From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from us-smtp-delivery-124.mimecast.com (us-smtp-delivery-124.mimecast.com [170.10.129.124]) by sourceware.org (Postfix) with ESMTPS id E8E7C385841F for ; Mon, 7 Mar 2022 19:14:44 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.1 sourceware.org E8E7C385841F Received: from mail-qv1-f69.google.com (mail-qv1-f69.google.com [209.85.219.69]) by relay.mimecast.com with ESMTP with STARTTLS (version=TLSv1.2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id us-mta-553-UcAhzfvBN0O-pP9rMtDZCQ-1; Mon, 07 Mar 2022 14:14:39 -0500 X-MC-Unique: UcAhzfvBN0O-pP9rMtDZCQ-1 Received: by mail-qv1-f69.google.com with SMTP id o7-20020a0cfa87000000b004352258d2d1so13285086qvn.10 for ; Mon, 07 Mar 2022 11:14:39 -0800 (PST) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20210112; h=x-gm-message-state:message-id:date:mime-version:user-agent:subject :content-language:to:cc:references:from:in-reply-to :content-transfer-encoding; bh=TJ7/GG6Lhjz3PtYlln+F4B8LjpC+QEum0gRYkEkmUTc=; b=lHh7IaMyc7o9IML3JmKlvXESv/Dhscbxfa+O4je8R7brzt7jQimepF5VDJlz05zGIT g5Ri7CG9jpQMTQzIQdwlwQGbj5Dz1fflDKgAeTuayiuQdAzejQJHOSBPjjlqeU9eYSdW iaX4Go3Fb5aw3BWsU55Y8IGJan9qHCu/AgmqLXeJd7UQ5ac8cC5HJXhqgbnN8TlTMZuW PSF8BYYKWMC5ICmaEQL4a3W44y6aXprjF4g2WnWuSXOLm6eV8CBElZ1yIsGqMZSSDfD1 GJ/9/rpzxFBYuG/t1C0CPU7F3egG4uPnq7RveHfkMISRJCc4DbDaW7PZ+ctUEhsX9IVq H9JA== X-Gm-Message-State: AOAM531q051yBridRHePwLdeLesF2pTGsd9mUExHLO86bbxmUBA02W2V YKDEx3f/WECNOaA2sMoZUoNKm9SkGbZE2PFXoLP/V9HO86J38mqkuqw5vGJQWaAEPb6PDq5qihd UMo6IgD2J/BpL0U2AJA== X-Received: by 2002:a05:622a:1703:b0:2de:4080:f7c1 with SMTP id h3-20020a05622a170300b002de4080f7c1mr10753067qtk.161.1646680478968; Mon, 07 Mar 2022 11:14:38 -0800 (PST) X-Google-Smtp-Source: ABdhPJyVNSPXhR70KouFk5Z0lTI44XlhJ1Z5oXYwYrpJQgaGD14AoTk0Le24rQbXDzEm1smGuEDlPA== X-Received: by 2002:a05:622a:1703:b0:2de:4080:f7c1 with SMTP id h3-20020a05622a170300b002de4080f7c1mr10753036qtk.161.1646680478472; Mon, 07 Mar 2022 11:14:38 -0800 (PST) Received: from [192.168.1.149] (130-44-159-43.s15913.c3-0.arl-cbr1.sbo-arl.ma.cable.rcncustomer.com. [130.44.159.43]) by smtp.gmail.com with ESMTPSA id g9-20020ac85d49000000b002dff273cce5sm9316333qtx.72.2022.03.07.11.14.37 (version=TLS1_3 cipher=TLS_AES_128_GCM_SHA256 bits=128/128); Mon, 07 Mar 2022 11:14:37 -0800 (PST) Message-ID: Date: Mon, 7 Mar 2022 14:14:36 -0500 MIME-Version: 1.0 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:91.0) Gecko/20100101 Thunderbird/91.6.1 Subject: Re: [PATCH] c++: detecting copy-init context during CTAD [PR102137] To: Patrick Palka Cc: gcc-patches@gcc.gnu.org References: <20220304182436.122498-1-ppalka@redhat.com> From: Jason Merrill In-Reply-To: X-Mimecast-Spam-Score: 0 X-Mimecast-Originator: redhat.com Content-Language: en-US Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 7bit X-Spam-Status: No, score=-12.8 required=5.0 tests=BAYES_00, DKIMWL_WL_HIGH, DKIM_SIGNED, DKIM_VALID, DKIM_VALID_AU, DKIM_VALID_EF, GIT_PATCH_0, NICE_REPLY_A, RCVD_IN_DNSWL_NONE, RCVD_IN_MSPIKE_H5, RCVD_IN_MSPIKE_WL, SPF_HELO_NONE, SPF_NONE, TXREP, T_SCC_BODY_TEXT_LINE autolearn=ham autolearn_force=no version=3.4.4 X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on server2.sourceware.org X-BeenThere: gcc-patches@gcc.gnu.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Gcc-patches mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 07 Mar 2022 19:14:47 -0000 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 void f() >> { >> T t = A(); >> } >> >> int main() >> { >> f(); // 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. > -- >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 > is unaffected and seems to work correctly thanks to code in build_aggr_init > that sets LOOKUP_ONLYCONVERTING when the initializer represents > copy-initialization. > > This patch factors out the copy-initialization test from build_aggr_init > and reuses it in tsubst_expr for sake of cp_finish_decl. And we might > as well use it in do_class_deduction too, since doing so lets us get rid > of its flags parameter (which is used only to control whether explicit > deduction guides are considered). > > Bootstrapped and regtestd on x86_64-pc-linux-gnu, does this look OK for > trunk? > > PR c++/102137 > PR c++/87820 > > gcc/cp/ChangeLog: > > * cp-tree.h (is_copy_initialization): Declare. > (do_auto_deduction): Remove flags parameter. > * decl.cc (cp_finish_decl): Adjust call to do_auto_deduction. > * init.cc (build_aggr_init): Split out copy-initialization > check into ... > (is_copy_initialization): ... here. > * pt.cc (convert_template_argument): Adjust call to > do_auto_deduction. > (tsubst_expr) : Pass LOOKUP_ONLYCONVERTING > to cp_finish_decl when is_copy_initialization. > (do_class_deduction): Remove flags parameter, and instead > call is_copy_initialization to determine if we're in a copy-init > context. > (do_auto_deduction): Adjust call to do_class_deduction. > > gcc/testsuite/ChangeLog: > > * g++.dg/cpp0x/explicit15.C: New test. > * g++.dg/cpp1z/class-deduction108.C: New test. > --- > gcc/cp/cp-tree.h | 4 +-- > gcc/cp/decl.cc | 2 +- > gcc/cp/init.cc | 20 +++++++++--- > gcc/cp/pt.cc | 23 ++++++++------ > gcc/testsuite/g++.dg/cpp0x/explicit15.C | 31 +++++++++++++++++++ > .../g++.dg/cpp1z/class-deduction108.C | 31 +++++++++++++++++++ > 6 files changed, 93 insertions(+), 18 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..be556d2c441 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); > @@ -7279,8 +7280,7 @@ extern tree do_auto_deduction (tree, tree, tree, > = tf_warning_or_error, > auto_deduction_context > = adc_unspecified, > - tree = NULL_TREE, > - int = LOOKUP_NORMAL); > + tree = NULL_TREE); > extern tree type_uses_auto (tree); > extern tree type_uses_auto_or_concept (tree); > extern void append_type_to_template_for_access_check (tree, tree, tree, > diff --git a/gcc/cp/decl.cc b/gcc/cp/decl.cc > index 199ac768d43..152f657e9f2 100644 > --- a/gcc/cp/decl.cc > +++ b/gcc/cp/decl.cc > @@ -8039,7 +8039,7 @@ cp_finish_decl (tree decl, tree init, bool init_const_expr_p, > outer_targs = DECL_TI_ARGS (decl); > type = TREE_TYPE (decl) = do_auto_deduction (type, d_init, auto_node, > tf_warning_or_error, adc, > - outer_targs, flags); > + outer_targs); > if (type == error_mark_node) > return; > if (TREE_CODE (type) == FUNCTION_TYPE) > 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 d94d4538faa..a3ff81e2c3a 100644 > --- a/gcc/cp/pt.cc > +++ b/gcc/cp/pt.cc > @@ -8567,8 +8567,7 @@ convert_template_argument (tree parm, > can happen in the context of -fnew-ttp-matching. */; > else if (tree a = type_uses_auto (t)) > { > - t = do_auto_deduction (t, arg, a, complain, adc_unify, args, > - LOOKUP_IMPLICIT); > + t = do_auto_deduction (t, arg, a, complain, adc_unify, args); > if (t == error_mark_node) > return error_mark_node; > } > @@ -18606,7 +18605,11 @@ tsubst_expr (tree t, tree args, tsubst_flags_t complain, tree in_decl, > TREE_TYPE (asmspec_tree) = char_array_type_node; > } > > - cp_finish_decl (decl, init, const_init, asmspec_tree, 0); > + > + int flags = LOOKUP_NORMAL; > + if (VAR_P (decl) && is_copy_initialization (init)) > + flags |= LOOKUP_ONLYCONVERTING; > + cp_finish_decl (decl, init, const_init, asmspec_tree, flags); > > if (ndecl != error_mark_node) > cp_finish_decomp (ndecl, first, cnt); > @@ -29832,8 +29835,7 @@ ctad_template_p (tree tmpl) > type. */ > > static tree > -do_class_deduction (tree ptype, tree tmpl, tree init, > - int flags, tsubst_flags_t complain) > +do_class_deduction (tree ptype, tree tmpl, tree init, tsubst_flags_t complain) > { > /* We should have handled this in the caller. */ > if (DECL_TEMPLATE_TEMPLATE_PARM_P (tmpl)) > @@ -29885,6 +29887,7 @@ do_class_deduction (tree ptype, tree tmpl, tree init, > > bool try_list_ctor = false; > bool list_init_p = false; > + bool copy_init_p = is_copy_initialization (init); > > releasing_vec rv_args = NULL; > vec *&args = *&rv_args; > @@ -29929,7 +29932,7 @@ do_class_deduction (tree ptype, tree tmpl, tree init, > /* Prune explicit deduction guides in copy-initialization context (but > not copy-list-initialization). */ > bool elided = false; > - if (!list_init_p && (flags & LOOKUP_ONLYCONVERTING)) > + if (!list_init_p && copy_init_p) > { > for (lkp_iterator iter (cands); !elided && iter; ++iter) > if (DECL_NONCONVERTING_P (STRIP_TEMPLATE (*iter))) > @@ -30007,7 +30010,7 @@ do_class_deduction (tree ptype, tree tmpl, tree init, > } > /* [over.match.list]/1: In copy-list-initialization, if an explicit > constructor is chosen, the initialization is ill-formed. */ > - else if (flags & LOOKUP_ONLYCONVERTING) > + else if (copy_init_p) > { > if (DECL_NONCONVERTING_P (fndecl)) > { > @@ -30045,7 +30048,7 @@ do_class_deduction (tree ptype, tree tmpl, tree init, > /* Replace occurrences of 'auto' in TYPE with the appropriate type deduced > from INIT. AUTO_NODE is the TEMPLATE_TYPE_PARM used for 'auto' in TYPE. > The CONTEXT determines the context in which auto deduction is performed > - and is used to control error diagnostics. FLAGS are the LOOKUP_* flags. > + and is used to control error diagnostics. > > OUTER_TARGS is used during template argument deduction (context == adc_unify) > to properly substitute the result. It's also used in the adc_unify and > @@ -30058,7 +30061,7 @@ do_class_deduction (tree ptype, tree tmpl, tree init, > tree > do_auto_deduction (tree type, tree init, tree auto_node, > tsubst_flags_t complain, auto_deduction_context context, > - tree outer_targs, int flags) > + tree outer_targs) > { > if (init == error_mark_node) > return error_mark_node; > @@ -30079,7 +30082,7 @@ do_auto_deduction (tree type, tree init, tree auto_node, > > if (tree tmpl = CLASS_PLACEHOLDER_TEMPLATE (auto_node)) > /* C++17 class template argument deduction. */ > - return do_class_deduction (type, tmpl, init, flags, complain); > + return do_class_deduction (type, tmpl, init, complain); > > if (init == NULL_TREE || TREE_TYPE (init) == NULL_TREE) > /* Nothing we can do with this, even in deduction context. */ > diff --git a/gcc/testsuite/g++.dg/cpp0x/explicit15.C b/gcc/testsuite/g++.dg/cpp0x/explicit15.C > new file mode 100644 > index 00000000000..f6c584d5f88 > --- /dev/null > +++ b/gcc/testsuite/g++.dg/cpp0x/explicit15.C > @@ -0,0 +1,31 @@ > +// PR c++/87820 > +// { dg-do compile { target c++11 } } > + > +struct A { > + explicit operator int(); > +}; > + > +template > +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}; > +} > + > +template > +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}; > +} > + > +template void f(); > +template void g(); > 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..9e36c53c1a4 > --- /dev/null > +++ b/gcc/testsuite/g++.dg/cpp1z/class-deduction108.C > @@ -0,0 +1,31 @@ > +// PR c++/102137 > +// { dg-do compile { target c++17 } } > + > +template > +struct A { > + A(); > + A(int); > +}; > + > +explicit A(...) -> A; > + > +template class T> > +void f() { > + T x1 = 0; // { dg-error "deduction|no match" } > + T x2 = {0}; // { dg-error "explicit deduction guide" } > + T x3(0); > + T x4{0}; > + T x5; > +} > + > +template > +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; > +} > + > +template void f(); > +template void g(int);