Considering the following code about the wrong octal number:

```
class Digit {
        int n = 079;
}
```

When compiling it, the javac would output:


```
Digit.java:2: error: integer number too large
        int n = 079;
                   ^
1 error
```


Because the lexer(JavaTokenizer) incorrectly reads `079` as a token.
So the parser(JavacParser) would throw the NumberFormatException when using the method Convert#string2int.
It is an issue of the lexer but it is caught by the parser, which is not our expectation.

Considering the following code about the wrong binary, hexadecimal and floating point numbers:

```
class Digit {
        double a = 9.f2;
        int n = 0b123;
        int n = 0x12r3;
}
```

The compiler would output:

```
Digit.java:2: error: ';' expected
        double a = 9.f2;
                              ^
Digit.java:3: error: ';' expected
        int n = 0b123;
                         ^
Digit.java:4: error: ';' expected
        int n = 0x12r3;
                           ^
Digit.java:4: error: <identifier> expected
        int n = 0x12r3;
                              ^
4 errors
```

The output of the wrong octal numbers should be similar to them, which is like the following output:

```

Digit.java:2: error: ';' expected
        int n = 079;
                       ^
1 error
```