public inbox for gcc-patches@gcc.gnu.org
 help / color / mirror / Atom feed
From: Jason Merrill <jason@redhat.com>
To: Patrick Palka <ppalka@redhat.com>, gcc-patches@gcc.gnu.org
Subject: Re: [PATCH] c++: scoped variable template-id of reference type [PR97340]
Date: Thu, 18 May 2023 14:56:56 -0400	[thread overview]
Message-ID: <f789ca2d-56b9-d6b1-9baa-99b2def9e8b7@redhat.com> (raw)
In-Reply-To: <20230518175927.4158045-1-ppalka@redhat.com>

On 5/18/23 13:59, Patrick Palka wrote:
> lookup_and_finish_template_variable calls convert_from_reference, which
> means for a variable template-id of reference type the function returns
> an INDIRECT_REF instead of the bare VAR_DECL.  But the downstream logic
> of two callers, tsubst_qualified_id and finish_class_member_access_expr,
> expect a DECL_P result and so we end up crashing when resolving the
> template-id's in the first testcase.  (Note that these two callers
> eventually call convert_from_reference as appropriate, so this earlier
> call seems at best redundant.)
> 
> This patch fixes this by pulling out the convert_from_reference call
> from lookup_and_finish_template_variable and into the callers that
> actually need it, which turns out to be tsubst_copy_and_build (without
> it we'd mishandle the second testcase).
> 
> Bootstrapped and regtested on x86_64-pc-linux-gnu, does this look OK for
> trunk?

OK.

> 	PR c++/97340
> 
> gcc/cp/ChangeLog:
> 
> 	* pt.cc (lookup_and_finish_template_variable): Don't call
> 	convert_from_reference.
> 	(tsubst_copy_and_build) <case TEMPLATE_ID_EXPR>: Call
> 	convert_from_reference on the result of
> 	lookup_and_finish_template_variable.
> 
> gcc/testsuite/ChangeLog:
> 
> 	* g++.dg/cpp1y/var-templ80.C: New test.
> 	* g++.dg/cpp1y/var-templ81.C: New test.
> ---
>   gcc/cp/pt.cc                             |  3 ++-
>   gcc/testsuite/g++.dg/cpp1y/var-templ80.C | 22 ++++++++++++++++++++++
>   gcc/testsuite/g++.dg/cpp1y/var-templ81.C | 14 ++++++++++++++
>   3 files changed, 38 insertions(+), 1 deletion(-)
>   create mode 100644 gcc/testsuite/g++.dg/cpp1y/var-templ80.C
>   create mode 100644 gcc/testsuite/g++.dg/cpp1y/var-templ81.C
> 
> diff --git a/gcc/cp/pt.cc b/gcc/cp/pt.cc
> index 351fc18b600..9e5b29f3099 100644
> --- a/gcc/cp/pt.cc
> +++ b/gcc/cp/pt.cc
> @@ -10394,7 +10394,7 @@ lookup_and_finish_template_variable (tree templ, tree targs,
>     complain &= ~tf_partial;
>     var = finish_template_variable (var, complain);
>     mark_used (var);
> -  return convert_from_reference (var);
> +  return var;
>   }
>   
>   /* If the set of template parameters PARMS contains a template parameter
> @@ -20462,6 +20462,7 @@ tsubst_copy_and_build (tree t,
>   	  {
>   	    tree r = lookup_and_finish_template_variable (templ, targs,
>   							  complain);
> +	    r = convert_from_reference (r);
>   	    r = maybe_wrap_with_location (r, EXPR_LOCATION (t));
>   	    RETURN (r);
>   	  }
> diff --git a/gcc/testsuite/g++.dg/cpp1y/var-templ80.C b/gcc/testsuite/g++.dg/cpp1y/var-templ80.C
> new file mode 100644
> index 00000000000..4439bee8292
> --- /dev/null
> +++ b/gcc/testsuite/g++.dg/cpp1y/var-templ80.C
> @@ -0,0 +1,22 @@
> +// PR c++/97340
> +// { dg-do compile { target c++14 } }
> +
> +template<class>
> +struct A {
> +  template<class>
> +  static constexpr const int& var = 0;
> +};
> +
> +template<class T>
> +struct B {
> +  static constexpr int x1 = A<T>::template var<int>;
> +  static constexpr int y1 = A<T>{}.template var<int>;
> +
> +  static constexpr int x2 = A<int>::template var<T>;
> +  static constexpr int y2 = A<int>{}.template var<T>;
> +
> +  static constexpr int x3 = A<int>::template var<int>;
> +  static constexpr int y3 = A<int>{}.template var<int>;
> +};
> +
> +template struct B<int>;
> diff --git a/gcc/testsuite/g++.dg/cpp1y/var-templ81.C b/gcc/testsuite/g++.dg/cpp1y/var-templ81.C
> new file mode 100644
> index 00000000000..f9d2e6b1eed
> --- /dev/null
> +++ b/gcc/testsuite/g++.dg/cpp1y/var-templ81.C
> @@ -0,0 +1,14 @@
> +// Verify we don't ICE on an invalid use of unary * for a variable
> +// template-id of reference type.
> +// { dg-do compile { target c++14 } }
> +
> +template<class>
> +static constexpr const int& var = 0;
> +
> +template<class T>
> +struct B {
> +  static constexpr int x = *var<T>; // { dg-error "argument of unary" }
> +  static constexpr const int& y = *var<T>; // { dg-error "argument of unary" }
> +};
> +
> +template struct B<int>;


      reply	other threads:[~2023-05-18 18:57 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-05-18 17:59 Patrick Palka
2023-05-18 18:56 ` Jason Merrill [this message]

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=f789ca2d-56b9-d6b1-9baa-99b2def9e8b7@redhat.com \
    --to=jason@redhat.com \
    --cc=gcc-patches@gcc.gnu.org \
    --cc=ppalka@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).