public inbox for gcc-patches@gcc.gnu.org
 help / color / mirror / Atom feed
From: Patrick Palka <ppalka@redhat.com>
To: gcc-patches@gcc.gnu.org
Subject: [PATCH] c++: CTAD within template argument [PR102933]
Date: Tue, 26 Oct 2021 13:44:09 -0400	[thread overview]
Message-ID: <20211026174409.3142598-1-ppalka@redhat.com> (raw)

Here when checking for erroneous occurrences of 'auto' inside a template
argument (which is allowed by the concepts TS for class templates),
extract_autos_r picks up the CTAD placeholder for X{T{0}} which causes
check_auto_in_tmpl_args to reject this valid template argument.  This
patch fixes this by making extract_autos_r ignore CTAD placeholders.

However, it seems we don't need to call check_auto_in_tmpl_args at all
outside of the concepts TS since using 'auto' as a type-id is otherwise
rejected more generally at parse time.  So this patch guards calls to
check_auto_in_tmpl_args with flag_concepts_ts instead of flag_concepts.

Relatedly, I think the concepts code paths in do_auto_deduction and
type_uses_auto are also necessary only for the concepts TS, so this
patch also restricts these code paths accordingly.

Bootstrapped and regtested on x86_64-pc-linux-gnu, does this look OK for
trunk and perhaps 11?

	PR c++/102933

gcc/cp/ChangeLog:

	* parser.c (cp_parser_template_id): Call check_auto_in_tmpl_args
	only for the concepts TS not also for standard concepts.
	(cp_parser_simple_type_specifier): Adjust diagnostic for using
	auto in parameter declaration.
	* pt.c (tsubst_qualified_id): Call check_auto_in_tmpl_args only
	for the concepts TS not also for standard concepts.
	(extract_autos_r): Ignore CTAD placeholders.
	(extract_autos): Use range-based for.
	(do_auto_deduction): Use extract_autos only for the concepts TS
	and not also for standard concepts.
	(type_uses_auto): Likewise with for_each_template_parm.
	(check_auto_in_tmpl_args): Assert that this function is used only
	for the concepts tS.  Simplify.

gcc/testsuite/ChangeLog:

	* g++.dg/cpp2a/nontype-class50.C: New test.
	* g++.dg/cpp2a/nontype-class50a.C: New test.
---
 gcc/cp/parser.c                               |  4 ++--
 gcc/cp/pt.c                                   | 24 +++++++++----------
 gcc/testsuite/g++.dg/cpp2a/nontype-class50.C  | 13 ++++++++++
 gcc/testsuite/g++.dg/cpp2a/nontype-class50a.C |  5 ++++
 4 files changed, 32 insertions(+), 14 deletions(-)
 create mode 100644 gcc/testsuite/g++.dg/cpp2a/nontype-class50.C
 create mode 100644 gcc/testsuite/g++.dg/cpp2a/nontype-class50a.C

diff --git a/gcc/cp/parser.c b/gcc/cp/parser.c
index 49d951cfb19..5052f534d40 100644
--- a/gcc/cp/parser.c
+++ b/gcc/cp/parser.c
@@ -18168,7 +18168,7 @@ cp_parser_template_id (cp_parser *parser,
      types.  We reject them in functions, but if what we have is an
      identifier, even with none_type we can't conclude it's NOT a
      type, we have to wait for template substitution.  */
-  if (flag_concepts && check_auto_in_tmpl_args (templ, arguments))
+  if (flag_concepts_ts && check_auto_in_tmpl_args (templ, arguments))
     template_id = error_mark_node;
   /* Build a representation of the specialization.  */
   else if (identifier_p (templ))
@@ -19505,7 +19505,7 @@ cp_parser_simple_type_specifier (cp_parser* parser,
 	  else if (!flag_concepts)
 	    pedwarn (token->location, 0,
 		     "use of %<auto%> in parameter declaration "
-		     "only available with %<-fconcepts-ts%>");
+		     "only available with %<-std=c++20%> or %<-fconcepts%>");
 	}
       else
 	type = make_auto ();
diff --git a/gcc/cp/pt.c b/gcc/cp/pt.c
index 287cf4ce9d0..3321601e6ff 100644
--- a/gcc/cp/pt.c
+++ b/gcc/cp/pt.c
@@ -16456,7 +16456,7 @@ tsubst_qualified_id (tree qualified_id, tree args,
 	 we want to catch is when we couldn't tell then, and can now,
 	 namely when templ prior to substitution was an
 	 identifier.  */
-      if (flag_concepts && check_auto_in_tmpl_args (expr, template_args))
+      if (flag_concepts_ts && check_auto_in_tmpl_args (expr, template_args))
 	return error_mark_node;
 
       if (variable_template_p (expr))
@@ -28560,7 +28560,7 @@ static int
 extract_autos_r (tree t, void *data)
 {
   hash_table<auto_hash> &hash = *(hash_table<auto_hash>*)data;
-  if (is_auto (t))
+  if (is_auto (t) && !template_placeholder_p (t))
     {
       /* All the autos were built with index 0; fix that up now.  */
       tree *p = hash.find_slot (t, INSERT);
@@ -28594,10 +28594,8 @@ extract_autos (tree type)
   for_each_template_parm (type, extract_autos_r, &hash, &visited, true);
 
   tree tree_vec = make_tree_vec (hash.elements());
-  for (hash_table<auto_hash>::iterator iter = hash.begin();
-       iter != hash.end(); ++iter)
+  for (tree elt : hash)
     {
-      tree elt = *iter;
       unsigned i = TEMPLATE_PARM_IDX (TEMPLATE_TYPE_PARM_INDEX (elt));
       TREE_VEC_ELT (tree_vec, i)
 	= build_tree_list (NULL_TREE, TYPE_NAME (elt));
@@ -29837,7 +29835,7 @@ do_auto_deduction (tree type, tree init, tree auto_node,
       tree parms = build_tree_list (NULL_TREE, type);
       tree tparms;
 
-      if (flag_concepts)
+      if (flag_concepts_ts)
 	tparms = extract_autos (type);
       else
 	{
@@ -30025,7 +30023,7 @@ type_uses_auto (tree type)
 {
   if (type == NULL_TREE)
     return NULL_TREE;
-  else if (flag_concepts)
+  else if (flag_concepts_ts)
     {
       /* The Concepts TS allows multiple autos in one type-specifier; just
 	 return the first one we find, do_auto_deduction will collect all of
@@ -30048,6 +30046,9 @@ type_uses_auto (tree type)
 bool
 check_auto_in_tmpl_args (tree tmpl, tree args)
 {
+  /* Only the concepts TS allows 'auto' as a type-id.  */
+  gcc_assert (flag_concepts_ts);
+
   /* If there were previous errors, nevermind.  */
   if (!args || TREE_CODE (args) != TREE_VEC)
     return false;
@@ -30057,11 +30058,10 @@ check_auto_in_tmpl_args (tree tmpl, tree args)
      We'll only be able to tell during template substitution, so we
      expect to be called again then.  If concepts are enabled and we
      know we have a type, we're ok.  */
-  if (flag_concepts
-      && (identifier_p (tmpl)
-	  || (DECL_P (tmpl)
-	      &&  (DECL_TYPE_TEMPLATE_P (tmpl)
-		   || DECL_TEMPLATE_TEMPLATE_PARM_P (tmpl)))))
+  if (identifier_p (tmpl)
+      || (DECL_P (tmpl)
+	  &&  (DECL_TYPE_TEMPLATE_P (tmpl)
+	       || DECL_TEMPLATE_TEMPLATE_PARM_P (tmpl))))
     return false;
 
   /* Quickly search for any occurrences of auto; usually there won't
diff --git a/gcc/testsuite/g++.dg/cpp2a/nontype-class50.C b/gcc/testsuite/g++.dg/cpp2a/nontype-class50.C
new file mode 100644
index 00000000000..1c2786a59e7
--- /dev/null
+++ b/gcc/testsuite/g++.dg/cpp2a/nontype-class50.C
@@ -0,0 +1,13 @@
+// PR c++/102933
+// { dg-do compile { target c++20 } }
+
+template<class T> struct X { T t; };
+
+template<X> void f();
+
+template<class T>
+void g() {
+  f<X{T{0}}>();
+}
+
+template void g<int>();
diff --git a/gcc/testsuite/g++.dg/cpp2a/nontype-class50a.C b/gcc/testsuite/g++.dg/cpp2a/nontype-class50a.C
new file mode 100644
index 00000000000..eb8a6ad9375
--- /dev/null
+++ b/gcc/testsuite/g++.dg/cpp2a/nontype-class50a.C
@@ -0,0 +1,5 @@
+// PR c++/102933
+// { dg-do compile { target c++20 } }
+// { dg-additional-options "-fconcepts-ts" }
+
+#include "nontype-class50.C"
-- 
2.33.1.711.g9d530dc002


             reply	other threads:[~2021-10-26 17:44 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-10-26 17:44 Patrick Palka [this message]
2021-10-27 20:51 ` Jason Merrill
2021-10-28 16:36   ` Patrick Palka

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=20211026174409.3142598-1-ppalka@redhat.com \
    --to=ppalka@redhat.com \
    --cc=gcc-patches@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).