found redundant code in file natSystemProperties.cc

found redundant code in file natSystemProperties.cc

[Martin Ettl]

Hello friends,

i have checked the sources of gcc-4.4.1 with the static code analysis tool cppcheck. It brought up an issue in file gcc-4.4.1/libjava/gnu/classpath/natSystemProperties.cc at line 278. Lets take a look at the sourcecode:


// ...

  while (buffer != NULL)
    {
      if (getcwd (buffer, buflen) != NULL)
	{
	  SET ("user.dir", buffer);
	  break;
	}
      if (errno != ERANGE)
	break;
      buflen = 2 * buflen;
      char *orig_buf = buffer;
      buffer = (char *) realloc (buffer, buflen);
      if (buffer == NULL)
	free (orig_buf);
    }
278 if (buffer != NULL)
    free (buffer);
....

Here the line 278 is redundant, because it is safe to deallocate a NULL-Pointer! In other words, the if-statement can be removed.

Best regards

Ettl Martin


-- 
Neu: GMX Doppel-FLAT mit Internet-Flatrate + Telefon-Flatrate
für nur 19,99 Euro/mtl.!* http://portal.gmx.net/de/go/dsl02

Re: found redundant code in file natSystemProperties.cc

[Andrew Haley]

Martin Ettl wrote:
> Hello friends,
> 
> i have checked the sources of gcc-4.4.1 with the static code analysis tool cppcheck. It brought up an issue in file gcc-4.4.1/libjava/gnu/classpath/natSystemProperties.cc at line 278. Lets take a look at the sourcecode:

> 278 if (buffer != NULL)
>     free (buffer);
> ....
> 
> Here the line 278 is redundant, because it is safe to deallocate a NULL-Pointer! In other words, the if-statement can be removed.

Well yes, but it saves a library call.

Harbison and Steele note "If the argument to free is a null pointer
then no action should occur, but this has been known to cause trouble
in some C implementations."  This is probably just ancient history,
though.

Andrew.