public inbox for gcc-cvs@sourceware.org
help / color / mirror / Atom feed
* [gcc r12-5252] c++: constexpr virtual and vbase thunk
@ 2021-11-15  7:47 Jason Merrill
  0 siblings, 0 replies; only message in thread
From: Jason Merrill @ 2021-11-15  7:47 UTC (permalink / raw)
  To: gcc-cvs

https://gcc.gnu.org/g:267318a2857a42922c3ca033dac4690172b17683

commit r12-5252-g267318a2857a42922c3ca033dac4690172b17683
Author: Jason Merrill <jason@redhat.com>
Date:   Wed Nov 10 16:42:04 2021 -0500

    c++: constexpr virtual and vbase thunk
    
    C++20 allows virtual functions to be constexpr.  I don't think that calling
    through a pointer to a vbase subobject is supposed to work in a constant
    expression, since an object with virtual bases can't be constant, but the
    call shouldn't ICE.
    
    gcc/cp/ChangeLog:
    
            * constexpr.c (cxx_eval_thunk_call): Error instead of ICE
            on vbase thunk to constexpr function.
    
    gcc/testsuite/ChangeLog:
    
            * g++.dg/cpp2a/constexpr-virtual20.C: New test.

Diff:
---
 gcc/cp/constexpr.c                               | 15 ++++++++++-----
 gcc/testsuite/g++.dg/cpp2a/constexpr-virtual20.C | 22 ++++++++++++++++++++++
 2 files changed, 32 insertions(+), 5 deletions(-)

diff --git a/gcc/cp/constexpr.c b/gcc/cp/constexpr.c
index 453007c686b..7c27131f506 100644
--- a/gcc/cp/constexpr.c
+++ b/gcc/cp/constexpr.c
@@ -2246,15 +2246,20 @@ cxx_eval_thunk_call (const constexpr_ctx *ctx, tree t, tree thunk_fndecl,
 {
   tree function = THUNK_TARGET (thunk_fndecl);
 
-  /* virtual_offset is only set in the presence of virtual bases, which make
-     the class non-literal, so we don't need to handle it here.  */
   if (THUNK_VIRTUAL_OFFSET (thunk_fndecl))
     {
-      gcc_assert (!DECL_DECLARED_CONSTEXPR_P (function));
       if (!ctx->quiet)
 	{
-	  error ("call to non-%<constexpr%> function %qD", function);
-	  explain_invalid_constexpr_fn (function);
+	  if (!DECL_DECLARED_CONSTEXPR_P (function))
+	    {
+	      error ("call to non-%<constexpr%> function %qD", function);
+	      explain_invalid_constexpr_fn (function);
+	    }
+	  else
+	    /* virtual_offset is only set for virtual bases, which make the
+	       class non-literal, so we don't need to handle it here.  */
+	    error ("calling constexpr member function %qD through virtual "
+		   "base subobject", function);
 	}
       *non_constant_p = true;
       return t;
diff --git a/gcc/testsuite/g++.dg/cpp2a/constexpr-virtual20.C b/gcc/testsuite/g++.dg/cpp2a/constexpr-virtual20.C
new file mode 100644
index 00000000000..3c411fa3fcc
--- /dev/null
+++ b/gcc/testsuite/g++.dg/cpp2a/constexpr-virtual20.C
@@ -0,0 +1,22 @@
+// Test for constexpr call through vbase thunk.
+// { dg-do compile { target c++20 } }
+
+class Rep {
+public:
+  constexpr virtual int foo() { return 1; }
+};
+
+class VBase {
+public:
+  constexpr virtual int foo() { return 2; }
+};
+
+class Main : public Rep, virtual public VBase {
+public:
+  constexpr virtual int foo() { return 5; }
+};
+
+int main() {
+  Main m;
+  static_assert(static_cast<VBase*>(&m)->foo() == 5); // { dg-error "" }
+}


^ permalink raw reply	[flat|nested] only message in thread

only message in thread, other threads:[~2021-11-15  7:47 UTC | newest]

Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-11-15  7:47 [gcc r12-5252] c++: constexpr virtual and vbase thunk Jason Merrill

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).