COMPILE_ERROR(ERROR_ARGS "dstr: must be string: %s",
-                          rb_builtin_type_name(TYPE(lit)));
-            return COMPILE_NG;
-        }
-        lit = rb_fstring(lit);
-        ADD_INSN1(ret, node, putobject, lit);
-        RB_OBJ_WRITTEN(iseq, Qundef, lit);
-        if (RSTRING_LEN(lit) == 0) first_lit = LAST_ELEMENT(ret);
+    if (str) {
+        lit = rb_str_new_mutable_parser_string(str);
+        debugp_param("nd_lit", lit);
+    }
+
+# define FLUSH_LIT() if (lit) { \
+        lit = rb_fstring(lit); \
+        ADD_INSN1(ret, lit_node, putobject, lit); \
+        RB_OBJ_WRITTEN(iseq, Qundef, lit); \
+        cnt++; \
     }
 
     while (list) {
         const NODE *const head = list->nd_head;
         if (nd_type_p(head, NODE_STR)) {
-            lit = rb_node_str_string_val(head);
-            ADD_INSN1(ret, head, putobject, lit);
-            RB_OBJ_WRITTEN(iseq, Qundef, lit);
-            lit = Qnil;
+            str = RNODE_STR(head)->string;
+            if (lit) {
+                rb_enc_str_buf_cat(lit, str->ptr, str->len, str->enc);
+            }
+            else {
+                lit = rb_str_new_mutable_parser_string(str);
+                lit_node = head;
+            }
         }
         else {
+            FLUSH_LIT();
+            lit = 0;
             CHECK(COMPILE(ret, "each string", head));
+            cnt++;
         }
-        cnt++;
         list = (struct RNode_LIST *)list->nd_next;
     }
-    if (NIL_P(lit) && first_lit) {
-        ELEM_REMOVE(first_lit);
-        --cnt;
-    }
+    FLUSH_LIT();
+# undef FLUSH_LIT
+
     *cntp = cnt;
 
     return COMPILE_OK;

parse.y

@@ -13157,9 +13157,6 @@ symbol_append(struct parser_params *p, NODE *symbols, NODE *symbol)
 static NODE *
 new_regexp(struct parser_params *p, NODE *node, int options, const YYLTYPE *loc)
 {
-    struct RNode_LIST *list;
-    NODE *prev;
-
     if (!node) {
         /* Check string is valid regex */
         rb_parser_string_t *str = STRING_NEW0();
@@ -13187,31 +13184,6 @@ new_regexp(struct parser_params *p, NODE *node, int options, const YYLTYPE *loc)
             /* Check string is valid regex */
             reg_compile(p, dreg->string, options);
         }
-        prev = node;
-        for (list = dreg->nd_next; list; list = RNODE_LIST(list->nd_next)) {
-            NODE *frag = list->nd_head;
-            enum node_type type = nd_type(frag);
-            if (type == NODE_STR || (type == NODE_DSTR && !RNODE_DSTR(frag)->nd_next)) {
-                rb_parser_string_t *tail = RNODE_STR(frag)->string;
-                if (prev && RNODE_DREGX(prev)->string) {
-                    rb_parser_string_t *lit = prev == node ? dreg->string : RNODE_STR(RNODE_LIST(prev)->nd_head)->string;
-                    if (!literal_concat0(p, lit, tail)) {
-                        return NEW_NIL(loc); /* dummy node on error */
-                    }
-                    rb_parser_str_resize(p, tail, 0);
-                    RNODE_LIST(prev)->nd_next = list->nd_next;
-                    rb_discard_node(p, list->nd_head);
-                    rb_discard_node(p, (NODE *)list);
-                    list = RNODE_LIST(prev);
-                }
-                else {
-                    prev = (NODE *)list;
-                }
-            }
-            else {
-                prev = 0;
-            }
-        }
         if (options & RE_OPTION_ONCE) {
             node = NEW_ONCE(node, loc);
         }