Re: [Help-gsl] Re: free

[Jochen Küpper]

John D Lamb <address@hidden> writes:
> Brian Gough wrote:
>> Jochen Küpper writes:

>>> Shouldn't the GSL free-routines be changed like the following patch?

>> My thinking on that was that for most people calling the free()
>> functions on a null pointer is usually an error rather than by design,
>> so it's helpful to get a segmentation fault.  

> A check in the _free codes would also add a small overhead.

Two good arguments; I agree that the current code is what it should
be. 

Maybe a comment could be added in the code:
,----
| Index: block/init_source.c
| ===================================================================
| RCS file: /cvs/gsl/gsl/block/init_source.c,v
| retrieving revision 1.7
| diff -u -r1.7 init_source.c
| --- block/init_source.c       26 Jun 2005 13:26:59 -0000      1.7
| +++ block/init_source.c       12 Jun 2006 21:20:39 -0000
| @@ -74,6 +74,11 @@
|  void
|  FUNCTION (gsl_block, free) (TYPE (gsl_block) * b)
|  {
| +  // no check for b == 0 performed:
| +  // - For most people calling this function on a null pointer is
| +  //   usually an error, rather than by design, so it is helpful
| +  //   to get a segmentation fault.
| +  // - The check imposes a small performance overhead.
|    free (b->data);
|    free (b);
|  }
`----
This might be rephrased shorter/better by a native speaker... Thinking
of it, maybe it should even go into the documentation.

Greetings,
Jochen
-- 
Einigkeit und Recht und Freiheit                http://www.Jochen-Kuepper.de
    Liberté, Égalité, Fraternité                GnuPG key: CC1B0B4D
        (Part 3 you find in my messages before fall 2003.)