Enables | close to variable token

scanner/lex.go

@@ -283,10 +283,22 @@ func lexStart(l *Lexer) stateFn {
 		absorbIdentifier(l)
 
 		next := l.peek()
-		if next != eof && !isSpace(next) &&
-			!isEndOfLine(next) && next != ';' &&
-			next != ')' && next != ',' && next != '+' &&
-			next != '[' && next != ']' && next != '(' {
+
+		// required to avoid the sintaxes below:
+		//   rm -rf $HOME/projects
+		//   rm -rf $GOPATH/test
+		// for being interpreted as:
+		//   rm -rf $HOME /projects
+		//   rm -rf $GOPATH /test
+		//
+		// Below are the valid suffixes for variables:
+		//   $HOME;
+		//   $HOME[    # used in: $HOME[0]
+		//   $HOME(    # used in: $callback()
+		//   $HOME+    # used in: $HOME+"/src"
+		//   $HOME]    # used in: $list[$index]
+		//   $HOME)    # used in: call($HOME)
+		if !isValidVariableSuffix(next) {
 			l.errorf("Unrecognized character in action: %#U", next)
 			return nil
 		}
@@ -550,3 +562,11 @@ func isAlpha(r rune) bool {
 func isEndOfLine(r rune) bool {
 	return r == '\r' || r == '\n'
 }
+
+func isValidVariableSuffix(r rune) bool {
+	return r == eof || isSpace(r) ||
+		isEndOfLine(r) || r == ';' ||
+		r == ')' || r == ',' || r == '+' ||
+		r == '[' || r == ']' || r == '(' ||
+		r == '|'
+}

scanner/lex_regression_test.go

@@ -207,4 +207,13 @@ func TestLexerIssue127(t *testing.T) {
 	}
 
 	testTable("test127", `rm -rf $HOME/.vim`, expected, t)
+
+	expected = []Token{
+		{typ: token.Ident, val: "rm"},
+		{typ: token.Arg, val: "-rf"},
+		{typ: token.Illegal, val: "test127:1:12: Unrecognized character in action: U+002E '.'"},
+		{typ: token.EOF},
+	}
+
+	testTable("test127", `rm -rf $HOME.vim`, expected, t)
 }

scanner/lex_test.go

@@ -1827,3 +1827,52 @@ func TestLexerLongAssignment(t *testing.T) {
 	jq ".GroupId" |
 	xargs echo -n)`, expected, t)
 }
+
+func TestLexerVariable(t *testing.T) {
+	base := []Token{
+		{typ: token.Variable, val: "$HOME"},
+	}
+
+	for r, tok := range map[rune]token.Token{
+		';': token.Semicolon,
+		'+': token.Plus,
+		'[': token.LBrack,
+		']': token.RBrack,
+		'(': token.LParen,
+		')': token.RParen,
+		',': token.Comma,
+		'|': token.Pipe,
+	} {
+		expected := append(base, Token{typ: tok, val: string(r)})
+
+		if r == ')' {
+			expected = append(expected, Token{typ: token.Semicolon, val: ";"})
+		}
+
+		expected = append(expected, Token{typ: token.EOF})
+		testTable("test variable", `$HOME`+string(r), expected, t)
+	}
+
+	// must fail
+	for _, r := range []rune{
+		'!', '@', '$', '%', '&', '*', '-', '=', '"', '\'',
+		'`', '{', '}', '<', '>', '.', ':', '?', '/', '\\',
+	} {
+		l := Lex("test", "$GOPATH"+string(r))
+
+		var results []Token
+
+		for tok := range l.Tokens {
+			results = append(results, tok)
+		}
+
+		if len(results) != 2 {
+			t.Fatalf("Expected 2 token, found %d", len(results))
+		}
+
+		if results[0].typ != token.Illegal {
+			t.Fatalf("Expected illegal token, but found %v", results[0])
+		}
+	}
+
+}