This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Attribute returns_null (was: New attribute: returns_nonnull)
- From: David Malcolm <dmalcolm at redhat dot com>
- To: Marc Glisse <marc dot glisse at inria dot fr>
- Cc: gcc-patches at gcc dot gnu dot org
- Date: Mon, 07 Oct 2013 14:37:37 -0400
- Subject: Re: Attribute returns_null (was: New attribute: returns_nonnull)
- Authentication-results: sourceware.org; auth=none
- References: <alpine dot DEB dot 2 dot 02 dot 1310071600430 dot 22663 at stedding dot saclay dot inria dot fr> <1381166322 dot 2732 dot 27 dot camel at surprise> <alpine dot DEB dot 2 dot 10 dot 1310071940290 dot 4148 at laptop-mg dot saclay dot inria dot fr>
On Mon, 2013-10-07 at 19:51 +0200, Marc Glisse wrote:
> On Mon, 7 Oct 2013, David Malcolm wrote:
>
> > On Mon, 2013-10-07 at 16:17 +0200, Marc Glisse wrote:
> >> Hello,
> >>
> >> this patch adds an attribute to let the compiler know that a function
> >> never returns NULL. I saw some ECF_* flags, but the attribute seems
> >> sufficient. I considered using nonnull(0), but then it would have been
> >> confusing that the version of nonnull without arguments applies only to
> >> parameters and not the return value.
> >
> > I can't comment on the patch itself, but could there also be an
> > attribute "returns_null", for functions that *always* return NULL?
> > This may sound weird, but I know of at least one API that exposes such
> > functions: CPython's exception-handling API: see e.g.
> > http://docs.python.org/2/c-api/exceptions.html#PyErr_NoMemory
> > and various other functions that have "Return value: Always NULL."
> > This allows the user to write one-liners like:
> >
> > return PyErr_NoMemory();
>
> I didn't think about it very long, so I probably missed the best reasons,
> but it doesn't sound like such a good idea. If PyErr_NoMemory always
> returns NULL, why not make that clear in the code? It could be an inline
> function, or even a macro that expands to
> (PyErr_SetNone(PyExc_MemoryError),NULL).
>
> To me, attributes are there for when the language is insufficient, kind of
> a last resort. Could you explain why you think it would be the best option
> here?
No, I can't :) (that API predates my involvement in that project).