| Message ID | 20230306235957.390533-1-polacek@redhat.com |
|---|---|
| State | New |
| Headers | show |
| Series | c++: noexcept and copy elision [PR109030] | expand |
On 3/6/23 18:59, Marek Polacek wrote: > When processing a noexcept, constructors aren't elided: build_over_call > has > /* It's unsafe to elide the constructor when handling > a noexcept-expression, it may evaluate to the wrong > value (c++/53025). */ > && (force_elide || cp_noexcept_operand == 0)) > so the assert I added recently needs to be relaxed a little bit. > > Bootstrapped/regtested on x86_64-pc-linux-gnu, ok for trunk? OK. > PR c++/109030 > > gcc/cp/ChangeLog: > > * constexpr.cc (cxx_eval_call_expression): Relax assert. > > gcc/testsuite/ChangeLog: > > * g++.dg/cpp0x/noexcept77.C: New test. > --- > gcc/cp/constexpr.cc | 6 +++++- > gcc/testsuite/g++.dg/cpp0x/noexcept77.C | 9 +++++++++ > 2 files changed, 14 insertions(+), 1 deletion(-) > create mode 100644 gcc/testsuite/g++.dg/cpp0x/noexcept77.C > > diff --git a/gcc/cp/constexpr.cc b/gcc/cp/constexpr.cc > index 364695b762c..5384d0e8e46 100644 > --- a/gcc/cp/constexpr.cc > +++ b/gcc/cp/constexpr.cc > @@ -2869,7 +2869,11 @@ cxx_eval_call_expression (const constexpr_ctx *ctx, tree t, > > /* We used to shortcut trivial constructor/op= here, but nowadays > we can only get a trivial function here with -fno-elide-constructors. */ > - gcc_checking_assert (!trivial_fn_p (fun) || !flag_elide_constructors); > + gcc_checking_assert (!trivial_fn_p (fun) > + || !flag_elide_constructors > + /* We don't elide constructors when processing > + a noexcept-expression. */ > + || cp_noexcept_operand); > > bool non_constant_args = false; > new_call.bindings > diff --git a/gcc/testsuite/g++.dg/cpp0x/noexcept77.C b/gcc/testsuite/g++.dg/cpp0x/noexcept77.C > new file mode 100644 > index 00000000000..16db8eb79ee > --- /dev/null > +++ b/gcc/testsuite/g++.dg/cpp0x/noexcept77.C > @@ -0,0 +1,9 @@ > +// PR c++/109030 > +// { dg-do compile { target c++11 } } > + > +struct foo { }; > + > +struct __as_receiver { > + foo empty_env; > +}; > +void sched(foo __fun) noexcept(noexcept(__as_receiver{__fun})) { } > > base-commit: dfb14cdd796ad9df6b5f2def047ef36b29385902
On Mon, 6 Mar 2023, Marek Polacek via Gcc-patches wrote: > When processing a noexcept, constructors aren't elided: build_over_call > has > /* It's unsafe to elide the constructor when handling > a noexcept-expression, it may evaluate to the wrong > value (c++/53025). */ > && (force_elide || cp_noexcept_operand == 0)) > so the assert I added recently needs to be relaxed a little bit. > > Bootstrapped/regtested on x86_64-pc-linux-gnu, ok for trunk? > > PR c++/109030 > > gcc/cp/ChangeLog: > > * constexpr.cc (cxx_eval_call_expression): Relax assert. > > gcc/testsuite/ChangeLog: > > * g++.dg/cpp0x/noexcept77.C: New test. > --- > gcc/cp/constexpr.cc | 6 +++++- > gcc/testsuite/g++.dg/cpp0x/noexcept77.C | 9 +++++++++ > 2 files changed, 14 insertions(+), 1 deletion(-) > create mode 100644 gcc/testsuite/g++.dg/cpp0x/noexcept77.C > > diff --git a/gcc/cp/constexpr.cc b/gcc/cp/constexpr.cc > index 364695b762c..5384d0e8e46 100644 > --- a/gcc/cp/constexpr.cc > +++ b/gcc/cp/constexpr.cc > @@ -2869,7 +2869,11 @@ cxx_eval_call_expression (const constexpr_ctx *ctx, tree t, > > /* We used to shortcut trivial constructor/op= here, but nowadays > we can only get a trivial function here with -fno-elide-constructors. */ > - gcc_checking_assert (!trivial_fn_p (fun) || !flag_elide_constructors); > + gcc_checking_assert (!trivial_fn_p (fun) > + || !flag_elide_constructors > + /* We don't elide constructors when processing > + a noexcept-expression. */ > + || cp_noexcept_operand); It seems weird that we're performing constant evaluation within an unevaluated operand. Would it make sense to also fix this a second way by avoiding constant evaluation from maybe_constant_init when cp_unevaluated_operand && !manifestly_const_eval, like in maybe_constant_value? IIUC since we could still have an evaluated subexpression withis noexcept, the two fixes would be complementary. > > bool non_constant_args = false; > new_call.bindings > diff --git a/gcc/testsuite/g++.dg/cpp0x/noexcept77.C b/gcc/testsuite/g++.dg/cpp0x/noexcept77.C > new file mode 100644 > index 00000000000..16db8eb79ee > --- /dev/null > +++ b/gcc/testsuite/g++.dg/cpp0x/noexcept77.C > @@ -0,0 +1,9 @@ > +// PR c++/109030 > +// { dg-do compile { target c++11 } } > + > +struct foo { }; > + > +struct __as_receiver { > + foo empty_env; > +}; > +void sched(foo __fun) noexcept(noexcept(__as_receiver{__fun})) { } > > base-commit: dfb14cdd796ad9df6b5f2def047ef36b29385902 > -- > 2.39.2 > >
On 3/9/23 14:32, Patrick Palka wrote: > On Mon, 6 Mar 2023, Marek Polacek via Gcc-patches wrote: > >> When processing a noexcept, constructors aren't elided: build_over_call >> has >> /* It's unsafe to elide the constructor when handling >> a noexcept-expression, it may evaluate to the wrong >> value (c++/53025). */ >> && (force_elide || cp_noexcept_operand == 0)) >> so the assert I added recently needs to be relaxed a little bit. >> >> Bootstrapped/regtested on x86_64-pc-linux-gnu, ok for trunk? >> >> PR c++/109030 >> >> gcc/cp/ChangeLog: >> >> * constexpr.cc (cxx_eval_call_expression): Relax assert. >> >> gcc/testsuite/ChangeLog: >> >> * g++.dg/cpp0x/noexcept77.C: New test. >> --- >> gcc/cp/constexpr.cc | 6 +++++- >> gcc/testsuite/g++.dg/cpp0x/noexcept77.C | 9 +++++++++ >> 2 files changed, 14 insertions(+), 1 deletion(-) >> create mode 100644 gcc/testsuite/g++.dg/cpp0x/noexcept77.C >> >> diff --git a/gcc/cp/constexpr.cc b/gcc/cp/constexpr.cc >> index 364695b762c..5384d0e8e46 100644 >> --- a/gcc/cp/constexpr.cc >> +++ b/gcc/cp/constexpr.cc >> @@ -2869,7 +2869,11 @@ cxx_eval_call_expression (const constexpr_ctx *ctx, tree t, >> >> /* We used to shortcut trivial constructor/op= here, but nowadays >> we can only get a trivial function here with -fno-elide-constructors. */ >> - gcc_checking_assert (!trivial_fn_p (fun) || !flag_elide_constructors); >> + gcc_checking_assert (!trivial_fn_p (fun) >> + || !flag_elide_constructors >> + /* We don't elide constructors when processing >> + a noexcept-expression. */ >> + || cp_noexcept_operand); > > It seems weird that we're performing constant evaluation within an > unevaluated operand. Would it make sense to also fix this a second way > by avoiding constant evaluation from maybe_constant_init when > cp_unevaluated_operand && !manifestly_const_eval, like in maybe_constant_value? Sounds good. > IIUC since we could still have an evaluated subexpression withis > noexcept, the two fixes would be complementary. > >> >> bool non_constant_args = false; >> new_call.bindings >> diff --git a/gcc/testsuite/g++.dg/cpp0x/noexcept77.C b/gcc/testsuite/g++.dg/cpp0x/noexcept77.C >> new file mode 100644 >> index 00000000000..16db8eb79ee >> --- /dev/null >> +++ b/gcc/testsuite/g++.dg/cpp0x/noexcept77.C >> @@ -0,0 +1,9 @@ >> +// PR c++/109030 >> +// { dg-do compile { target c++11 } } >> + >> +struct foo { }; >> + >> +struct __as_receiver { >> + foo empty_env; >> +}; >> +void sched(foo __fun) noexcept(noexcept(__as_receiver{__fun})) { } >> >> base-commit: dfb14cdd796ad9df6b5f2def047ef36b29385902 >> -- >> 2.39.2 >> >> >
On Thu, 9 Mar 2023, Jason Merrill wrote: > On 3/9/23 14:32, Patrick Palka wrote: > > On Mon, 6 Mar 2023, Marek Polacek via Gcc-patches wrote: > > > > > When processing a noexcept, constructors aren't elided: build_over_call > > > has > > > /* It's unsafe to elide the constructor when handling > > > a noexcept-expression, it may evaluate to the wrong > > > value (c++/53025). */ > > > && (force_elide || cp_noexcept_operand == 0)) > > > so the assert I added recently needs to be relaxed a little bit. > > > > > > Bootstrapped/regtested on x86_64-pc-linux-gnu, ok for trunk? > > > > > > PR c++/109030 > > > > > > gcc/cp/ChangeLog: > > > > > > * constexpr.cc (cxx_eval_call_expression): Relax assert. > > > > > > gcc/testsuite/ChangeLog: > > > > > > * g++.dg/cpp0x/noexcept77.C: New test. > > > --- > > > gcc/cp/constexpr.cc | 6 +++++- > > > gcc/testsuite/g++.dg/cpp0x/noexcept77.C | 9 +++++++++ > > > 2 files changed, 14 insertions(+), 1 deletion(-) > > > create mode 100644 gcc/testsuite/g++.dg/cpp0x/noexcept77.C > > > > > > diff --git a/gcc/cp/constexpr.cc b/gcc/cp/constexpr.cc > > > index 364695b762c..5384d0e8e46 100644 > > > --- a/gcc/cp/constexpr.cc > > > +++ b/gcc/cp/constexpr.cc > > > @@ -2869,7 +2869,11 @@ cxx_eval_call_expression (const constexpr_ctx *ctx, > > > tree t, > > > /* We used to shortcut trivial constructor/op= here, but nowadays > > > we can only get a trivial function here with > > > -fno-elide-constructors. */ > > > - gcc_checking_assert (!trivial_fn_p (fun) || !flag_elide_constructors); > > > + gcc_checking_assert (!trivial_fn_p (fun) > > > + || !flag_elide_constructors > > > + /* We don't elide constructors when processing > > > + a noexcept-expression. */ > > > + || cp_noexcept_operand); > > > > It seems weird that we're performing constant evaluation within an > > unevaluated operand. Would it make sense to also fix this a second way > > by avoiding constant evaluation from maybe_constant_init when > > cp_unevaluated_operand && !manifestly_const_eval, like in > > maybe_constant_value? > > Sounds good. Hmm, while working on this I noticed we currently don't reject a version of g++.dg/cpp2a/constexpr-inst1.C that list initializes an aggregate instead of int (ever since r12-4425-g1595fe44e11a96): struct A { int m; }; template<typename T> constexpr int f() { return T::value; } template<bool B, typename T> void h(decltype(A{B ? f<T>() : 0})); // was int{...} template<bool B, typename T> void h(...); void x() { h<false, int>(0); // OK? } ISTM we should instantiate f<int> here for the same reason we do in the original version of the testcase, and for that to happen we need to pass manifestly_const_eval=true in massage_init_elt. Does that seem reasonable?
On Wed, 15 Mar 2023, Patrick Palka wrote: > On Thu, 9 Mar 2023, Jason Merrill wrote: > > > On 3/9/23 14:32, Patrick Palka wrote: > > > On Mon, 6 Mar 2023, Marek Polacek via Gcc-patches wrote: > > > > > > > When processing a noexcept, constructors aren't elided: build_over_call > > > > has > > > > /* It's unsafe to elide the constructor when handling > > > > a noexcept-expression, it may evaluate to the wrong > > > > value (c++/53025). */ > > > > && (force_elide || cp_noexcept_operand == 0)) > > > > so the assert I added recently needs to be relaxed a little bit. > > > > > > > > Bootstrapped/regtested on x86_64-pc-linux-gnu, ok for trunk? > > > > > > > > PR c++/109030 > > > > > > > > gcc/cp/ChangeLog: > > > > > > > > * constexpr.cc (cxx_eval_call_expression): Relax assert. > > > > > > > > gcc/testsuite/ChangeLog: > > > > > > > > * g++.dg/cpp0x/noexcept77.C: New test. > > > > --- > > > > gcc/cp/constexpr.cc | 6 +++++- > > > > gcc/testsuite/g++.dg/cpp0x/noexcept77.C | 9 +++++++++ > > > > 2 files changed, 14 insertions(+), 1 deletion(-) > > > > create mode 100644 gcc/testsuite/g++.dg/cpp0x/noexcept77.C > > > > > > > > diff --git a/gcc/cp/constexpr.cc b/gcc/cp/constexpr.cc > > > > index 364695b762c..5384d0e8e46 100644 > > > > --- a/gcc/cp/constexpr.cc > > > > +++ b/gcc/cp/constexpr.cc > > > > @@ -2869,7 +2869,11 @@ cxx_eval_call_expression (const constexpr_ctx *ctx, > > > > tree t, > > > > /* We used to shortcut trivial constructor/op= here, but nowadays > > > > we can only get a trivial function here with > > > > -fno-elide-constructors. */ > > > > - gcc_checking_assert (!trivial_fn_p (fun) || !flag_elide_constructors); > > > > + gcc_checking_assert (!trivial_fn_p (fun) > > > > + || !flag_elide_constructors > > > > + /* We don't elide constructors when processing > > > > + a noexcept-expression. */ > > > > + || cp_noexcept_operand); > > > > > > It seems weird that we're performing constant evaluation within an > > > unevaluated operand. Would it make sense to also fix this a second way > > > by avoiding constant evaluation from maybe_constant_init when > > > cp_unevaluated_operand && !manifestly_const_eval, like in > > > maybe_constant_value? > > > > Sounds good. > > Hmm, while working on this I noticed we currently don't reject a version of > g++.dg/cpp2a/constexpr-inst1.C that list initializes an aggregate instead of > int (ever since r12-4425-g1595fe44e11a96): > > struct A { int m; }; > template<typename T> constexpr int f() { return T::value; } > template<bool B, typename T> void h(decltype(A{B ? f<T>() : 0})); // was int{...} > template<bool B, typename T> void h(...); > void x() { > h<false, int>(0); // OK? > } > > ISTM we should instantiate f<int> here for the same reason we do in the > original version of the testcase, and for that to happen we need to > pass manifestly_const_eval=true in massage_init_elt. Does that seem > reasonable? > FWIW the reason this came up is because I tried contriving a testcase for the aforementioned maybe_constant_init change, and I came up with: struct __as_receiver { int empty_env; }; template<class T> constexpr int f(T t) { return t.fail; }; using type = decltype(__as_receiver{f(0)}); // OK, f<int> no longer instantiated which we used to reject and afterwards accept. But since the elements of an initializer list are potentially constant evaluated, I wonder if that that means f<int> should be instantiated here after all despite the unevaluated context? Here's the full patch for reference: -- >8 -- Subject: [PATCH] c++: maybe_constant_init and unevaluated operands [PR109030] This testcase in this PR (already fixed by r13-6526-ge4692319fd5fc7) illustrates that maybe_constant_init can be called on an unevaluated operand (from massage_init_elt), so this entry point should limit constant evaluation in that case, like maybe_constant_value does. PR c++/109030 gcc/cp/ChangeLog: * constexpr.cc (maybe_constant_init_1): For an unevaluated non-manifestly-constant operand, don't constant evaluate and instead call fold_to_constant. gcc/testsuite/ChangeLog: * g++.dg/cpp0x/decltype83.C: New test. --- gcc/cp/constexpr.cc | 2 ++ gcc/testsuite/g++.dg/cpp0x/decltype83.C | 14 ++++++++++++++ 2 files changed, 16 insertions(+) create mode 100644 gcc/testsuite/g++.dg/cpp0x/decltype83.C diff --git a/gcc/cp/constexpr.cc b/gcc/cp/constexpr.cc index 8683c00596a..f325af375c8 100644 --- a/gcc/cp/constexpr.cc +++ b/gcc/cp/constexpr.cc @@ -8795,6 +8795,8 @@ maybe_constant_init_1 (tree t, tree decl, bool allow_non_constant, && (TREE_STATIC (decl) || DECL_EXTERNAL (decl))); if (is_static) manifestly_const_eval = true; + if (cp_unevaluated_operand && !manifestly_const_eval) + return fold_to_constant (t); t = cxx_eval_outermost_constant_expr (t, allow_non_constant, !is_static, mce_value (manifestly_const_eval), false, decl); diff --git a/gcc/testsuite/g++.dg/cpp0x/decltype83.C b/gcc/testsuite/g++.dg/cpp0x/decltype83.C new file mode 100644 index 00000000000..17005a92eb5 --- /dev/null +++ b/gcc/testsuite/g++.dg/cpp0x/decltype83.C @@ -0,0 +1,14 @@ +// { dg-do compile { target c++11 } } + +struct __as_receiver { + int empty_env; +}; + +template<class T> +constexpr int f(T t) { + return t.fail; +}; + +int main() { + using type = decltype(__as_receiver{f(0)}); // OK, f<int> not instantiated +}
On 3/16/23 10:09, Patrick Palka wrote: > On Wed, 15 Mar 2023, Patrick Palka wrote: > >> On Thu, 9 Mar 2023, Jason Merrill wrote: >> >>> On 3/9/23 14:32, Patrick Palka wrote: >>>> On Mon, 6 Mar 2023, Marek Polacek via Gcc-patches wrote: >>>> >>>>> When processing a noexcept, constructors aren't elided: build_over_call >>>>> has >>>>> /* It's unsafe to elide the constructor when handling >>>>> a noexcept-expression, it may evaluate to the wrong >>>>> value (c++/53025). */ >>>>> && (force_elide || cp_noexcept_operand == 0)) >>>>> so the assert I added recently needs to be relaxed a little bit. >>>>> >>>>> Bootstrapped/regtested on x86_64-pc-linux-gnu, ok for trunk? >>>>> >>>>> PR c++/109030 >>>>> >>>>> gcc/cp/ChangeLog: >>>>> >>>>> * constexpr.cc (cxx_eval_call_expression): Relax assert. >>>>> >>>>> gcc/testsuite/ChangeLog: >>>>> >>>>> * g++.dg/cpp0x/noexcept77.C: New test. >>>>> --- >>>>> gcc/cp/constexpr.cc | 6 +++++- >>>>> gcc/testsuite/g++.dg/cpp0x/noexcept77.C | 9 +++++++++ >>>>> 2 files changed, 14 insertions(+), 1 deletion(-) >>>>> create mode 100644 gcc/testsuite/g++.dg/cpp0x/noexcept77.C >>>>> >>>>> diff --git a/gcc/cp/constexpr.cc b/gcc/cp/constexpr.cc >>>>> index 364695b762c..5384d0e8e46 100644 >>>>> --- a/gcc/cp/constexpr.cc >>>>> +++ b/gcc/cp/constexpr.cc >>>>> @@ -2869,7 +2869,11 @@ cxx_eval_call_expression (const constexpr_ctx *ctx, >>>>> tree t, >>>>> /* We used to shortcut trivial constructor/op= here, but nowadays >>>>> we can only get a trivial function here with >>>>> -fno-elide-constructors. */ >>>>> - gcc_checking_assert (!trivial_fn_p (fun) || !flag_elide_constructors); >>>>> + gcc_checking_assert (!trivial_fn_p (fun) >>>>> + || !flag_elide_constructors >>>>> + /* We don't elide constructors when processing >>>>> + a noexcept-expression. */ >>>>> + || cp_noexcept_operand); >>>> >>>> It seems weird that we're performing constant evaluation within an >>>> unevaluated operand. Would it make sense to also fix this a second way >>>> by avoiding constant evaluation from maybe_constant_init when >>>> cp_unevaluated_operand && !manifestly_const_eval, like in >>>> maybe_constant_value? >>> >>> Sounds good. >> >> Hmm, while working on this I noticed we currently don't reject a version of >> g++.dg/cpp2a/constexpr-inst1.C that list initializes an aggregate instead of >> int (ever since r12-4425-g1595fe44e11a96): >> >> struct A { int m; }; >> template<typename T> constexpr int f() { return T::value; } >> template<bool B, typename T> void h(decltype(A{B ? f<T>() : 0})); // was int{...} >> template<bool B, typename T> void h(...); >> void x() { >> h<false, int>(0); // OK? >> } >> >> ISTM we should instantiate f<int> here for the same reason we do in the >> original version of the testcase, and for that to happen we need to >> pass manifestly_const_eval=true in massage_init_elt. Does that seem >> reasonable? >> > > FWIW the reason this came up is because I tried contriving a testcase > for the aforementioned maybe_constant_init change, and I came up with: > > struct __as_receiver { > int empty_env; > }; > > template<class T> > constexpr int f(T t) { > return t.fail; > }; > > using type = decltype(__as_receiver{f(0)}); // OK, f<int> no longer instantiated > > which we used to reject and afterwards accept. But since the elements > of an initializer list are potentially constant evaluated, I wonder if > that that means f<int> should be instantiated here after all despite the > unevaluated context? The relevant section of the standard would seem to be https://eel.is/c++draft/expr.const#20 ; an immediate subexpression of a braced-init-list is potentially constant-evaluated even though it isn't potentially-evaluated or manifestly constant-evaluated. It seems like the call to fold_non_dependent_expr in check_narrowing ought to cause instantiation in this case, why doesn't it? > Here's the full patch for reference: > > -- >8 -- > > Subject: [PATCH] c++: maybe_constant_init and unevaluated operands [PR109030] > > This testcase in this PR (already fixed by r13-6526-ge4692319fd5fc7) > illustrates that maybe_constant_init can be called on an unevaluated > operand (from massage_init_elt), so this entry point should limit > constant evaluation in that case, like maybe_constant_value does. > > PR c++/109030 > > gcc/cp/ChangeLog: > > * constexpr.cc (maybe_constant_init_1): For an unevaluated > non-manifestly-constant operand, don't constant evaluate > and instead call fold_to_constant. > > gcc/testsuite/ChangeLog: > > * g++.dg/cpp0x/decltype83.C: New test. > --- > gcc/cp/constexpr.cc | 2 ++ > gcc/testsuite/g++.dg/cpp0x/decltype83.C | 14 ++++++++++++++ > 2 files changed, 16 insertions(+) > create mode 100644 gcc/testsuite/g++.dg/cpp0x/decltype83.C > > diff --git a/gcc/cp/constexpr.cc b/gcc/cp/constexpr.cc > index 8683c00596a..f325af375c8 100644 > --- a/gcc/cp/constexpr.cc > +++ b/gcc/cp/constexpr.cc > @@ -8795,6 +8795,8 @@ maybe_constant_init_1 (tree t, tree decl, bool allow_non_constant, > && (TREE_STATIC (decl) || DECL_EXTERNAL (decl))); > if (is_static) > manifestly_const_eval = true; > + if (cp_unevaluated_operand && !manifestly_const_eval) > + return fold_to_constant (t); > t = cxx_eval_outermost_constant_expr (t, allow_non_constant, !is_static, > mce_value (manifestly_const_eval), > false, decl); > diff --git a/gcc/testsuite/g++.dg/cpp0x/decltype83.C b/gcc/testsuite/g++.dg/cpp0x/decltype83.C > new file mode 100644 > index 00000000000..17005a92eb5 > --- /dev/null > +++ b/gcc/testsuite/g++.dg/cpp0x/decltype83.C > @@ -0,0 +1,14 @@ > +// { dg-do compile { target c++11 } } > + > +struct __as_receiver { > + int empty_env; > +}; > + > +template<class T> > +constexpr int f(T t) { > + return t.fail; > +}; > + > +int main() { > + using type = decltype(__as_receiver{f(0)}); // OK, f<int> not instantiated > +}
On Thu, 16 Mar 2023, Jason Merrill wrote: > On 3/16/23 10:09, Patrick Palka wrote: > > On Wed, 15 Mar 2023, Patrick Palka wrote: > > > > > On Thu, 9 Mar 2023, Jason Merrill wrote: > > > > > > > On 3/9/23 14:32, Patrick Palka wrote: > > > > > On Mon, 6 Mar 2023, Marek Polacek via Gcc-patches wrote: > > > > > > > > > > > When processing a noexcept, constructors aren't elided: > > > > > > build_over_call > > > > > > has > > > > > > /* It's unsafe to elide the constructor when handling > > > > > > a noexcept-expression, it may evaluate to the wrong > > > > > > value (c++/53025). */ > > > > > > && (force_elide || cp_noexcept_operand == 0)) > > > > > > so the assert I added recently needs to be relaxed a little bit. > > > > > > > > > > > > Bootstrapped/regtested on x86_64-pc-linux-gnu, ok for trunk? > > > > > > > > > > > > PR c++/109030 > > > > > > > > > > > > gcc/cp/ChangeLog: > > > > > > > > > > > > * constexpr.cc (cxx_eval_call_expression): Relax assert. > > > > > > > > > > > > gcc/testsuite/ChangeLog: > > > > > > > > > > > > * g++.dg/cpp0x/noexcept77.C: New test. > > > > > > --- > > > > > > gcc/cp/constexpr.cc | 6 +++++- > > > > > > gcc/testsuite/g++.dg/cpp0x/noexcept77.C | 9 +++++++++ > > > > > > 2 files changed, 14 insertions(+), 1 deletion(-) > > > > > > create mode 100644 gcc/testsuite/g++.dg/cpp0x/noexcept77.C > > > > > > > > > > > > diff --git a/gcc/cp/constexpr.cc b/gcc/cp/constexpr.cc > > > > > > index 364695b762c..5384d0e8e46 100644 > > > > > > --- a/gcc/cp/constexpr.cc > > > > > > +++ b/gcc/cp/constexpr.cc > > > > > > @@ -2869,7 +2869,11 @@ cxx_eval_call_expression (const constexpr_ctx > > > > > > *ctx, > > > > > > tree t, > > > > > > /* We used to shortcut trivial constructor/op= here, but > > > > > > nowadays > > > > > > we can only get a trivial function here with > > > > > > -fno-elide-constructors. */ > > > > > > - gcc_checking_assert (!trivial_fn_p (fun) || > > > > > > !flag_elide_constructors); > > > > > > + gcc_checking_assert (!trivial_fn_p (fun) > > > > > > + || !flag_elide_constructors > > > > > > + /* We don't elide constructors when processing > > > > > > + a noexcept-expression. */ > > > > > > + || cp_noexcept_operand); > > > > > > > > > > It seems weird that we're performing constant evaluation within an > > > > > unevaluated operand. Would it make sense to also fix this a second > > > > > way > > > > > by avoiding constant evaluation from maybe_constant_init when > > > > > cp_unevaluated_operand && !manifestly_const_eval, like in > > > > > maybe_constant_value? > > > > > > > > Sounds good. > > > > > > Hmm, while working on this I noticed we currently don't reject a version > > > of > > > g++.dg/cpp2a/constexpr-inst1.C that list initializes an aggregate instead > > > of > > > int (ever since r12-4425-g1595fe44e11a96): > > > > > > struct A { int m; }; > > > template<typename T> constexpr int f() { return T::value; } > > > template<bool B, typename T> void h(decltype(A{B ? f<T>() : 0})); // > > > was int{...} > > > template<bool B, typename T> void h(...); > > > void x() { > > > h<false, int>(0); // OK? > > > } > > > > > > ISTM we should instantiate f<int> here for the same reason we do in the > > > original version of the testcase, and for that to happen we need to > > > pass manifestly_const_eval=true in massage_init_elt. Does that seem > > > reasonable? > > > > > > > FWIW the reason this came up is because I tried contriving a testcase > > for the aforementioned maybe_constant_init change, and I came up with: > > > > struct __as_receiver { > > int empty_env; > > }; > > > > template<class T> > > constexpr int f(T t) { > > return t.fail; > > }; > > > > using type = decltype(__as_receiver{f(0)}); // OK, f<int> no longer > > instantiated > > > > which we used to reject and afterwards accept. But since the elements > > of an initializer list are potentially constant evaluated, I wonder if > > that that means f<int> should be instantiated here after all despite the > > unevaluated context? > > The relevant section of the standard would seem to be > https://eel.is/c++draft/expr.const#20 ; an immediate subexpression of a > braced-init-list is potentially constant-evaluated even though it isn't > potentially-evaluated or manifestly constant-evaluated. > > It seems like the call to fold_non_dependent_expr in check_narrowing ought to > cause instantiation in this case, why doesn't it? Looks like check_narrowing isn't called at all in this aggr init case. The call from e.g. convert_like_internal isn't reached because the conversion for the initializer element is ck_identity, and don't ever set conversion::check_narrowing for ck_identity conversions I think. Yet for using 'type = decltype(int{f(0)});' (similar to the example in [temp.inst]/8) we do call check_narrowing directly from finish_compound_literal, despite the conversion effectively being an identity conversion. > > > Here's the full patch for reference: > > > > -- >8 -- > > > > Subject: [PATCH] c++: maybe_constant_init and unevaluated operands > > [PR109030] > > > > This testcase in this PR (already fixed by r13-6526-ge4692319fd5fc7) > > illustrates that maybe_constant_init can be called on an unevaluated > > operand (from massage_init_elt), so this entry point should limit > > constant evaluation in that case, like maybe_constant_value does. > > > > PR c++/109030 > > > > gcc/cp/ChangeLog: > > > > * constexpr.cc (maybe_constant_init_1): For an unevaluated > > non-manifestly-constant operand, don't constant evaluate > > and instead call fold_to_constant. > > > > gcc/testsuite/ChangeLog: > > > > * g++.dg/cpp0x/decltype83.C: New test. > > --- > > gcc/cp/constexpr.cc | 2 ++ > > gcc/testsuite/g++.dg/cpp0x/decltype83.C | 14 ++++++++++++++ > > 2 files changed, 16 insertions(+) > > create mode 100644 gcc/testsuite/g++.dg/cpp0x/decltype83.C > > > > diff --git a/gcc/cp/constexpr.cc b/gcc/cp/constexpr.cc > > index 8683c00596a..f325af375c8 100644 > > --- a/gcc/cp/constexpr.cc > > +++ b/gcc/cp/constexpr.cc > > @@ -8795,6 +8795,8 @@ maybe_constant_init_1 (tree t, tree decl, bool > > allow_non_constant, > > && (TREE_STATIC (decl) || DECL_EXTERNAL (decl))); > > if (is_static) > > manifestly_const_eval = true; > > + if (cp_unevaluated_operand && !manifestly_const_eval) > > + return fold_to_constant (t); > > t = cxx_eval_outermost_constant_expr (t, allow_non_constant, > > !is_static, > > mce_value (manifestly_const_eval), > > false, decl); > > diff --git a/gcc/testsuite/g++.dg/cpp0x/decltype83.C > > b/gcc/testsuite/g++.dg/cpp0x/decltype83.C > > new file mode 100644 > > index 00000000000..17005a92eb5 > > --- /dev/null > > +++ b/gcc/testsuite/g++.dg/cpp0x/decltype83.C > > @@ -0,0 +1,14 @@ > > +// { dg-do compile { target c++11 } } > > + > > +struct __as_receiver { > > + int empty_env; > > +}; > > + > > +template<class T> > > +constexpr int f(T t) { > > + return t.fail; > > +}; > > + > > +int main() { > > + using type = decltype(__as_receiver{f(0)}); // OK, f<int> not > > instantiated > > +} > >
On 3/16/23 11:48, Patrick Palka wrote: > On Thu, 16 Mar 2023, Jason Merrill wrote: > >> On 3/16/23 10:09, Patrick Palka wrote: >>> On Wed, 15 Mar 2023, Patrick Palka wrote: >>> >>>> On Thu, 9 Mar 2023, Jason Merrill wrote: >>>> >>>>> On 3/9/23 14:32, Patrick Palka wrote: >>>>>> On Mon, 6 Mar 2023, Marek Polacek via Gcc-patches wrote: >>>>>> >>>>>>> When processing a noexcept, constructors aren't elided: >>>>>>> build_over_call >>>>>>> has >>>>>>> /* It's unsafe to elide the constructor when handling >>>>>>> a noexcept-expression, it may evaluate to the wrong >>>>>>> value (c++/53025). */ >>>>>>> && (force_elide || cp_noexcept_operand == 0)) >>>>>>> so the assert I added recently needs to be relaxed a little bit. >>>>>>> >>>>>>> Bootstrapped/regtested on x86_64-pc-linux-gnu, ok for trunk? >>>>>>> >>>>>>> PR c++/109030 >>>>>>> >>>>>>> gcc/cp/ChangeLog: >>>>>>> >>>>>>> * constexpr.cc (cxx_eval_call_expression): Relax assert. >>>>>>> >>>>>>> gcc/testsuite/ChangeLog: >>>>>>> >>>>>>> * g++.dg/cpp0x/noexcept77.C: New test. >>>>>>> --- >>>>>>> gcc/cp/constexpr.cc | 6 +++++- >>>>>>> gcc/testsuite/g++.dg/cpp0x/noexcept77.C | 9 +++++++++ >>>>>>> 2 files changed, 14 insertions(+), 1 deletion(-) >>>>>>> create mode 100644 gcc/testsuite/g++.dg/cpp0x/noexcept77.C >>>>>>> >>>>>>> diff --git a/gcc/cp/constexpr.cc b/gcc/cp/constexpr.cc >>>>>>> index 364695b762c..5384d0e8e46 100644 >>>>>>> --- a/gcc/cp/constexpr.cc >>>>>>> +++ b/gcc/cp/constexpr.cc >>>>>>> @@ -2869,7 +2869,11 @@ cxx_eval_call_expression (const constexpr_ctx >>>>>>> *ctx, >>>>>>> tree t, >>>>>>> /* We used to shortcut trivial constructor/op= here, but >>>>>>> nowadays >>>>>>> we can only get a trivial function here with >>>>>>> -fno-elide-constructors. */ >>>>>>> - gcc_checking_assert (!trivial_fn_p (fun) || >>>>>>> !flag_elide_constructors); >>>>>>> + gcc_checking_assert (!trivial_fn_p (fun) >>>>>>> + || !flag_elide_constructors >>>>>>> + /* We don't elide constructors when processing >>>>>>> + a noexcept-expression. */ >>>>>>> + || cp_noexcept_operand); >>>>>> >>>>>> It seems weird that we're performing constant evaluation within an >>>>>> unevaluated operand. Would it make sense to also fix this a second >>>>>> way >>>>>> by avoiding constant evaluation from maybe_constant_init when >>>>>> cp_unevaluated_operand && !manifestly_const_eval, like in >>>>>> maybe_constant_value? >>>>> >>>>> Sounds good. >>>> >>>> Hmm, while working on this I noticed we currently don't reject a version >>>> of >>>> g++.dg/cpp2a/constexpr-inst1.C that list initializes an aggregate instead >>>> of >>>> int (ever since r12-4425-g1595fe44e11a96): >>>> >>>> struct A { int m; }; >>>> template<typename T> constexpr int f() { return T::value; } >>>> template<bool B, typename T> void h(decltype(A{B ? f<T>() : 0})); // >>>> was int{...} >>>> template<bool B, typename T> void h(...); >>>> void x() { >>>> h<false, int>(0); // OK? >>>> } >>>> >>>> ISTM we should instantiate f<int> here for the same reason we do in the >>>> original version of the testcase, and for that to happen we need to >>>> pass manifestly_const_eval=true in massage_init_elt. Does that seem >>>> reasonable? >>>> >>> >>> FWIW the reason this came up is because I tried contriving a testcase >>> for the aforementioned maybe_constant_init change, and I came up with: >>> >>> struct __as_receiver { >>> int empty_env; >>> }; >>> >>> template<class T> >>> constexpr int f(T t) { >>> return t.fail; >>> }; >>> >>> using type = decltype(__as_receiver{f(0)}); // OK, f<int> no longer >>> instantiated >>> >>> which we used to reject and afterwards accept. But since the elements >>> of an initializer list are potentially constant evaluated, I wonder if >>> that that means f<int> should be instantiated here after all despite the >>> unevaluated context? >> >> The relevant section of the standard would seem to be >> https://eel.is/c++draft/expr.const#20 ; an immediate subexpression of a >> braced-init-list is potentially constant-evaluated even though it isn't >> potentially-evaluated or manifestly constant-evaluated. >> >> It seems like the call to fold_non_dependent_expr in check_narrowing ought to >> cause instantiation in this case, why doesn't it? > > Looks like check_narrowing isn't called at all in this aggr init case. > The call from e.g. convert_like_internal isn't reached because the > conversion for the initializer element is ck_identity, and don't ever > set conversion::check_narrowing for ck_identity conversions I think. Ah, yes, that makes sense; an identity conversion can never be narrowing, so we don't care about the constant value. So not instantiating seems correct, and the patch is OK. > Yet for using 'type = decltype(int{f(0)});' (similar to the example in > [temp.inst]/8) we do call check_narrowing directly from > finish_compound_literal, despite the conversion effectively being an > identity conversion. Hmm, maybe check_narrowing should defer constant evaluation until after deciding that the target type is not a superset of the source type... >>> Here's the full patch for reference: >>> >>> -- >8 -- >>> >>> Subject: [PATCH] c++: maybe_constant_init and unevaluated operands >>> [PR109030] >>> >>> This testcase in this PR (already fixed by r13-6526-ge4692319fd5fc7) >>> illustrates that maybe_constant_init can be called on an unevaluated >>> operand (from massage_init_elt), so this entry point should limit >>> constant evaluation in that case, like maybe_constant_value does. >>> >>> PR c++/109030 >>> >>> gcc/cp/ChangeLog: >>> >>> * constexpr.cc (maybe_constant_init_1): For an unevaluated >>> non-manifestly-constant operand, don't constant evaluate >>> and instead call fold_to_constant. >>> >>> gcc/testsuite/ChangeLog: >>> >>> * g++.dg/cpp0x/decltype83.C: New test. >>> --- >>> gcc/cp/constexpr.cc | 2 ++ >>> gcc/testsuite/g++.dg/cpp0x/decltype83.C | 14 ++++++++++++++ >>> 2 files changed, 16 insertions(+) >>> create mode 100644 gcc/testsuite/g++.dg/cpp0x/decltype83.C >>> >>> diff --git a/gcc/cp/constexpr.cc b/gcc/cp/constexpr.cc >>> index 8683c00596a..f325af375c8 100644 >>> --- a/gcc/cp/constexpr.cc >>> +++ b/gcc/cp/constexpr.cc >>> @@ -8795,6 +8795,8 @@ maybe_constant_init_1 (tree t, tree decl, bool >>> allow_non_constant, >>> && (TREE_STATIC (decl) || DECL_EXTERNAL (decl))); >>> if (is_static) >>> manifestly_const_eval = true; >>> + if (cp_unevaluated_operand && !manifestly_const_eval) >>> + return fold_to_constant (t); >>> t = cxx_eval_outermost_constant_expr (t, allow_non_constant, >>> !is_static, >>> mce_value (manifestly_const_eval), >>> false, decl); >>> diff --git a/gcc/testsuite/g++.dg/cpp0x/decltype83.C >>> b/gcc/testsuite/g++.dg/cpp0x/decltype83.C >>> new file mode 100644 >>> index 00000000000..17005a92eb5 >>> --- /dev/null >>> +++ b/gcc/testsuite/g++.dg/cpp0x/decltype83.C >>> @@ -0,0 +1,14 @@ >>> +// { dg-do compile { target c++11 } } >>> + >>> +struct __as_receiver { >>> + int empty_env; >>> +}; >>> + >>> +template<class T> >>> +constexpr int f(T t) { >>> + return t.fail; >>> +}; >>> + >>> +int main() { >>> + using type = decltype(__as_receiver{f(0)}); // OK, f<int> not >>> instantiated >>> +} >> >> >
diff --git a/gcc/cp/constexpr.cc b/gcc/cp/constexpr.cc index 364695b762c..5384d0e8e46 100644 --- a/gcc/cp/constexpr.cc +++ b/gcc/cp/constexpr.cc @@ -2869,7 +2869,11 @@ cxx_eval_call_expression (const constexpr_ctx *ctx, tree t, /* We used to shortcut trivial constructor/op= here, but nowadays we can only get a trivial function here with -fno-elide-constructors. */ - gcc_checking_assert (!trivial_fn_p (fun) || !flag_elide_constructors); + gcc_checking_assert (!trivial_fn_p (fun) + || !flag_elide_constructors + /* We don't elide constructors when processing + a noexcept-expression. */ + || cp_noexcept_operand); bool non_constant_args = false; new_call.bindings diff --git a/gcc/testsuite/g++.dg/cpp0x/noexcept77.C b/gcc/testsuite/g++.dg/cpp0x/noexcept77.C new file mode 100644 index 00000000000..16db8eb79ee --- /dev/null +++ b/gcc/testsuite/g++.dg/cpp0x/noexcept77.C @@ -0,0 +1,9 @@ +// PR c++/109030 +// { dg-do compile { target c++11 } } + +struct foo { }; + +struct __as_receiver { + foo empty_env; +}; +void sched(foo __fun) noexcept(noexcept(__as_receiver{__fun})) { }