public inbox for gcc-patches@gcc.gnu.org
 help / color / mirror / Atom feed
From: Jonathan Wakely <jwakely@redhat.com>
To: libstdc++@gcc.gnu.org, gcc-patches@gcc.gnu.org
Subject: [committed] libstdc++: Fix regression in std::stable_sort
Date: Wed, 24 Aug 2022 15:33:39 +0100	[thread overview]
Message-ID: <20220824143339.121092-1-jwakely@redhat.com> (raw)

Tested powerpc64le-linux, pushed to trunk.

No backport needed as the regression is only on trunk.

-- >8 --

The recent change to split out the cold path of std::stable_sort caused
a regression for some Qt code. The problem is that the library now adds
a value of type ptrdiff_t to the iterator, which is ambiguous with
-pedantic. The addition could either convert the iterator to a built-in
pointer and add the ptrdiff_t to that, or it could convert the ptrdiff_t
to the iterator's difference_type and use the iterator's own operator+.

The fix is to cast the ptrdiff_t value to the difference type first.

libstdc++-v3/ChangeLog:

	* include/bits/stl_algo.h (__stable_sort): Cast size to
	iterator's difference type.
	* testsuite/25_algorithms/stable_sort/4.cc: New test.
---
 libstdc++-v3/include/bits/stl_algo.h          |  5 +-
 .../testsuite/25_algorithms/stable_sort/4.cc  | 51 +++++++++++++++++++
 2 files changed, 54 insertions(+), 2 deletions(-)
 create mode 100644 libstdc++-v3/testsuite/25_algorithms/stable_sort/4.cc

diff --git a/libstdc++-v3/include/bits/stl_algo.h b/libstdc++-v3/include/bits/stl_algo.h
index c6078054514..57fa1c1dc55 100644
--- a/libstdc++-v3/include/bits/stl_algo.h
+++ b/libstdc++-v3/include/bits/stl_algo.h
@@ -5026,8 +5026,9 @@ _GLIBCXX_BEGIN_NAMESPACE_ALGO
       _TmpBuf __buf(__first, (__last - __first + 1) / 2);
 
       if (__builtin_expect(__buf.requested_size() == __buf.size(), true))
-	std::__stable_sort_adaptive(__first, __first + __buf.size(), __last,
-				    __buf.begin(), __comp);
+	std::__stable_sort_adaptive(__first,
+				    __first + _DistanceType(__buf.size()),
+				    __last, __buf.begin(), __comp);
       else if (__builtin_expect(__buf.begin() == 0, false))
 	std::__inplace_stable_sort(__first, __last, __comp);
       else
diff --git a/libstdc++-v3/testsuite/25_algorithms/stable_sort/4.cc b/libstdc++-v3/testsuite/25_algorithms/stable_sort/4.cc
new file mode 100644
index 00000000000..b7bda4eeaca
--- /dev/null
+++ b/libstdc++-v3/testsuite/25_algorithms/stable_sort/4.cc
@@ -0,0 +1,51 @@
+// { dg-options "-pedantic" }
+// { dg-do compile }
+
+#include <algorithm>
+
+/* This type is reduced from QTypedArrayData::iterator which has an implicit
+ * conversion to its pointer type and a difference type of int.
+ * The expression Iter() + ptrdiff_t(0) is ambiguous with -pedantic because it
+ * could either convert the RHS to int and use Iter::operator+(int)
+ * or it could convert the LHS to pointer and use built-in pointer arithmetic.
+ */
+struct Iter
+{
+  struct value_type { bool operator<(value_type) const; };
+  typedef value_type* pointer;
+  typedef value_type& reference;
+  typedef std::random_access_iterator_tag iterator_category;
+  typedef int difference_type;
+
+  reference operator*() const;
+  pointer operator->() const;
+
+  reference operator[](difference_type) const;
+
+  Iter& operator++();
+  Iter& operator--();
+  Iter operator++(int);
+  Iter operator--(int);
+
+  Iter& operator+=(difference_type);
+  Iter& operator-=(difference_type);
+
+  Iter operator+(difference_type) const;
+  Iter operator-(difference_type) const;
+
+  difference_type operator-(Iter) const;
+
+  operator pointer() const; // XXX this causes the ambiguity
+
+  bool operator==(Iter) const;
+  bool operator!=(Iter) const;
+
+  bool operator<(Iter) const;
+};
+
+Iter operator+(Iter::difference_type, Iter);
+
+int main()
+{
+  std::stable_sort(Iter(), Iter());
+}
-- 
2.37.2


                 reply	other threads:[~2022-08-24 14:33 UTC|newest]

Thread overview: [no followups] expand[flat|nested]  mbox.gz  Atom feed

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=20220824143339.121092-1-jwakely@redhat.com \
    --to=jwakely@redhat.com \
    --cc=gcc-patches@gcc.gnu.org \
    --cc=libstdc++@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).