RFC: Handling realloc(0)

[Howard Chu]

Nick Clifton via Binutils wrote:
> Hi Guys,
> 
>   A recent binutils bug report (PR 27797) has brought to light the fact
>   that the behaviour of realloc(0) is not well defined.  So I am
>   proposing the attached patch to change the behaviour of the
>   bfd_realloc and bfd_realloc_or_free functions.  With this patch
>   applied:
> 
>     bfd_realloc (NULL, 0)          will return bfd_malloc (1)

Is bfd_malloc(0) also treated as bfd_malloc(1) ?

>     bfd_realloc (ptr, 0)           will return ptr

Shouldn't this return NULL? a standard realloc(ptr, 0) == free(ptr).

>     bfd_realloc_of_free (NULL, 0)  will return NULL (and no error set)
>     bfd_realloc_or_free (ptr, 0)   will return NULL and ptr will have been freed.
> 
>   What do people think ?  Any objections ?
> 
> Cheers
>   Nick
> 
> bfd/ChangeLog
> 2021-04-30  Nick Clifton  <nickc@redhat.com>
> 
> 	* libbfd.c (bfd_realloc): Provide some documentation.  Treat a
> 	size of 0 as a size of 1.
> 	(bfd_realloc_or_free): Provide some documentation.  Treat a size
> 	of 0 as a request to free.
> 	* libbfd-in.h (bfd_realloc): Remove prototype.
> 	(bfd_realloc_or_free): Remove prototype.
> 	* libbfd.h: Regenerate.
> 


-- 
  -- Howard Chu
  CTO, Symas Corp.           http://www.symas.com
  Director, Highland Sun     http://highlandsun.com/hyc/
  Chief Architect, OpenLDAP  http://www.openldap.org/project/