[C++ PATCH] Fix ICE in joust with -Wconversion (PR c++/81167)

[C++ PATCH] Fix ICE in joust with -Wconversion (PR c++/81167)

[Jakub Jelinek]

Hi!

As mentioned in the PR, we ICE on this testcase because
w->fn is a conversion operator, w->convs[0]->type is a reference
to a class type, but because that conversion is ck_ref_bind,
source_type looks through it and finds ck_identity with
the class type.  Then we because w->fn is not a constructor
do source = TREE_TYPE (source);, assuming we got a pointer
type like on the other 5 testcases in check-c++-all with -Wconversion
that cover this code, so source is NULL and we die in calling
warning with bogus arguments.

The following patch fixes it by only using TREE_TYPE on pointer/reference
types.  I must say I don't understand this fully, but conversion
operators should be used on class types, so that is what we are looking for
with source, right?  No idea about the ! DECL_CONSTRUCTOR_P (w->fn)
check though and nothing in the testsuite covers that.

Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk?

2018-01-18  Jakub Jelinek  <jakub@redhat.com>

	PR c++/81167
	* call.c (joust): Only use TREE_TYPE (source) if source is
	a POINTER_TYPE_P.

	* g++.dg/cpp0x/pr81167.C: New test.

--- gcc/cp/call.c.jj	2018-01-17 22:00:13.626228171 +0100
+++ gcc/cp/call.c	2018-01-18 21:01:51.596136303 +0100
@@ -10082,7 +10082,7 @@ joust (struct z_candidate *cand1, struct
       else if (warn)
 	{
 	  tree source = source_type (w->convs[0]);
-	  if (! DECL_CONSTRUCTOR_P (w->fn))
+	  if (! DECL_CONSTRUCTOR_P (w->fn) && POINTER_TYPE_P (source))
 	    source = TREE_TYPE (source);
 	  if (warning (OPT_Wconversion, "choosing %qD over %qD", w->fn, l->fn)
 	      && warning (OPT_Wconversion, "  for conversion from %qH to %qI",
--- gcc/testsuite/g++.dg/cpp0x/pr81167.C.jj	2018-01-18 21:07:48.723189500 +0100
+++ gcc/testsuite/g++.dg/cpp0x/pr81167.C	2018-01-18 21:07:18.610187374 +0100
@@ -0,0 +1,24 @@
+// PR c++/81167
+// { dg-do compile { target c++11 } }
+// { dg-options "-Wconversion" }
+
+struct bar;
+
+struct foo
+{
+  foo () {}
+  foo (const bar &) {}
+};
+
+struct bar
+{
+  operator foo () && { return foo (); }
+};
+
+void test ()
+{
+  foo f = bar ();
+// { dg-warning "choosing 'bar::operator foo\\(\\) &&' over 'foo::foo\\(const bar&\\)'" "" { target *-*-* } .-1 }
+// { dg-warning "for conversion from 'bar' to 'foo'" "" { target *-*-* } .-2 }
+// { dg-message "because conversion sequence for the argument is better" "" { target *-*-* } .-3 }
+}

	Jakub

Re: [C++ PATCH] Fix ICE in joust with -Wconversion (PR c++/81167)

[Jason Merrill]

On Thu, Jan 18, 2018 at 6:35 PM, Jakub Jelinek <jakub@redhat.com> wrote:
> Hi!
>
> As mentioned in the PR, we ICE on this testcase because
> w->fn is a conversion operator, w->convs[0]->type is a reference
> to a class type, but because that conversion is ck_ref_bind,
> source_type looks through it and finds ck_identity with
> the class type.  Then we because w->fn is not a constructor
> do source = TREE_TYPE (source);, assuming we got a pointer
> type like on the other 5 testcases in check-c++-all with -Wconversion
> that cover this code, so source is NULL and we die in calling
> warning with bogus arguments.
>
> The following patch fixes it by only using TREE_TYPE on pointer/reference
> types.  I must say I don't understand this fully, but conversion
> operators should be used on class types, so that is what we are looking for
> with source, right?  No idea about the ! DECL_CONSTRUCTOR_P (w->fn)
> check though and nothing in the testsuite covers that.

I think that the test for POINTER_TYPE_P can replace the
DECL_CONSTRUCTOR_P test.

Jason