Create parser plugin "topLevelAwait" (#10449)

* Create parser plugin "topLevelAwait"

* Update test262 whitelist

* Update ts typings

* Fix "sourceType: unambiguous" with TLA

* Ambiguous tokens after await

* Update await %x(0)

* typo [skip ci]

* Typo [skip ci]

Co-Authored-By: Brian Ng <bng412@gmail.com>

packages/babel-parser/src/index.js

@@ -25,15 +25,32 @@ export function parse(input: string, options?: Options): File {
       const parser = getParser(options, input);
       const ast = parser.parse();
 
-      // Rather than try to parse as a script first, we opt to parse as a module and convert back
-      // to a script where possible to avoid having to do a full re-parse of the input content.
-      if (!parser.sawUnambiguousESM) ast.program.sourceType = "script";
+      if (parser.sawUnambiguousESM) {
+        return ast;
+      }
+
+      if (parser.ambiguousScriptDifferentAst) {
+        // Top level await introduces code which can be both a valid script and
+        // a valid module, but which produces different ASTs:
+        //    await
+        //    0
+        // can be parsed either as an AwaitExpression, or as two ExpressionStatements.
+        try {
+          options.sourceType = "script";
+          return getParser(options, input).parse();
+        } catch {}
+      } else {
+        // This is both a valid module and a valid script, but
+        // we parse it as a script by default
+        ast.program.sourceType = "script";
+      }
+
       return ast;
     } catch (moduleError) {
       try {
         options.sourceType = "script";
         return getParser(options, input).parse();
-      } catch (scriptError) {}
+      } catch {}
 
       throw moduleError;
     }

packages/babel-parser/src/parser/base.js

@@ -13,6 +13,7 @@ export default class BaseParser {
   plugins: PluginsMap;
   filename: ?string;
   sawUnambiguousESM: boolean = false;
+  ambiguousScriptDifferentAst: boolean = false;
 
   // Initialized by Tokenizer
   state: State;

packages/babel-parser/src/parser/expression.js

@@ -2208,6 +2208,7 @@ export default class ExpressionParser extends LValParser {
   isAwaitAllowed(): boolean {
     if (this.scope.inFunction) return this.scope.inAsync;
     if (this.options.allowAwaitOutsideFunction) return true;
+    if (this.hasPlugin("topLevelAwait")) return this.inModule;
     return false;
   }
 
@@ -2234,9 +2235,33 @@ export default class ExpressionParser extends LValParser {
       );
     }
 
+    if (!this.scope.inFunction && !this.options.allowAwaitOutsideFunction) {
+      if (
+        this.hasPrecedingLineBreak() ||
+        // All the following expressions are ambiguous:
+        //   await + 0, await - 0, await ( 0 ), await [ 0 ], await / 0 /u, await ``
+        this.match(tt.plusMin) ||
+        this.match(tt.parenL) ||
+        this.match(tt.bracketL) ||
+        this.match(tt.backQuote) ||
+        // Sometimes the tokenizer generates tt.slash for regexps, and this is
+        // handler by parseExprAtom
+        this.match(tt.regexp) ||
+        this.match(tt.slash) ||
+        // This code could be parsed both as a modulo operator or as an intrinsic:
+        //   await %x(0)
+        (this.hasPlugin("v8intrinsic") && this.match(tt.modulo))
+      ) {
+        this.ambiguousScriptDifferentAst = true;
+      } else {
+        this.sawUnambiguousESM = true;
+      }
+    }
+
     if (!this.state.soloAwait) {
       node.argument = this.parseMaybeUnary();
     }
+
     return this.finishNode(node, "AwaitExpression");
   }
 

packages/babel-parser/src/parser/statement.js

@@ -495,11 +495,7 @@ export default class StatementParser extends ExpressionParser {
     this.state.labels.push(loopLabel);
 
     let awaitAt = -1;
-    if (
-      (this.scope.inAsync ||
-        (!this.scope.inFunction && this.options.allowAwaitOutsideFunction)) &&
-      this.eatContextual("await")
-    ) {
+    if (this.isAwaitAllowed() && this.eatContextual("await")) {
       awaitAt = this.state.lastTokStart;
     }
     this.scope.enter(SCOPE_OTHER);