From f6b94f0cba4f58722858fe680bfad4a859ccaa09 Mon Sep 17 00:00:00 2001
From: Aaron Patterson <tenderlove@ruby-lang.org>
Date: Mon, 6 Jun 2022 17:27:56 -0700
Subject: [PATCH] Emit special instruction for array literal + .hash

This commit introduces a new instruction `opt_newarray_hash` which is
used when there is an array literal with a `.hash` method called on the
literal.  For example:

```
[a, b, c].hash
```

Will emit an `opt_newarray_hash` instruction.  This instruction falls
back to a method call if the hash method has been monkeypatched, similar
to `opt_newarray_min`.

Co-authored-by: John Hawthorn <jhawthorn@github.com>
---
 array.c          | 30 ++++++++++++++++++------------
 compile.c        |  4 ++++
 defs/id.def      |  1 +
 insns.def        | 12 ++++++++++++
 internal/array.h |  1 +
 vm.c             |  1 +
 vm_core.h        |  1 +
 vm_insnhelper.c  | 12 ++++++++++++
 8 files changed, 50 insertions(+), 12 deletions(-)

diff --git a/array.c b/array.c
index 7b3f5bd0b0..e714652520 100644
--- a/array.c
+++ b/array.c
@@ -5306,6 +5306,23 @@ rb_ary_eql(VALUE ary1, VALUE ary2)
     return rb_exec_recursive_paired(recursive_eql, ary1, ary2, ary2);
 }
 
+VALUE
+rb_ary_hash_values(long len, const VALUE *elements)
+{
+    long i;
+    st_index_t h;
+    VALUE n;
+
+    h = rb_hash_start(len);
+    h = rb_hash_uint(h, (st_index_t)rb_ary_hash_values);
+    for (i=0; i<len; i++) {
+	n = rb_hash(elements[i]);
+	h = rb_hash_uint(h, NUM2LONG(n));
+    }
+    h = rb_hash_end(h);
+    return ST2FIX(h);
+}
+
 /*
  *  call-seq:
  *    array.hash -> integer
@@ -5322,18 +5339,7 @@ rb_ary_eql(VALUE ary1, VALUE ary2)
 static VALUE
 rb_ary_hash(VALUE ary)
 {
-    long i;
-    st_index_t h;
-    VALUE n;
-
-    h = rb_hash_start(RARRAY_LEN(ary));
-    h = rb_hash_uint(h, (st_index_t)rb_ary_hash);
-    for (i=0; i<RARRAY_LEN(ary); i++) {
-	n = rb_hash(RARRAY_AREF(ary, i));
-	h = rb_hash_uint(h, NUM2LONG(n));
-    }
-    h = rb_hash_end(h);
-    return ST2FIX(h);
+    return rb_ary_hash_values(RARRAY_LEN(ary), RARRAY_CONST_PTR(ary));
 }
 
 /*
diff --git a/compile.c b/compile.c
index 67641fcfb6..c7a9a7deee 100644
--- a/compile.c
+++ b/compile.c
@@ -3489,6 +3489,10 @@ iseq_specialized_instruction(rb_iseq_t *iseq, INSN *iobj)
 		    iobj->insn_id = BIN(opt_newarray_min);
 		    ELEM_REMOVE(&niobj->link);
 		    return COMPILE_OK;
+		  case idHash:
+		    iobj->insn_id = BIN(opt_newarray_hash);
+		    ELEM_REMOVE(&niobj->link);
+		    return COMPILE_OK;
 		}
 	    }
 	}
diff --git a/defs/id.def b/defs/id.def
index 097e34e405..1ce6710094 100644
--- a/defs/id.def
+++ b/defs/id.def
@@ -2,6 +2,7 @@
 firstline, predefined = __LINE__+1, %[\
   max
   min
+  hash
   freeze
   nil?
   inspect
diff --git a/insns.def b/insns.def
index b5dea9c10c..cac1bf372d 100644
--- a/insns.def
+++ b/insns.def
@@ -867,6 +867,18 @@ opt_newarray_min
     val = vm_opt_newarray_min(ec, num, STACK_ADDR_FROM_TOP(num));
 }
 
+DEFINE_INSN
+opt_newarray_hash
+(rb_num_t num)
+(...)
+(VALUE val)
+/* Same discussion as opt_newarray_max. */
+// attr bool leaf = false; /* has rb_funcall() */
+// attr rb_snum_t sp_inc = 1 - (rb_snum_t)num;
+{
+    val = vm_opt_newarray_hash(ec, num, STACK_ADDR_FROM_TOP(num));
+}
+
 /* super(args) # args.size => num */
 DEFINE_INSN
 invokesuper
diff --git a/internal/array.h b/internal/array.h
index 60f66f31bf..acd286941d 100644
--- a/internal/array.h
+++ b/internal/array.h
@@ -21,6 +21,7 @@
 #define RARRAY_PTR_IN_USE_FLAG FL_USER14
 
 /* array.c */
+VALUE rb_ary_hash_values(long len, const VALUE *elements);
 VALUE rb_ary_last(int, const VALUE *, VALUE);
 void rb_ary_set_len(VALUE, long);
 void rb_ary_delete_same(VALUE, VALUE);
diff --git a/vm.c b/vm.c
index 76041b0b91..7204d5171d 100644
--- a/vm.c
+++ b/vm.c
@@ -1969,6 +1969,7 @@ vm_init_redefined_flag(void)
     OP(UMinus, UMINUS), (C(String));
     OP(Max, MAX), (C(Array));
     OP(Min, MIN), (C(Array));
+    OP(Hash, HASH), (C(Array));
     OP(Call, CALL), (C(Proc));
     OP(And, AND), (C(Integer));
     OP(Or, OR), (C(Integer));
diff --git a/vm_core.h b/vm_core.h
index 2fca0aecd2..f93b191262 100644
--- a/vm_core.h
+++ b/vm_core.h
@@ -576,6 +576,7 @@ enum ruby_basic_operators {
     BOP_UMINUS,
     BOP_MAX,
     BOP_MIN,
+    BOP_HASH,
     BOP_CALL,
     BOP_AND,
     BOP_OR,
diff --git a/vm_insnhelper.c b/vm_insnhelper.c
index 1e40088ffa..b409cf509b 100644
--- a/vm_insnhelper.c
+++ b/vm_insnhelper.c
@@ -4922,6 +4922,18 @@ vm_opt_newarray_min(rb_execution_context_t *ec, rb_num_t num, const VALUE *ptr)
     }
 }
 
+static VALUE
+vm_opt_newarray_hash(rb_execution_context_t *ec, rb_num_t num, const VALUE *ptr)
+{
+    // If Array#hash is _not_ monkeypatched, use the optimized call
+    if (BASIC_OP_UNREDEFINED_P(BOP_HASH, ARRAY_REDEFINED_OP_FLAG)) {
+        return rb_ary_hash_values(num, ptr);
+    }
+    else {
+        return rb_vm_call_with_refinements(ec, rb_ary_new4(num, ptr), idHash, 0, NULL, RB_NO_KEYWORDS);
+    }
+}
+
 #undef id_cmp
 
 #define IMEMO_CONST_CACHE_SHAREABLE IMEMO_FL_USER0
-- 
2.36.0