Re: [PATCH] c++: ICE with noexcept and local specialization [PR114114]

[Marek Polacek <polacek@redhat.com>]

On Fri, Mar 15, 2024 at 12:12:49PM -0400, Patrick Palka wrote:
> On Fri, 15 Mar 2024, Marek Polacek wrote:
> 
> > On Fri, Mar 15, 2024 at 10:35:07AM -0400, Patrick Palka wrote:
> > > On Tue, 5 Mar 2024, Marek Polacek wrote:
> > > 
> > > > Bootstrapped/regtested on x86_64-pc-linux-gnu, ok for trunk?
> > > > 
> > > > -- >8 --
> > > > Here we ICE because we call register_local_specialization while
> > > > local_specializations is null, so
> > > > 
> > > >   local_specializations->put ();
> > > > 
> > > > crashes on null this.  It's null since maybe_instantiate_noexcept calls
> > > > push_to_top_level which creates a new scope.  Normally, I would have
> > > > guessed that we need a new local_specialization_stack.  But here we're
> > > > dealing with an operand of a noexcept, which is an unevaluated operand,
> > > > and those aren't registered in the hash map.  maybe_instantiate_noexcept
> > > > wasn't signalling that it's substituting an unevaluated operand though.
> > > 
> > > It thought it was noexcept-exprs rather than noexcept-specs that are
> > > unevaluated contexts?
> > 
> > Yes, sigh.  It would have to be noexcept(noexcept(x)).  I was looking at
> > cp_parser_unary_expression/RID_NOEXCEPT but that's a noexcept-expr.  So
> > what can we do here, set a new local_specialization_stack?  That wasn't
> > that straightforward when I tried.  Or maybe just
> 
> Maybe we can avoid doing push_to_top_level (which clears
> local_specializations) from maybe_instantiate_noexcept if
> current_function_decl == fn?

Thanks, I agree that not doing push_to_top_level in the first place
is a better fix.  I just sent a patch that does that.
 
> Relatedly I wonder if we can avoid calling regenerate_decl_from_template
> for local class member functions since they can't be redeclared?

Good point.  I've tried the below, but that breaks a lot of contracts tests.
I have not pursued it further than that.

diff --git a/gcc/cp/pt.cc b/gcc/cp/pt.cc
index a7ba8b5af92..5352453a5d3 100644
--- a/gcc/cp/pt.cc
+++ b/gcc/cp/pt.cc
@@ -26623,6 +26623,12 @@ regenerate_decl_from_template (tree decl, tree tmpl, tree args)
       if (DECL_UNIQUE_FRIEND_P (decl))
 	goto done;
 
+      /* [class.mem.general]/5 says that a member shall not be declared twice
+	 in the member-specification (unless it's a nested class or member class
+	 template or an enumeration).  */
+      if (DECL_CLASS_SCOPE_P (decl))
+	goto done;
+
       /* Use the source location of the definition.  */
       DECL_SOURCE_LOCATION (decl) = DECL_SOURCE_LOCATION (tmpl);
 

Marek