lexer: allow extglob wildcards as function names

interp/interp_test.go

@@ -3117,6 +3117,20 @@ hello, world
 hello, world
 `,
 	},
+	{
+		// globbing wildcard as function name
+		`@() { echo "$@"; }; @ lala; function +() { echo "$@"; }; + foo`,
+		"lala\nfoo\n",
+	},
+	{
+		`      @() { echo "$@"; }; @ lala;`,
+		"lala\n",
+	},
+	{
+		// globbing wildcard as function name but with space after the name
+		`+ () { echo "$@"; }; + foo; @ () { echo "$@"; }; @ lala; ? () { echo "$@"; }; ? bar`,
+		"foo\nlala\nbar\n",
+	},
 }
 
 var runTestsWindows = []runTest{

syntax/lexer.go

@@ -290,7 +290,7 @@ skipSpace:
 				p.advanceLitNone(r)
 			}
 		case '?', '*', '+', '@', '!':
-			if p.peekByte('(') {
+			if p.tokenizeGlob() {
 				switch r {
 				case '?':
 					p.tok = globQuest
@@ -346,6 +346,28 @@ skipSpace:
 	}
 }
 
+// tokenizeGlob determines whether the expression should be tokenized as a glob literal
+func (p *Parser) tokenizeGlob() bool {
+	if p.val == "function" {
+		return false
+	}
+	// NOTE: empty pattern list is a valid globbing syntax, eg @()
+	// but we'll operate on the "likelihood" that it is a function;
+	// only tokenize if its a non-empty pattern list
+	if p.peekBytes("()") {
+		return false
+	}
+	return p.peekByte('(')
+}
+
+func (p *Parser) peekBytes(s string) bool {
+	for p.bsp+(len(p.bs)-1) >= len(p.bs) {
+		p.fill()
+	}
+	bw := p.bsp + len(s)
+	return bw < len(p.bs) && bytes.HasPrefix(p.bs[p.bsp:bw], []byte(s))
+}
+
 func (p *Parser) peekByte(b byte) bool {
 	if p.bsp == len(p.bs) {
 		p.fill()
@@ -882,7 +904,7 @@ loop:
 			tok = _Lit
 			break loop
 		case '?', '*', '+', '@', '!':
-			if p.peekByte('(') {
+			if p.tokenizeGlob() {
 				tok = _Lit
 				break loop
 			}