public inbox for gcc-cvs@sourceware.org
help / color / mirror / Atom feed
From: Patrick Palka <ppalka@gcc.gnu.org>
To: gcc-cvs@gcc.gnu.org
Subject: [gcc r14-1519] c++: is_specialization_of_friend confusion [PR109923]
Date: Sat,  3 Jun 2023 13:27:24 +0000 (GMT)	[thread overview]
Message-ID: <20230603132724.C421A3858422@sourceware.org> (raw)

https://gcc.gnu.org/g:38af81df6aac11d49f9378d1c628ccc3c815c3ff

commit r14-1519-g38af81df6aac11d49f9378d1c628ccc3c815c3ff
Author: Patrick Palka <ppalka@redhat.com>
Date:   Sat Jun 3 09:26:43 2023 -0400

    c++: is_specialization_of_friend confusion [PR109923]
    
    The check for a non-template member function of a class template in
    is_specialization_of_friend is overbroad, and accidentally holds for a
    non-template hidden friend too, which for the testcase below causes the
    predicate to bogusly return true for
    
      decl = void non_templ_friend(A<int>, A<void>)
      friend_decl = void non_templ_friend(A<void>, A<void>)
    
    This patch refines the check appropriately.
    
            PR c++/109923
    
    gcc/cp/ChangeLog:
    
            * pt.cc (is_specialization_of_friend): Fix overbroad check for
            a non-template member function of a class template.
    
    gcc/testsuite/ChangeLog:
    
            * g++.dg/template/friend79.C: New test.

Diff:
---
 gcc/cp/pt.cc                             |  1 +
 gcc/testsuite/g++.dg/template/friend79.C | 20 ++++++++++++++++++++
 2 files changed, 21 insertions(+)

diff --git a/gcc/cp/pt.cc b/gcc/cp/pt.cc
index dae3815f62a..6b20c58ce66 100644
--- a/gcc/cp/pt.cc
+++ b/gcc/cp/pt.cc
@@ -1319,6 +1319,7 @@ is_specialization_of_friend (tree decl, tree friend_decl)
      of a template class, we want to check if DECL is a specialization
      if this.  */
   if (TREE_CODE (friend_decl) == FUNCTION_DECL
+      && DECL_CLASS_SCOPE_P (friend_decl)
       && DECL_TEMPLATE_INFO (friend_decl)
       && !DECL_USE_TEMPLATE (friend_decl))
     {
diff --git a/gcc/testsuite/g++.dg/template/friend79.C b/gcc/testsuite/g++.dg/template/friend79.C
new file mode 100644
index 00000000000..7603adf5bea
--- /dev/null
+++ b/gcc/testsuite/g++.dg/template/friend79.C
@@ -0,0 +1,20 @@
+// PR c++/109923
+
+template<class T>
+struct A {
+private:
+  int x;
+
+public:
+  A() : x(0) { }
+
+  friend void non_templ_friend(A val, A<void> weird) {
+    val.x++; // always works
+    weird.x++; // { dg-error "private" } should only work when T=void
+  }
+};
+
+int main() {
+  non_templ_friend(A<void>(), A<void>()); // { dg-bogus "" }
+  non_templ_friend(A<int>(), A<void>()); // { dg-message "required from here" }
+}

                 reply	other threads:[~2023-06-03 13:27 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=20230603132724.C421A3858422@sourceware.org \
    --to=ppalka@gcc.gnu.org \
    --cc=gcc-cvs@gcc.gnu.org \
    /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).