Re: gEDA-dev: PCB MyFree

[ldoolitt-at-NOSPAMrecycle_dot_lbl_dot_gov]

DJ -

On Thu, Jun 22, 2006 at 03:56:12PM -0400, DJ Delorie wrote:
> 
> > void *xcalloc_function(const char *file, int line, size_t nmemb, size_t size)
> > {
> > 	void *rv;
> > #ifdef __DMALLOC_H__
> > 	rv = dmalloc_malloc(file, line, (nmemb)*(size), DMALLOC_FUNC_CALLOC, 0, 0);
> > #else
> > 	rv = malloc((nmemb)*(size));
> 
> xcalloc should call calloc, not malloc.

Oops, you're right.  I cribbed from the dmalloc_malloc() line,
that cheats and adds the DMALLOC_FUNC_CALLOC.

> > This has to be duplicated
> > four times (calloc, malloc, realloc, strdup).
> 
> Five.  xfree().

Right again.

> > #if DMALLOC_VERSION_MAJOR != 5 || DMALLOC_VERSION_MINOR != 4
> > #warning this code untested against this dmalloc version
> > #endif
> 
> Or use those defines to add ot the DMALLOC_H conditional, so that for
> older versions (what about dmalloc 6? ;) we just call malloc() and let
> the dmalloc.h #define it however it wants.

.. and then the line numbers dmalloc reports are useless.
Fine, I can live with that.  A warning still might be good:

#warning dmalloc version unknown, will work but record useless line numbers

Sounds like a plan.

  - Larry
_______________________________________________
geda-dev mailing list
geda-dev@moria.seul.org
http://www.seul.org/cgi-bin/mailman/listinfo/geda-dev