public inbox for gcc-patches@gcc.gnu.org
 help / color / mirror / Atom feed
From: Nathan Sidwell <nathan@acm.org>
To: GCC Patches <gcc-patches@gcc.gnu.org>
Cc: Jason Merrill <jason@redhat.com>
Subject: [c++] Implement DR 976
Date: Sat, 10 Jun 2023 17:28:42 -0400	[thread overview]
Message-ID: <4af0dc0a-b06b-372c-f2c3-e58b2141e027@acm.org> (raw)

[-- Attachment #1: Type: text/plain, Size: 397 bytes --]

DR 976 affects conversion operator deduction, swapping reference stripping and 
cv-qual removal.  This allows 'Y::operator P const & ()' to deduce T against a 
call wanting plain A  (previously that would fail as 'P const' cannot be deduced 
from 'A').

It also affects deductions for array- or function-producing conversions, which I 
suspect is rarer.

pushed to trunk

nathan
-- 
Nathan Sidwell

[-- Attachment #2: 0001-c-Adjust-conversion-deduction-PR61663-DR976.patch --]
[-- Type: text/x-patch, Size: 3195 bytes --]

From 80f075b410125bddb31459428760645baba1a69f Mon Sep 17 00:00:00 2001
From: Nathan Sidwell <nathan@acm.org>
Date: Sat, 10 Jun 2023 12:42:17 -0400
Subject: [PATCH] c++: Adjust conversion deduction [PR61663][DR976]

Drop the return type's reference before doing cvqual and related decays.

	gcc/cp/
	PR c++/61663
	* pt.cc (maybe_adjust_types_for_deduction): Implement DR976.
	gcc/testsuite/
	* g++.dg/template/pr61663.C: New.
---
 gcc/cp/pt.cc                            | 11 +++--
 gcc/testsuite/g++.dg/template/pr61663.C | 63 +++++++++++++++++++++++++
 2 files changed, 69 insertions(+), 5 deletions(-)
 create mode 100644 gcc/testsuite/g++.dg/template/pr61663.C

diff --git a/gcc/cp/pt.cc b/gcc/cp/pt.cc
index 6b20c58ce66..6a2cf2c123f 100644
--- a/gcc/cp/pt.cc
+++ b/gcc/cp/pt.cc
@@ -22725,10 +22725,16 @@ maybe_adjust_types_for_deduction (tree tparms,
       break;
 
     case DEDUCE_CONV:
+      /* [temp.deduct.conv] First remove a reference type on parm.
+	 DRs 322 & 976 affected this.  */
+      if (TYPE_REF_P (*parm))
+	*parm = TREE_TYPE (*parm);
+
       /* Swap PARM and ARG throughout the remainder of this
 	 function; the handling is precisely symmetric since PARM
 	 will initialize ARG rather than vice versa.  */
       std::swap (parm, arg);
+
       break;
 
     case DEDUCE_EXACT:
@@ -22795,11 +22801,6 @@ maybe_adjust_types_for_deduction (tree tparms,
       result |= UNIFY_ALLOW_OUTER_MORE_CV_QUAL;
     }
 
-  /* DR 322. For conversion deduction, remove a reference type on parm
-     too (which has been swapped into ARG).  */
-  if (strict == DEDUCE_CONV && TYPE_REF_P (*arg))
-    *arg = TREE_TYPE (*arg);
-
   return result;
 }
 
diff --git a/gcc/testsuite/g++.dg/template/pr61663.C b/gcc/testsuite/g++.dg/template/pr61663.C
new file mode 100644
index 00000000000..2964fa6c309
--- /dev/null
+++ b/gcc/testsuite/g++.dg/template/pr61663.C
@@ -0,0 +1,63 @@
+// { dg-do compile { target c++11 } }
+// PR c++/61663
+// DR 976, strip ref from conv op return type before doing
+// fn and ary decay or CV qual removal
+
+struct F 
+{
+  template<class T>
+  operator const T&();
+};
+
+void Foo () 
+{
+  F f;
+  int i = f;
+}
+
+template<typename T> struct X {};
+
+struct Y
+{
+  template<typename T> operator X<T> () &&; // #3
+  template<typename T> operator X<T> const & () const &; // #4
+};
+
+void Use (X<void>);
+Y Val ();
+Y const &Ref ();
+
+// { dg-final { scan-assembler "_Z5Frob3v:.*_ZNO1Ycv1XIT_EIvEEv.*_Z3Use1XIvE" } }
+void Frob3 ()
+{
+  Use (Val ()); // #3
+}
+
+// { dg-final { scan-assembler "_Z5Frob4v:.*_ZNKR1YcvRK1XIT_EIvEEv.*_Z3Use1XIvE" } }
+void Frob4 ()
+{
+  Use (Ref ()); // #4
+}
+
+struct Z 
+{
+  template<typename T> using FnRef = void (&) (T);
+  template<typename T> using AryRef = T (&)[];
+
+  template<typename T> operator FnRef<T> ();
+  template<typename T> operator AryRef<T> ();
+};
+
+// { dg-final { scan-assembler "_Z5Frob5R1Z:.*_ZN1ZcvRFvT_EIiEEv.*_ZN1ZcvRA_T_IiEEv" } }
+void Frob5 (Z &z)
+{
+  void (*fnptr)(int) = z;
+  int *iptr = z;
+}
+
+// { dg-final { scan-assembler "_Z5Frob6R1Z:.*_ZN1ZcvRFvT_EIfEEv.*_ZN1ZcvRA_T_IfEEv" } }
+void Frob6 (Z &z)
+{
+  void (&fnref)(float) = z;
+  float (&aryref)[] = z;
+}
-- 
2.40.1


                 reply	other threads:[~2023-06-10 21:28 UTC|newest]

Thread overview: [no followups] expand[flat|nested]  mbox.gz  Atom feed

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=4af0dc0a-b06b-372c-f2c3-e58b2141e027@acm.org \
    --to=nathan@acm.org \
    --cc=gcc-patches@gcc.gnu.org \
    --cc=jason@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).