From: Jason Merrill <jason@redhat.com>
To: Patrick Palka <ppalka@redhat.com>
Cc: gcc-patches@gcc.gnu.org
Subject: Re: [PATCH] c++: detecting copy-init context during CTAD [PR102137]
Date: Tue, 8 Mar 2022 15:07:31 -0500 [thread overview]
Message-ID: <ff5e5ce7-c845-01c0-a844-b6dcf3e33dbd@redhat.com> (raw)
In-Reply-To: <6962ed32-1424-07c8-c165-0e9441b93433@idea>
On 3/8/22 14:38, Patrick Palka wrote:
> On Tue, 8 Mar 2022, Jason Merrill wrote:
>
>> On 3/8/22 11:36, Patrick Palka wrote:
>>> On Mon, 7 Mar 2022, Jason Merrill wrote:
>>>
>>>> On 3/7/22 10:47, Patrick Palka wrote:
>>>>> On Fri, 4 Mar 2022, Jason Merrill wrote:
>>>>>
>>>>>> On 3/4/22 14:24, Patrick Palka wrote:
>>>>>>> Here we're failing to communicate to cp_finish_decl from tsubst_expr
>>>>>>> that we're in a copy-initialization context (via the
>>>>>>> LOOKUP_ONLYCONVERTING
>>>>>>> flag), which causes do_class_deduction to always consider explicit
>>>>>>> deduction guides when performing CTAD for a templated variable
>>>>>>> initializer.
>>>>>>>
>>>>>>> We could fix this by passing LOOKUP_ONLYCONVERTING appropriately
>>>>>>> when
>>>>>>> calling cp_finish_decl from tsubst_expr, but it seems
>>>>>>> do_class_deduction
>>>>>>> can determine if we're in a copy-init context by simply inspecting
>>>>>>> the
>>>>>>> initializer, and thus render its flags parameter unnecessary, which
>>>>>>> is
>>>>>>> what this patch implements. (If we were to fix this in tsubst_expr
>>>>>>> instead, I think we'd have to inspect the initializer in the same
>>>>>>> way
>>>>>>> in order to detect a copy-init context?)
>>>>>>
>>>>>> Hmm, does this affect conversions as well?
>>>>>>
>>>>>> Looks like it does:
>>>>>>
>>>>>> struct A
>>>>>> {
>>>>>> explicit operator int();
>>>>>> };
>>>>>>
>>>>>> template <class T> void f()
>>>>>> {
>>>>>> T t = A();
>>>>>> }
>>>>>>
>>>>>> int main()
>>>>>> {
>>>>>> f<int>(); // wrongly accepted
>>>>>> }
>>>>>>
>>>>>> The reverse, initializing via an explicit constructor, is caught by
>>>>>> code
>>>>>> in
>>>>>> build_aggr_init much like the code your patch adds to
>>>>>> do_auto_deduction;
>>>>>> perhaps we should move/copy that code to cp_finish_decl?
>>>>>
>>>>> Ah, makes sense. Moving that code from build_aggr_init to
>>>>> cp_finish_decl broke things, but using it in both spots seems to work
>>>>> well. And I suppose we might as well use it in do_class_deduction too,
>>>>> since doing so lets us remove the flags parameter.
>>>>
>>>> Before removing the flags parameter please try asserting that it now
>>>> matches
>>>> is_copy_initialization and see if anything breaks.
>>>
>>> I added to do_class_deduction:
>>>
>>> gcc_assert (bool(flags & LOOKUP_ONLYCONVERTING) == is_copy_initialization
>>> (init));
>>>
>>> Turns out removing the flags parameter breaks CTAD for new-expressions
>>> of the form 'new TT(x)' because in this case build_new passes just 'x'
>>> as the initializer to do_auto_deduction (as opposed to a single TREE_LIST),
>>> for which is_copy_initialization returns true even though it's really
>>> direct initalization.
>>>
>>> Also turns out we're similarly not passing the right LOOKUP_* flags to
>>> cp_finish_decl from instantiate_body, which breaks consideration of
>>> explicit conversions/deduction guides when instantiating the initializer
>>> of a static data member. I added some xfailed testcases for these
>>> situations.
>>
>> Maybe we want to check is_copy_initialization in cp_finish_decl?
>
> That seems to work nicely :) All xfailed tests for the static data
> member initialization case now also pass.
>
> Bootstrapped and regtested on x86_64-pc-linux-gnu, does this look OK for
> trunk?
>
> -- >8 --
>
> Subject: [PATCH] c++: detecting copy-init context during CTAD [PR102137]
>
> Here we're failing to communicate to cp_finish_decl from tsubst_expr
> that we're in a copy-initialization context (via the LOOKUP_ONLYCONVERTING
> flag), which causes us to always consider explicit deduction guides when
> performing CTAD for a templated variable initializer.
>
> It turns out this bug also affects consideration of explicit conversion
> operators for the same reason. But consideration of explicit constructors
> seems to do the right thing thanks to code in build_aggr_init that sets
> LOOKUP_ONLYCONVERTING when the initializer represents copy-initialization.
>
> This patch fixes this by making cp_finish_decl set LOOKUP_ONLYCONVERTING
> by inspecting the initializer like build_aggr_init does, so that callers
> don't need to explicitly pass this flag.
>
> PR c++/102137
> PR c++/87820
>
> gcc/cp/ChangeLog:
>
> * cp-tree.h (is_copy_initialization): Declare.
> * decl.cc (cp_finish_decl): Set LOOKUP_ONLYCONVERTING
> when is_copy_initialization is true.
> * init.cc (build_aggr_init): Split out copy-initialization
> check into ...
> (is_copy_initialization): ... here.
> * pt.cc (instantiate_decl): Pass 0 instead of
> LOOKUP_ONLYCONVERTING as flags to cp_finish_decl.
Any reason not to use LOOKUP_NORMAL, both here and in tsubst_expr? OK
either way.
> gcc/testsuite/ChangeLog:
>
> * g++.dg/cpp0x/explicit15.C: New test.
> * g++.dg/cpp1z/class-deduction108.C: New test.
> ---
> gcc/cp/cp-tree.h | 1 +
> gcc/cp/decl.cc | 3 +
> gcc/cp/init.cc | 20 +++--
> gcc/cp/pt.cc | 3 +-
> gcc/testsuite/g++.dg/cpp0x/explicit15.C | 83 +++++++++++++++++++
> .../g++.dg/cpp1z/class-deduction108.C | 78 +++++++++++++++++
> 6 files changed, 181 insertions(+), 7 deletions(-)
> create mode 100644 gcc/testsuite/g++.dg/cpp0x/explicit15.C
> create mode 100644 gcc/testsuite/g++.dg/cpp1z/class-deduction108.C
>
> diff --git a/gcc/cp/cp-tree.h b/gcc/cp/cp-tree.h
> index ac723901098..fd76909ca75 100644
> --- a/gcc/cp/cp-tree.h
> +++ b/gcc/cp/cp-tree.h
> @@ -7039,6 +7039,7 @@ extern void emit_mem_initializers (tree);
> extern tree build_aggr_init (tree, tree, int,
> tsubst_flags_t);
> extern int is_class_type (tree, int);
> +extern bool is_copy_initialization (tree);
> extern tree build_zero_init (tree, tree, bool);
> extern tree build_value_init (tree, tsubst_flags_t);
> extern tree build_value_init_noctor (tree, tsubst_flags_t);
> diff --git a/gcc/cp/decl.cc b/gcc/cp/decl.cc
> index 199ac768d43..5452f7d9cb3 100644
> --- a/gcc/cp/decl.cc
> +++ b/gcc/cp/decl.cc
> @@ -7968,6 +7968,9 @@ cp_finish_decl (tree decl, tree init, bool init_const_expr_p,
> if (type == error_mark_node)
> return;
>
> + if (VAR_P (decl) && is_copy_initialization (init))
> + flags |= LOOKUP_ONLYCONVERTING;
> +
> /* Warn about register storage specifiers except when in GNU global
> or local register variable extension. */
> if (VAR_P (decl) && DECL_REGISTER (decl) && asmspec_tree == NULL_TREE)
> diff --git a/gcc/cp/init.cc b/gcc/cp/init.cc
> index 545d904c0f9..d5faa0dc519 100644
> --- a/gcc/cp/init.cc
> +++ b/gcc/cp/init.cc
> @@ -2019,11 +2019,7 @@ build_aggr_init (tree exp, tree init, int flags, tsubst_flags_t complain)
> return stmt_expr;
> }
>
> - if (init && init != void_type_node
> - && TREE_CODE (init) != TREE_LIST
> - && !(TREE_CODE (init) == TARGET_EXPR
> - && TARGET_EXPR_DIRECT_INIT_P (init))
> - && !DIRECT_LIST_INIT_P (init))
> + if (is_copy_initialization (init))
> flags |= LOOKUP_ONLYCONVERTING;
>
> is_global = begin_init_stmts (&stmt_expr, &compound_stmt);
> @@ -2331,6 +2327,20 @@ is_class_type (tree type, int or_else)
> return 1;
> }
>
> +/* Returns true iff the variable initializer INIT represents
> + copy-initialization (and therefore we must set LOOKUP_ONLYCONVERTING
> + when processing it). */
> +
> +bool
> +is_copy_initialization (tree init)
> +{
> + return (init && init != void_type_node
> + && TREE_CODE (init) != TREE_LIST
> + && !(TREE_CODE (init) == TARGET_EXPR
> + && TARGET_EXPR_DIRECT_INIT_P (init))
> + && !DIRECT_LIST_INIT_P (init));
> +}
> +
> /* Build a reference to a member of an aggregate. This is not a C++
> `&', but really something which can have its address taken, and
> then act as a pointer to member, for example TYPE :: FIELD can have
> diff --git a/gcc/cp/pt.cc b/gcc/cp/pt.cc
> index 402365285cf..e8b5d8fbb73 100644
> --- a/gcc/cp/pt.cc
> +++ b/gcc/cp/pt.cc
> @@ -26602,8 +26602,7 @@ instantiate_decl (tree d, bool defer_ok, bool expl_inst_class_mem_p)
> const_init
> = DECL_INITIALIZED_BY_CONSTANT_EXPRESSION_P (code_pattern);
> cp_finish_decl (d, init, /*init_const_expr_p=*/const_init,
> - /*asmspec_tree=*/NULL_TREE,
> - LOOKUP_ONLYCONVERTING);
> + /*asmspec_tree=*/NULL_TREE, 0);
> }
> if (enter_context)
> pop_nested_class ();
> diff --git a/gcc/testsuite/g++.dg/cpp0x/explicit15.C b/gcc/testsuite/g++.dg/cpp0x/explicit15.C
> new file mode 100644
> index 00000000000..076d8b31631
> --- /dev/null
> +++ b/gcc/testsuite/g++.dg/cpp0x/explicit15.C
> @@ -0,0 +1,83 @@
> +// PR c++/87820
> +// { dg-do compile { target c++11 } }
> +
> +struct A {
> + constexpr explicit operator int() const { return 0; }
> +};
> +
> +template<class T>
> +void f() {
> + A a;
> + T t1 = a; // { dg-error "cannot convert" }
> + T t2 = {a}; // { dg-error "cannot convert" }
> + T t3(a);
> + T t4{a};
> + T t5 = T(a);
> + T t6 = T{a};
> + new T(a);
> + new T{a};
> +}
> +
> +template<class T>
> +void g() {
> + T t;
> + int n1 = t; // { dg-error "cannot convert" }
> + int n2 = {t}; // { dg-error "cannot convert" }
> + int n3(t);
> + int n4{t};
> + int n5 = int(t);
> + int n6 = int{t};
> + new int(t);
> + new int{t};
> +}
> +
> +template void f<int>();
> +template void g<A>();
> +
> +template<class T>
> +struct B {
> + static constexpr A a{};
> + static constexpr T t1 = a; // { dg-error "cannot convert" }
> + static constexpr T t2 = {a}; // { dg-error "cannot convert" }
> + static constexpr T t4{a}; // { dg-bogus "cannot convert" }
> + static constexpr T t5 = T(a);
> + static constexpr T t6 = T{a};
> +};
> +
> +template<class T>
> +struct C {
> + static constexpr T t{};
> + static constexpr int n1 = t; // { dg-error "cannot convert" }
> + static constexpr int n2 = {t}; // { dg-error "cannot convert" }
> + static constexpr int n4{t}; // { dg-bogus "cannot convert" }
> + static constexpr int n5 = int(t);
> + static constexpr int n6 = int{t};
> +};
> +
> +template struct B<int>;
> +template struct C<A>;
> +
> +#if __cpp_inline_variables
> +template<class T>
> +struct D {
> + static inline A a;
> + static inline T t1 = a; // { dg-error "cannot convert" "" { target c++17 } }
> + static inline T t2 = {a}; // { dg-error "cannot convert" "" { target c++17 } }
> + static inline T t4{a};
> + static inline T t5 = T(a);
> + static inline T t6 = T{a};
> +};
> +
> +template<class T>
> +struct E {
> + static inline T t;
> + static inline int n1 = t; // { dg-error "cannot convert" "" { target c++17 } }
> + static inline int n2 = {t}; // { dg-error "cannot convert" "" { target c++17 } }
> + static inline int n4{t};
> + static inline int n5 = int(t);
> + static inline int n6 = int{t};
> +};
> +
> +template struct D<int>;
> +template struct E<A>;
> +#endif
> diff --git a/gcc/testsuite/g++.dg/cpp1z/class-deduction108.C b/gcc/testsuite/g++.dg/cpp1z/class-deduction108.C
> new file mode 100644
> index 00000000000..e82c4bafb04
> --- /dev/null
> +++ b/gcc/testsuite/g++.dg/cpp1z/class-deduction108.C
> @@ -0,0 +1,78 @@
> +// PR c++/102137
> +// { dg-do compile { target c++17 } }
> +
> +template<class T>
> +struct A {
> + constexpr A() { }
> + constexpr A(int) { }
> +};
> +
> +explicit A(...) -> A<int>;
> +
> +template<template<class> class TT>
> +void f() {
> + TT x1 = 0; // { dg-error "deduction|no match" }
> + TT x2 = {0}; // { dg-error "explicit deduction guide" }
> + TT x3(0);
> + TT x4{0};
> + TT x5;
> + new TT(0);
> + new TT{0};
> + new TT();
> + new TT{};
> + new TT;
> +}
> +
> +template<class T>
> +void g(T t) {
> + A a1 = t; // { dg-error "deduction|no match" }
> + A a2 = {t}; // { dg-error "explicit deduction guide" }
> + A a3(t);
> + A a4{t};
> + A a5;
> + new A(t);
> + new A{t};
> +}
> +
> +template void f<A>();
> +template void g(int);
> +
> +template<template<class> class TT>
> +struct B {
> + static inline TT x1 = 0; // { dg-error "deduction|no match" }
> + static inline TT x2 = {0}; // { dg-error "explicit deduction guide" }
> + static inline TT x4{0};
> + static inline TT x5;
> +};
> +
> +template<class T>
> +struct C {
> + static inline T t;
> + static inline A a1 = t; // { dg-error "deduction|no match" }
> + static inline A a2 = {t}; // { dg-error "explicit deduction guide" }
> + static inline A a4{t};
> + static inline A a5{};
> +};
> +
> +template struct B<A>;
> +template struct C<int>;
> +
> +template<template<class> class TT>
> +struct E {
> + static constexpr TT x1 = 0; // { dg-error "deduction|no match" }
> + static constexpr TT x2 = {0}; // { dg-error "explicit deduction guide" }
> + static constexpr TT x4{0};
> + static constexpr TT x5{};
> +};
> +
> +template<class T>
> +struct F {
> + static constexpr T t{};
> + static constexpr A a1 = t; // { dg-error "deduction|no match" }
> + static constexpr A a2 = {t}; // { dg-error "explicit deduction guide" }
> + static constexpr A a4{t};
> + static constexpr A a5{};
> +};
> +
> +template struct E<A>;
> +template struct F<int>;
next prev parent reply other threads:[~2022-03-08 20:07 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-03-04 18:24 Patrick Palka
2022-03-04 22:34 ` Jason Merrill
2022-03-07 14:47 ` Patrick Palka
2022-03-07 19:14 ` Jason Merrill
2022-03-08 15:36 ` Patrick Palka
2022-03-08 16:24 ` Jason Merrill
2022-03-08 18:38 ` Patrick Palka
2022-03-08 20:07 ` Jason Merrill [this message]
2022-03-08 21:17 ` Patrick Palka
2022-03-08 22:26 ` Jason Merrill
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=ff5e5ce7-c845-01c0-a844-b6dcf3e33dbd@redhat.com \
--to=jason@redhat.com \
--cc=gcc-patches@gcc.gnu.org \
--cc=ppalka@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).