| Message ID | CAFk2RUbzGttMNp8br5nk3vox4F7=UpkUpnsL_rPQrYb3M9JXzA@mail.gmail.com |
|---|---|
| State | New |
| Headers | show |
| Series | c++: Fix access checking for __is_assignable and __is_constructible (c++/94197) | expand |
On Mon, 16 Mar 2020 at 23:25, Ville Voutilainen <ville.voutilainen@gmail.com> wrote: > > Tested on Linux-PPC64. > > This ain't no regression. But it seems to hamper attempts to fix library > regressions (see https://gcc.gnu.org/bugzilla/show_bug.cgi?id=94033). It occurred to me that this can be done in one place. 2020-03-17 Ville Voutilainen <ville.voutilainen@gmail.com> gcc/ PR c++/94197 * cp/method.c (assignable_expr): Use cp_unevaluated. (is_xible_helper): Push a non-deferred access check for the stub objects created by assignable_expr and constructible_expr. testsuite/ PR c++/94197 * g++.dg/ext/pr94197.C: New.
On 3/16/20 6:13 PM, Ville Voutilainen wrote: > On Mon, 16 Mar 2020 at 23:25, Ville Voutilainen > <ville.voutilainen@gmail.com> wrote: >> >> Tested on Linux-PPC64. >> >> This ain't no regression. But it seems to hamper attempts to fix library >> regressions (see https://gcc.gnu.org/bugzilla/show_bug.cgi?id=94033). > > It occurred to me that this can be done in one place. > > 2020-03-17 Ville Voutilainen <ville.voutilainen@gmail.com> > > gcc/ > > PR c++/94197 > * cp/method.c (assignable_expr): Use cp_unevaluated. > (is_xible_helper): Push a non-deferred access check for > the stub objects created by assignable_expr and constructible_expr. It occurs to me that we don't need to defer access control in instantiate_template_1 if we're instantiating an alias template, but this patch is OK as is. Jason
Shouldn't the test use { dg-do compile { target c++11 } } instead of:
+// { dg-do compile }
+// { dg-options "-std=c++11" }
?
On 17/03/20 13:02 +0000, Jonathan Wakely wrote: >Shouldn't the test use { dg-do compile { target c++11 } } instead of: > >+// { dg-do compile } >+// { dg-options "-std=c++11" } > >? > With that change I see: UNSUPPORTED: g++.dg/ext/pr94197.C -std=c++98 PASS: g++.dg/ext/pr94197.C -std=c++14 (test for excess errors) PASS: g++.dg/ext/pr94197.C -std=c++17 (test for excess errors) PASS: g++.dg/ext/pr94197.C -std=c++2a (test for excess errors) rather than: PASS: g++.dg/ext/pr94197.C (test for excess errors)
On 3/17/20 9:04 AM, Jonathan Wakely wrote: > On 17/03/20 13:02 +0000, Jonathan Wakely wrote: >> Shouldn't the test use { dg-do compile { target c++11 } } instead of: >> >> +// { dg-do compile } >> +// { dg-options "-std=c++11" } >> >> ? >> > > With that change I see: > > UNSUPPORTED: g++.dg/ext/pr94197.C -std=c++98 > PASS: g++.dg/ext/pr94197.C -std=c++14 (test for excess errors) > PASS: g++.dg/ext/pr94197.C -std=c++17 (test for excess errors) > PASS: g++.dg/ext/pr94197.C -std=c++2a (test for excess errors) > > rather than: > > PASS: g++.dg/ext/pr94197.C (test for excess errors) Yes, good point. Jason
On Tue, 17 Mar 2020 at 16:42, Jason Merrill <jason@redhat.com> wrote: > > On 3/17/20 9:04 AM, Jonathan Wakely wrote: > > On 17/03/20 13:02 +0000, Jonathan Wakely wrote: > >> Shouldn't the test use { dg-do compile { target c++11 } } instead of: > >> > >> +// { dg-do compile } > >> +// { dg-options "-std=c++11" } > > Yes, good point. Ack, changing to { dg-do compile { target c++11 } } and committing with the (correct :P) expectation that the change is still ok with that change.
On Tue, 17 Mar 2020 at 16:52, Ville Voutilainen <ville.voutilainen@gmail.com> wrote: > > On Tue, 17 Mar 2020 at 16:42, Jason Merrill <jason@redhat.com> wrote: > > > > On 3/17/20 9:04 AM, Jonathan Wakely wrote: > > > On 17/03/20 13:02 +0000, Jonathan Wakely wrote: > > >> Shouldn't the test use { dg-do compile { target c++11 } } instead of: > > >> > > >> +// { dg-do compile } > > >> +// { dg-options "-std=c++11" } > > > > Yes, good point. > > Ack, changing to { dg-do compile { target c++11 } } and committing > with the (correct :P) expectation that > the change is still ok with that change. Jason, are backports ok too?
On 3/17/20 11:25 AM, Ville Voutilainen wrote: > On Tue, 17 Mar 2020 at 16:52, Ville Voutilainen > <ville.voutilainen@gmail.com> wrote: >> >> On Tue, 17 Mar 2020 at 16:42, Jason Merrill <jason@redhat.com> wrote: >>> >>> On 3/17/20 9:04 AM, Jonathan Wakely wrote: >>>> On 17/03/20 13:02 +0000, Jonathan Wakely wrote: >>>>> Shouldn't the test use { dg-do compile { target c++11 } } instead of: >>>>> >>>>> +// { dg-do compile } >>>>> +// { dg-options "-std=c++11" } >>> >>> Yes, good point. >> >> Ack, changing to { dg-do compile { target c++11 } } and committing >> with the (correct :P) expectation that >> the change is still ok with that change. > > Jason, are backports ok too? Yes. Jason
diff --git a/gcc/cp/method.c b/gcc/cp/method.c index 790d5704092..b429ea48049 100644 --- a/gcc/cp/method.c +++ b/gcc/cp/method.c @@ -1739,11 +1739,11 @@ check_nontriv (tree *tp, int *, void *) static tree assignable_expr (tree to, tree from) { - ++cp_unevaluated_operand; + cp_unevaluated cp_uneval_guard; + deferring_access_check_sentinel acs (dk_no_deferred); to = build_stub_object (to); from = build_stub_object (from); tree r = cp_build_modify_expr (input_location, to, NOP_EXPR, from, tf_none); - --cp_unevaluated_operand; return r; } @@ -1759,6 +1759,7 @@ constructible_expr (tree to, tree from) { tree expr; cp_unevaluated cp_uneval_guard; + deferring_access_check_sentinel acs (dk_no_deferred); if (CLASS_TYPE_P (to)) { tree ctype = to; diff --git a/gcc/testsuite/g++.dg/ext/pr94197.C b/gcc/testsuite/g++.dg/ext/pr94197.C new file mode 100644 index 00000000000..768bfbac0a6 --- /dev/null +++ b/gcc/testsuite/g++.dg/ext/pr94197.C @@ -0,0 +1,75 @@ +// { dg-do compile } +// { dg-options "-std=c++11" } + +template<typename T> + T&& declval() noexcept; + +template<bool B> +struct bool_constant +{ + static constexpr bool value = B; + using type = bool_constant; +}; + +using true_type = bool_constant<true>; +using false_type = bool_constant<false>; + +template<bool, typename T, typename Arg> + struct __is_nt_constructible_impl + : public false_type + { }; + +template<typename T, typename Arg> + struct __is_nt_constructible_impl<true, T, Arg> + : public bool_constant<noexcept(static_cast<T>(declval<Arg>()))> + { }; + +template<typename T, typename Arg> + using __is_nothrow_constructible_impl + = __is_nt_constructible_impl<__is_constructible(T, Arg), T, Arg>; + +template<typename T> + struct __is_nothrow_copy_constructible_impl + : public __is_nothrow_constructible_impl<T, const T&> + { }; + +template<typename T> + struct is_nothrow_copy_constructible + : public __is_nothrow_copy_constructible_impl<T>::type + { }; + +template<bool, typename T, typename Arg> + struct __is_nt_assignable_impl + : public false_type + { }; + +template<typename T, typename Arg> + struct __is_nt_assignable_impl<true, T, Arg> + : public bool_constant<noexcept(declval<T&>() = declval<Arg>())> + { }; + +template<typename T, typename Arg> + using __is_nothrow_assignable_impl + = __is_nt_assignable_impl<__is_assignable(T, Arg), T, Arg>; + +template<typename T> + struct __is_nothrow_copy_assignable_impl + : public __is_nothrow_assignable_impl<T, const T&> + { }; + +template<typename T> + struct is_nothrow_copy_assignable + : public __is_nothrow_copy_assignable_impl<T>::type + { }; + +struct NType +{ + NType(); +private: + NType(const NType&); + NType& operator=(const NType&); +}; + + +static_assert( !is_nothrow_copy_constructible<NType>::value, "" ); +static_assert( !is_nothrow_copy_assignable<NType>::value, "" );
Tested on Linux-PPC64. This ain't no regression. But it seems to hamper attempts to fix library regressions (see https://gcc.gnu.org/bugzilla/show_bug.cgi?id=94033). 2020-03-16 Ville Voutilainen <ville.voutilainen@gmail.com> gcc/ PR c++/94197 * cp/method.c (assignable_expr, constructible_expr): Push a deferred access check for the stub object. testsuite/ PR c++/94197 * g++.dg/ext/pr94197.C: New.