On Mon, Jun 17, 2019 at 05:48:23PM +0200, Jakub Jelinek wrote: > On Sun, Jun 16, 2019 at 12:36:58PM -0400, Marek Polacek wrote: > > On Sun, Jun 16, 2019 at 06:18:56PM +0200, Jakub Jelinek wrote: > > > On Sun, Jun 16, 2019 at 12:10:37PM -0400, Marek Polacek wrote: > > > > > Guess I will really have to make the changes to treat [[noreturn]] > > > > > similarly > > > > > to e.g. [[nodiscard]], so that cxx11_attribute_p works. > > > > > > > > Thus. Changes I've made: > > > > * don't treat [[noreturn]] as an equivalent to > > > > __attribute__((noreturn)); > > > > * for that I had to adjust decl_attributes, it wasn't preserving the > > > > C++11 form (a list in another list); fix shadowing while at it; > > > > * the above turned up two spots that were wrongly accessing TREE_PURPOSE > > > > directly instead of using get_attribute_name; > > > > * give error only for [[noreturn]] but not for __attribute__((noreturn)) > > > > or [[gnu::noreturn]]. > > > > > > > > Bootstrapped/regtested on x86_64-linux, ok for trunk? > > > > > > I'd prefer to defer review of this to Jason, just want to note that I > > > don't > > > see any testsuite coverage on mixing declarations with different forms of > > > attributes ([[noreturn]] on one decl and __attribute__((noreturn)) or > > > [[gnu::noreturn]] on another one or vice versa. > > > > Added now. I suppose it should compile fine, which it does. > > I meant also the tests of the new diagnostics, say if you have > a decl without any of those attributes, then gnu:: one (or __attribute__ > one; that merge decls should be ok) and on third decl [[noreturn]] (shall > that diagnose anything or not? As there is no way to differentiate it from > the gnu:: attribute on the very first one, I'd say it shouldn't, with the > use of the gnu:: or __attribute__ we are already outside of the standard.
I've added noreturn-11.C for that. With my patch we don't diagnose void f1 (); void f1 [[gnu::noreturn]] (); void f1 [[noreturn]] (); but that seems fine to me, too. Of course there's the problem that we only check the previous decl, not the first one, but I guess we'll have to live with it; it will detect the bogus cases anyway. -- Marek Polacek • Red Hat, Inc. • 300 A St, Boston, MA
