mruby/ops.h: add new instructions `OP_GETIDX` and `OP_SETIDX`.

Which represent `obj[int]` and `obj[int]=val` respectively where `obj`
is either `string`, `array` or `hash`, so that index access could be
faster. When `obj` is not assumed type or `R(a+1)` is not integer, the
instructions fallback to method calls.

include/mruby/ops.h

@@ -47,6 +47,8 @@ OPCODE(GETMCNST,   BB)       /* R(a) = R(a)::Syms(b) */
 OPCODE(SETMCNST,   BB)       /* R(a+1)::Syms(b) = R(a) */
 OPCODE(GETUPVAR,   BBB)      /* R(a) = uvget(b,c) */
 OPCODE(SETUPVAR,   BBB)      /* uvset(b,c,R(a)) */
+OPCODE(GETIDX,     B)        /* R(a) = R(a)[R(a+1)] */
+OPCODE(SETIDX,     B)        /* R(a)[R(a+1)] = R(a+2) */
 OPCODE(JMP,        S)        /* pc+=a */
 OPCODE(JMPIF,      BS)       /* if R(a) pc+=b */
 OPCODE(JMPNOT,     BS)       /* if !R(a) pc+=b */

mrbgems/mruby-compiler/core/codegen.c

@@ -832,6 +832,10 @@ static mrb_bool
 gen_binop(codegen_scope *s, mrb_sym op, uint16_t dst)
 {
   if (no_peephole(s)) return FALSE;
+  else if (op == MRB_OPSYM_2(s->mrb, aref)) {
+    genop_1(s, OP_GETIDX, dst);
+    return TRUE;
+  }
   else {
     struct mrb_insn_data data = mrb_last_insn(s);
     mrb_int n, n0;

src/codedump.c

@@ -265,6 +265,12 @@ codedump(mrb_state *mrb, const mrb_irep *irep)
       printf("OP_SETCV\t%s\tR%d\t", mrb_sym_dump(mrb, irep->syms[b]), a);
       print_lv_a(mrb, irep, a);
       break;
+    CASE(OP_GETIDX, B):
+      printf("OP_GETIDX\tR%d\tR%d\n", a, a+1);
+      break;
+    CASE(OP_SETIDX, B):
+      printf("OP_SETIDX\tR%d\tR%d\tR%d\n", a, a+1, a+2);
+      break;
     CASE(OP_JMP, S):
       i = pc - irep->iseq;
       printf("OP_JMP\t\t%03d\n", (int)i+(int16_t)a);

src/string.c

@@ -1146,7 +1146,7 @@ range_arg:
   return STR_OUT_OF_RANGE;
 }
 
-static mrb_value
+mrb_value
 mrb_str_aref(mrb_state *mrb, mrb_value str, mrb_value indx, mrb_value alen)
 {
   mrb_int beg, len;

src/vm.c

@@ -1063,6 +1063,7 @@ get_send_args(mrb_state *mrb, mrb_int argc, mrb_value *regs)
 mrb_value mrb_obj_missing(mrb_state *mrb, mrb_value mod);
 void mrb_hash_check_kdict(mrb_state *mrb, mrb_value self);
 void mrb_method_added(mrb_state *mrb, struct RClass *c, mrb_sym mid);
+mrb_value mrb_str_aref(mrb_state *mrb, mrb_value str, mrb_value idx, mrb_value len);
 
 MRB_API mrb_value
 mrb_vm_exec(mrb_state *mrb, const struct RProc *proc, const mrb_code *pc)
@@ -1252,6 +1253,34 @@ RETRY_TRY_BLOCK:
       NEXT;
     }
 
+    CASE(OP_GETIDX, B) {
+      mrb_value va = regs[a], vb = regs[a+1];
+      switch (mrb_type(va)) {
+      case MRB_TT_ARRAY:
+        if (!mrb_integer_p(vb)) goto getidx_fallback;
+        regs[a] = mrb_ary_entry(va, mrb_integer(vb));
+        break;
+      case MRB_TT_HASH:
+        regs[a] = mrb_hash_get(mrb, va, vb);
+        break;
+      case MRB_TT_STRING:
+        regs[a] = mrb_str_aref(mrb, va, vb, mrb_undef_value());
+        break;
+      default:
+      getidx_fallback:
+        c = 1;
+        mid = MRB_OPSYM(aref);
+        goto L_SEND_SYM;
+      }
+      NEXT;
+    }
+
+    CASE(OP_SETIDX, B) {
+      c = 1;
+      mid = MRB_OPSYM(aset);
+      goto L_SEND_SYM;
+    }
+
     CASE(OP_GETCONST, BB) {
       mrb_value val;
       mrb_sym sym = syms[b];