From: "Sergei Trofimovich" <slyfox@gentoo.org>
To: gentoo-commits@lists.gentoo.org
Subject: [gentoo-commits] proj/gcc-patches:master commit in: 10.1.0/gentoo/
Date: Wed, 24 Jun 2020 20:15:49 +0000 (UTC) [thread overview]
Message-ID: <1593029682.21ed34524e3ab3139c811b6d61f50522ebd2b184.slyfox@gentoo> (raw)
commit: 21ed34524e3ab3139c811b6d61f50522ebd2b184
Author: Sergei Trofimovich <slyfox <AT> gentoo <DOT> org>
AuthorDate: Wed Jun 24 20:14:42 2020 +0000
Commit: Sergei Trofimovich <slyfox <AT> gentoo <DOT> org>
CommitDate: Wed Jun 24 20:14:42 2020 +0000
URL: https://gitweb.gentoo.org/proj/gcc-patches.git/commit/?id=21ed3452
10.1.0: backport PR95508, ICE on array subscript implicit conversion
Reported-by: hsk17 <AT> mail.de
Bug: https://bugs.gentoo.org/729434
Signed-off-by: Sergei Trofimovich <slyfox <AT> gentoo.org>
.../35_all_ICE-array-subscript-PR95508.patch | 106 +++++++++++++++++++++
10.1.0/gentoo/README.history | 1 +
2 files changed, 107 insertions(+)
diff --git a/10.1.0/gentoo/35_all_ICE-array-subscript-PR95508.patch b/10.1.0/gentoo/35_all_ICE-array-subscript-PR95508.patch
new file mode 100644
index 0000000..37abd53
--- /dev/null
+++ b/10.1.0/gentoo/35_all_ICE-array-subscript-PR95508.patch
@@ -0,0 +1,106 @@
+https://gcc.gnu.org/PR95508
+https://bugs.gentoo.org/729434
+
+From 1bab254fd30c2b94a675b9057349fc80946375b1 Mon Sep 17 00:00:00 2001
+From: Marek Polacek <polacek@redhat.com>
+Date: Wed, 17 Jun 2020 14:38:05 -0400
+Subject: [PATCH] c++: ICE with IMPLICIT_CONV_EXPR in array subscript [PR95508]
+
+Since r10-7096 convert_like, when called in a template, creates an
+IMPLICIT_CONV_EXPR when we're converting to/from array type.
+
+In this test, we have e[f], and we're converting f (of type class A) to
+int, so convert_like in build_new_op_1 created the IMPLICIT_CONV_EXPR
+that got into cp_build_array_ref which calls maybe_constant_value. My
+patch above failed to adjust this spot to call fold_non_dependent_expr
+instead, which can handle codes like I_C_E in a template. Fixed by
+using a new function maybe_fold_non_dependent_expr, which, if the expr
+can't be evaluated to a constant, returns the original expression.
+
+gcc/cp/ChangeLog:
+
+ PR c++/95508
+ * constexpr.c (maybe_fold_non_dependent_expr): New.
+ * cp-tree.h (maybe_fold_non_dependent_expr): Declare.
+ * typeck.c (cp_build_array_ref): Call maybe_fold_non_dependent_expr
+ instead of maybe_constant_value.
+
+gcc/testsuite/ChangeLog:
+
+ PR c++/95508
+ * g++.dg/template/conv16.C: New test.
+---
+ gcc/cp/constexpr.c | 13 +++++++++++++
+ gcc/cp/cp-tree.h | 2 ++
+ gcc/cp/typeck.c | 2 +-
+ gcc/testsuite/g++.dg/template/conv16.C | 17 +++++++++++++++++
+ 4 files changed, 33 insertions(+), 1 deletion(-)
+ create mode 100644 gcc/testsuite/g++.dg/template/conv16.C
+
+--- a/gcc/cp/constexpr.c
++++ b/gcc/cp/constexpr.c
+@@ -7043,6 +7043,19 @@ fold_non_dependent_expr (tree t,
+ return maybe_constant_value (t, object, manifestly_const_eval);
+ }
+
++/* Like fold_non_dependent_expr, but if EXPR couldn't be folded to a constant,
++ return the original expression. */
++
++tree
++maybe_fold_non_dependent_expr (tree expr,
++ tsubst_flags_t complain/*=tf_warning_or_error*/)
++{
++ tree t = fold_non_dependent_expr (expr, complain);
++ if (t && TREE_CONSTANT (t))
++ return t;
++
++ return expr;
++}
+
+ /* Like maybe_constant_init but first fully instantiate the argument. */
+
+--- a/gcc/cp/cp-tree.h
++++ b/gcc/cp/cp-tree.h
+@@ -7955,6 +7955,8 @@ extern tree maybe_constant_init (tree, tree = NULL_TREE, bool = false);
+ extern tree fold_non_dependent_expr (tree,
+ tsubst_flags_t = tf_warning_or_error,
+ bool = false, tree = NULL_TREE);
++extern tree maybe_fold_non_dependent_expr (tree,
++ tsubst_flags_t = tf_warning_or_error);
+ extern tree fold_non_dependent_init (tree,
+ tsubst_flags_t = tf_warning_or_error,
+ bool = false);
+--- a/gcc/cp/typeck.c
++++ b/gcc/cp/typeck.c
+@@ -3553,7 +3553,7 @@ cp_build_array_ref (location_t loc, tree array, tree idx,
+ pointer arithmetic.) */
+ idx = cp_perform_integral_promotions (idx, complain);
+
+- idx = maybe_constant_value (idx);
++ idx = maybe_fold_non_dependent_expr (idx, complain);
+
+ /* An array that is indexed by a non-constant
+ cannot be stored in a register; we must be able to do
+--- /dev/null
++++ b/gcc/testsuite/g++.dg/template/conv16.C
+@@ -0,0 +1,17 @@
++// PR c++/95508
++// { dg-do compile }
++
++template <typename>
++struct A;
++template <typename>
++struct B {
++ operator int () { return 0; }
++};
++template <>
++struct A<unsigned> : B<int> {};
++struct D {
++ template <typename>
++ int foo () { return e[f]; }
++ int e[6];
++ A<unsigned> f;
++};
+--
+2.27.0
+
diff --git a/10.1.0/gentoo/README.history b/10.1.0/gentoo/README.history
index 480a54e..72e0086 100644
--- a/10.1.0/gentoo/README.history
+++ b/10.1.0/gentoo/README.history
@@ -1,6 +1,7 @@
3 TODO
33_all_avx512-scalar-PR95528.patch
34_all_cet-cross-x86.patch
+ 35_all_ICE-array-subscript-PR95508.patch
2 11 June 2020
+ 29_all_fix-float-hang-PR95118.patch
next reply other threads:[~2020-06-24 20:15 UTC|newest]
Thread overview: 15+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-06-24 20:15 Sergei Trofimovich [this message]
-- strict thread matches above, loose matches on Subject: below --
2020-07-23 8:50 [gentoo-commits] proj/gcc-patches:master commit in: 10.1.0/gentoo/ Sergei Trofimovich
2020-07-19 21:35 Sergei Trofimovich
2020-07-14 7:17 Sergei Trofimovich
2020-07-04 7:56 Sergei Trofimovich
2020-07-04 7:56 Sergei Trofimovich
2020-07-03 22:12 Sergei Trofimovich
2020-06-23 6:34 Sergei Trofimovich
2020-06-14 11:07 Sergei Trofimovich
2020-06-11 23:00 Sergei Trofimovich
2020-06-11 22:49 Sergei Trofimovich
2020-06-05 21:02 Sergei Trofimovich
2020-05-19 22:02 Sergei Trofimovich
2020-05-14 17:58 Sergei Trofimovich
2020-05-07 20:51 Sergei Trofimovich
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=1593029682.21ed34524e3ab3139c811b6d61f50522ebd2b184.slyfox@gentoo \
--to=slyfox@gentoo.org \
--cc=gentoo-commits@lists.gentoo.org \
--cc=gentoo-dev@lists.gentoo.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