Parse and emit `array.len` without a type annotation (#5151)

Test that we can still parse the old annotated form as well.

15 files changed, 68 insertions, 36 deletions

diff --git a/src/passes/Print.cpp b/src/passes/Print.cpp
index 0f4a314ae..d0f19f78e 100644
--- a/src/passes/Print.cpp
+++ b/src/passes/Print.cpp
@@ -2236,14 +2236,7 @@ struct PrintExpressionContents
     printMedium(o, "array.set ");
     TypeNamePrinter(o, wasm).print(curr->ref->type.getHeapType());
   }
-  void visitArrayLen(ArrayLen* curr) {
-    printMedium(o, "array.len ");
-    if (curr->ref->type == Type::unreachable) {
-      TypeNamePrinter(o, wasm).print(HeapType::array);
-    } else {
-      TypeNamePrinter(o, wasm).print(curr->ref->type.getHeapType());
-    }
-  }
+  void visitArrayLen(ArrayLen* curr) { printMedium(o, "array.len"); }
   void visitArrayCopy(ArrayCopy* curr) {
     if (printUnreachableOrNullReplacement(curr->srcRef) ||
         printUnreachableOrNullReplacement(curr->destRef)) {
diff --git a/src/wasm-binary.h b/src/wasm-binary.h
index 709c0c9f9..db123515a 100644
--- a/src/wasm-binary.h
+++ b/src/wasm-binary.h
@@ -1111,8 +1111,9 @@ enum ASTNodes {
   ArrayGetS = 0x14,
   ArrayGetU = 0x15,
   ArraySet = 0x16,
-  ArrayLen = 0x17,
+  ArrayLenAnnotated = 0x17,
   ArrayCopy = 0x18,
+  ArrayLen = 0x19,
   ArrayInitStatic = 0x1a,
   ArrayNew = 0x1b,
   ArrayNewDefault = 0x1c,
diff --git a/src/wasm/wasm-binary.cpp b/src/wasm/wasm-binary.cpp
index 13c28f151..54c904459 100644
--- a/src/wasm/wasm-binary.cpp
+++ b/src/wasm/wasm-binary.cpp
@@ -7111,11 +7111,12 @@ bool WasmBinaryBuilder::maybeVisitArraySet(Expression*& out, uint32_t code) {
 }
 
 bool WasmBinaryBuilder::maybeVisitArrayLen(Expression*& out, uint32_t code) {
-  if (code != BinaryConsts::ArrayLen) {
+  if (code == BinaryConsts::ArrayLenAnnotated) {
+    // Ignore the type annotation and don't bother validating it.
+    getU32LEB();
+  } else if (code != BinaryConsts::ArrayLen) {
     return false;
   }
-  // Ignore the type annotation and don't bother validating it.
-  getU32LEB();
   auto* ref = popNonVoidExpression();
   out = Builder(wasm).makeArrayLen(ref);
   return true;
diff --git a/src/wasm/wasm-s-parser.cpp b/src/wasm/wasm-s-parser.cpp
index 41a0619f9..78ac84f42 100644
--- a/src/wasm/wasm-s-parser.cpp
+++ b/src/wasm/wasm-s-parser.cpp
@@ -3027,9 +3027,14 @@ Expression* SExpressionWasmBuilder::makeArraySet(Element& s) {
 }
 
 Expression* SExpressionWasmBuilder::makeArrayLen(Element& s) {
-  // Ignore the type annotation and don't bother validating it.
-  parseHeapType(*s[1]);
-  auto ref = parseExpression(*s[2]);
+  // There may or may not be a type annotation.
+  Index i = 1;
+  try {
+    parseHeapType(*s[i]);
+    ++i;
+  } catch (...) {
+  }
+  auto ref = parseExpression(*s[i]);
   return Builder(wasm).makeArrayLen(ref);
 }
 
diff --git a/src/wasm/wasm-stack.cpp b/src/wasm/wasm-stack.cpp
index 7c1c121d5..f30e4be82 100644
--- a/src/wasm/wasm-stack.cpp
+++ b/src/wasm/wasm-stack.cpp
@@ -2169,8 +2169,6 @@ void BinaryInstWriter::visitArraySet(ArraySet* curr) {
 
 void BinaryInstWriter::visitArrayLen(ArrayLen* curr) {
   o << int8_t(BinaryConsts::GCPrefix) << U32LEB(BinaryConsts::ArrayLen);
-  // Unused type index.
-  o << U32LEB(0);
 }
 
 void BinaryInstWriter::visitArrayCopy(ArrayCopy* curr) {
diff --git a/test/example/c-api-kitchen-sink.txt b/test/example/c-api-kitchen-sink.txt
index 5c6400ce5..983c80d90 100644
--- a/test/example/c-api-kitchen-sink.txt
+++ b/test/example/c-api-kitchen-sink.txt
@@ -2213,7 +2213,7 @@ BinaryenFeatureAll: 126975
        (i32.const 42)
       )
       (drop
-       (array.len $[mut:i8]
+       (array.len
         (global.get $i8Array-global)
        )
       )
diff --git a/test/heap-types.wast.from-wast b/test/heap-types.wast.from-wast
index 4994857e2..dc41451cd 100644
--- a/test/heap-types.wast.from-wast
+++ b/test/heap-types.wast.from-wast
@@ -150,7 +150,7 @@
    )
   )
   (drop
-   (array.len $vector
+   (array.len
     (local.get $x)
    )
   )
@@ -383,7 +383,7 @@
  )
  (func $unreachables-array-6
   (drop
-   (array.len array
+   (array.len
     (unreachable)
    )
   )
diff --git a/test/heap-types.wast.fromBinary b/test/heap-types.wast.fromBinary
index dc55e1e8f..f16e07c7f 100644
--- a/test/heap-types.wast.fromBinary
+++ b/test/heap-types.wast.fromBinary
@@ -148,7 +148,7 @@
    )
   )
   (drop
-   (array.len $vector
+   (array.len
     (local.get $x)
    )
   )
diff --git a/test/heap-types.wast.fromBinary.noDebugInfo b/test/heap-types.wast.fromBinary.noDebugInfo
index 9c5194750..53c3b0bd6 100644
--- a/test/heap-types.wast.fromBinary.noDebugInfo
+++ b/test/heap-types.wast.fromBinary.noDebugInfo
@@ -148,7 +148,7 @@
    )
   )
   (drop
-   (array.len $[mut:f64]
+   (array.len
     (local.get $0)
    )
   )
diff --git a/test/lit/arrays.wast b/test/lit/arrays.wast
index b79ca1844..31aead1cd 100644
--- a/test/lit/arrays.wast
+++ b/test/lit/arrays.wast
@@ -11,25 +11,25 @@
 
 (module
  (type $byte-array (array (mut i8)))
+ ;; CHECK:      (type $arrayref_=>_i32 (func (param arrayref) (result i32)))
+
  ;; CHECK:      (type $ref|array|_=>_i32 (func (param (ref array)) (result i32)))
 
  ;; CHECK:      (type $nullref_=>_i32 (func (param nullref) (result i32)))
 
- ;; CHECK:      (type $arrayref_=>_i32 (func (param arrayref) (result i32)))
-
  ;; CHECK:      (func $len (param $a (ref array)) (result i32)
- ;; CHECK-NEXT:  (array.len array
+ ;; CHECK-NEXT:  (array.len
  ;; CHECK-NEXT:   (local.get $a)
  ;; CHECK-NEXT:  )
  ;; CHECK-NEXT: )
+ ;; ROUNDTRIP:      (type $arrayref_=>_i32 (func (param arrayref) (result i32)))
+
  ;; ROUNDTRIP:      (type $ref|array|_=>_i32 (func (param (ref array)) (result i32)))
 
  ;; ROUNDTRIP:      (type $nullref_=>_i32 (func (param nullref) (result i32)))
 
- ;; ROUNDTRIP:      (type $arrayref_=>_i32 (func (param arrayref) (result i32)))
-
  ;; ROUNDTRIP:      (func $len (param $a (ref array)) (result i32)
- ;; ROUNDTRIP-NEXT:  (array.len array
+ ;; ROUNDTRIP-NEXT:  (array.len
  ;; ROUNDTRIP-NEXT:   (local.get $a)
  ;; ROUNDTRIP-NEXT:  )
  ;; ROUNDTRIP-NEXT: )
@@ -41,12 +41,12 @@
  )
 
  ;; CHECK:      (func $impossible-len (param $none nullref) (result i32)
- ;; CHECK-NEXT:  (array.len none
+ ;; CHECK-NEXT:  (array.len
  ;; CHECK-NEXT:   (local.get $none)
  ;; CHECK-NEXT:  )
  ;; CHECK-NEXT: )
  ;; ROUNDTRIP:      (func $impossible-len (param $none nullref) (result i32)
- ;; ROUNDTRIP-NEXT:  (array.len none
+ ;; ROUNDTRIP-NEXT:  (array.len
  ;; ROUNDTRIP-NEXT:   (local.get $none)
  ;; ROUNDTRIP-NEXT:  )
  ;; ROUNDTRIP-NEXT: )
@@ -57,7 +57,7 @@
  )
 
  ;; CHECK:      (func $unreachable-len (param $a arrayref) (result i32)
- ;; CHECK-NEXT:  (array.len array
+ ;; CHECK-NEXT:  (array.len
  ;; CHECK-NEXT:   (unreachable)
  ;; CHECK-NEXT:  )
  ;; CHECK-NEXT: )
@@ -69,4 +69,20 @@
    (unreachable)
   )
  )
+
+ ;; CHECK:      (func $unannotated-len (param $a arrayref) (result i32)
+ ;; CHECK-NEXT:  (array.len
+ ;; CHECK-NEXT:   (local.get $a)
+ ;; CHECK-NEXT:  )
+ ;; CHECK-NEXT: )
+ ;; ROUNDTRIP:      (func $unannotated-len (param $a arrayref) (result i32)
+ ;; ROUNDTRIP-NEXT:  (array.len
+ ;; ROUNDTRIP-NEXT:   (local.get $a)
+ ;; ROUNDTRIP-NEXT:  )
+ ;; ROUNDTRIP-NEXT: )
+ (func $unannotated-len (param $a arrayref) (result i32)
+  (array.len
+   (local.get $a)
+  )
+ )
 )
diff --git a/test/lit/binary/annotated-array-len.test b/test/lit/binary/annotated-array-len.test
new file mode 100644
index 000000000..a8eb08fe7
--- /dev/null
+++ b/test/lit/binary/annotated-array-len.test
@@ -0,0 +1,18 @@
+;; NOTE: Assertions have been generated by update_lit_checks.py --all-items and should not be edited.
+
+;; Test the we can properly parse the annotated array.len format that we no
+;; longer emit.
+
+;; RUN: wasm-dis %s.wasm -all --nominal | filecheck %s
+
+;; CHECK:      (type $none_=>_i32 (func_subtype (result i32) func))
+
+;; CHECK:      (type $[mut:i8] (array_subtype (mut i8) data))
+
+;; CHECK:      (func $0 (type $none_=>_i32) (result i32)
+;; CHECK-NEXT:  (array.len
+;; CHECK-NEXT:   (array.new_default $[mut:i8]
+;; CHECK-NEXT:    (i32.const 0)
+;; CHECK-NEXT:   )
+;; CHECK-NEXT:  )
+;; CHECK-NEXT: )
diff --git a/test/lit/binary/annotated-array-len.test.wasm b/test/lit/binary/annotated-array-len.test.wasm
new file mode 100644
index 000000000..b199f58de
--- /dev/null
+++ b/test/lit/binary/annotated-array-len.test.wasm
diff --git a/test/lit/passes/gufa-refs.wast b/test/lit/passes/gufa-refs.wast
index 09de96fca..ec58e10a0 100644
--- a/test/lit/passes/gufa-refs.wast
+++ b/test/lit/passes/gufa-refs.wast
@@ -4150,7 +4150,7 @@
 
   ;; CHECK:      (func $arrays (type $ref|$B|_=>_none) (param $B (ref $B))
   ;; CHECK-NEXT:  (drop
-  ;; CHECK-NEXT:   (array.len $B
+  ;; CHECK-NEXT:   (array.len
   ;; CHECK-NEXT:    (array.init_static $B
   ;; CHECK-NEXT:     (ref.null none)
   ;; CHECK-NEXT:     (ref.null none)
diff --git a/test/lit/passes/optimize-instructions-gc.wast b/test/lit/passes/optimize-instructions-gc.wast
index c0d106250..fb50d103f 100644
--- a/test/lit/passes/optimize-instructions-gc.wast
+++ b/test/lit/passes/optimize-instructions-gc.wast
@@ -766,7 +766,7 @@
   ;; CHECK-NEXT:   )
   ;; CHECK-NEXT:  )
   ;; CHECK-NEXT:  (drop
-  ;; CHECK-NEXT:   (array.len $array
+  ;; CHECK-NEXT:   (array.len
   ;; CHECK-NEXT:    (local.get $y)
   ;; CHECK-NEXT:   )
   ;; CHECK-NEXT:  )
@@ -801,7 +801,7 @@
   ;; NOMNL-NEXT:   )
   ;; NOMNL-NEXT:  )
   ;; NOMNL-NEXT:  (drop
-  ;; NOMNL-NEXT:   (array.len $array
+  ;; NOMNL-NEXT:   (array.len
   ;; NOMNL-NEXT:    (local.get $y)
   ;; NOMNL-NEXT:   )
   ;; NOMNL-NEXT:  )
diff --git a/test/passes/Oz_fuzz-exec_all-features.txt b/test/passes/Oz_fuzz-exec_all-features.txt
index e19110848..a38193d08 100644
--- a/test/passes/Oz_fuzz-exec_all-features.txt
+++ b/test/passes/Oz_fuzz-exec_all-features.txt
@@ -123,7 +123,7 @@
  (func $1 (; has Stack IR ;)
   (local $0 (ref $bytes))
   (call $log
-   (array.len $bytes
+   (array.len
     (local.tee $0
      (array.new $bytes
       (i32.const 42)
@@ -319,7 +319,7 @@
  (func $24 (; has Stack IR ;)
   (local $0 (ref $bytes))
   (call $log
-   (array.len $bytes
+   (array.len
     (local.tee $0
      (array.init_static $bytes
       (i32.const 42)