Re: [PATCH 02/28] glib-compat: Introduce g_memdup2() wrapper

[Vladimir Sementsov-Ogievskiy]

03.09.2021 14:56, Daniel P. Berrangé wrote:
> On Fri, Sep 03, 2021 at 02:51:21PM +0300, Vladimir Sementsov-Ogievskiy wrote:
> > That was most probably already discussed, so sorry my question:
> >
> > Why we can't just do
> >
> > #if ! GLIB_CHECK_VERSION(2, 68, 0)
> > static inline gpointer g_memdup2(gconstpointer mem, gsize byte_size)
> > {
> >      gpointer new_mem;
> >
> >      if (mem && byte_size != 0) {
> >          new_mem = g_malloc(byte_size);
> >          memcpy(new_mem, mem, byte_size);
> >      } else {
> >          new_mem = NULL;
> >      }
> >
> >      return new_mem;
> > }
> > #endif
> >
> > ?
>
> This doesn't play with GLIB_VERSION_MAX_ALLOWED - any use of
> g_memdup2 will trigger compile warnings since we're using an
> API that only exists in a glib version newer than our declared
> baseline.
>
> The inline wrapper  + macro is a trick that lets us backport
> new features, while avoiding the compile warnings.

Ah right, with macro, compiler doesn't see g_memdup2() invocations in code. 
Thanks!

> This is documented in the include/glib-compat.h file that Philippe
> is modifying.
>
> Regards,
> Daniel


--
Best regards,
Vladimir