libstdc++: Fix non-portable results from 64-bit std::subtract_with_carry_engine [PR107466]

Message ID 20240111230548.623208-1-jwakely@redhat.com
State Accepted
Headers
Series libstdc++: Fix non-portable results from 64-bit std::subtract_with_carry_engine [PR107466] |

Checks

Context Check Description
snail/gcc-patch-check success Github commit url

Commit Message

Jonathan Wakely Jan. 11, 2024, 11:04 p.m. UTC
  This fixes a regression introduced by the LWG 3809 change, so is needed
on trunk and gcc-13 and gcc-12.

Tested x86_64-linux and aarch64-linux.

-- >8 --

I implemented the resolution of LWG 3809 in r13-4364-ga64775a0edd469 but
more recently it was noted that the change causes possible truncation
for 64-bit seeds. Whether the truncation occurs (and to what value)
depends on the width of uint_least32_t which is not portable. The new
problem was filed as LWG 4014. I proposed a new change which reduces the
seed by the LCG's modulus before the conversion to uint_least32_t. This
ensures a consistent result across paltforms, and restores the old
behaviour for std::subtract_with_carry_engine specializations using a
64-bit result type (such as std::ranlux48_base).

libstdc++-v3/ChangeLog:

	PR libstdc++/107466
	* include/bits/random.tcc (subtract_with_carry_engine::seed):
	Implement proposed resolution of LWG 4014.
	* testsuite/26_numerics/random/pr60037-neg.cc: Adjust dg-error
	line number.
	* testsuite/26_numerics/random/subtract_with_carry_engine/cons/lwg3809.cc:
	Check for expected result of 64-bit engine with seed that
	doesn't fit in 32-bits.
---
 libstdc++-v3/include/bits/random.tcc          |  5 ++++-
 .../26_numerics/random/pr60037-neg.cc         |  2 +-
 .../cons/lwg3809.cc                           | 19 +++++++++++++++++--
 3 files changed, 22 insertions(+), 4 deletions(-)
  

Patch

diff --git a/libstdc++-v3/include/bits/random.tcc b/libstdc++-v3/include/bits/random.tcc
index 7f4bf5ea183..ade416390b3 100644
--- a/libstdc++-v3/include/bits/random.tcc
+++ b/libstdc++-v3/include/bits/random.tcc
@@ -541,8 +541,11 @@  _GLIBCXX_BEGIN_NAMESPACE_VERSION
     subtract_with_carry_engine<_UIntType, __w, __s, __r>::
     seed(result_type __value)
     {
+      // _GLIBCXX_RESOLVE_LIB_DEFECTS
+      // 3809. Is std::subtract_with_carry_engine<uint16_t> supposed to work?
+      // 4014. LWG 3809 changes behavior of some existing code
       std::linear_congruential_engine<uint_least32_t, 40014u, 0u, 2147483563u>
-	__lcg(__value == 0u ? default_seed : __value);
+	__lcg(__value == 0u ? default_seed : __value % 2147483563u);
 
       const size_t __n = (__w + 31) / 32;
 
diff --git a/libstdc++-v3/testsuite/26_numerics/random/pr60037-neg.cc b/libstdc++-v3/testsuite/26_numerics/random/pr60037-neg.cc
index c58f480640f..59cf84adb48 100644
--- a/libstdc++-v3/testsuite/26_numerics/random/pr60037-neg.cc
+++ b/libstdc++-v3/testsuite/26_numerics/random/pr60037-neg.cc
@@ -12,4 +12,4 @@  auto x = std::generate_canonical<std::size_t,
 
 // { dg-error "static assertion failed: template argument must be a floating point type" "" { target *-*-* } 169 }
 
-// { dg-error "static assertion failed: template argument must be a floating point type" "" { target *-*-* } 3348 }
+// { dg-error "static assertion failed: template argument must be a floating point type" "" { target *-*-* } 3351 }
diff --git a/libstdc++-v3/testsuite/26_numerics/random/subtract_with_carry_engine/cons/lwg3809.cc b/libstdc++-v3/testsuite/26_numerics/random/subtract_with_carry_engine/cons/lwg3809.cc
index d91ee7448f6..b6fb57f8eeb 100644
--- a/libstdc++-v3/testsuite/26_numerics/random/subtract_with_carry_engine/cons/lwg3809.cc
+++ b/libstdc++-v3/testsuite/26_numerics/random/subtract_with_carry_engine/cons/lwg3809.cc
@@ -2,10 +2,11 @@ 
 #include <random>
 #include <testsuite_hooks.h>
 
-// LWG 3809. Is std::subtract_with_carry_engine<uint16_t> supposed to work?
 // PR 107466 - invalid -Wnarrowing error with std::subtract_with_carry_engine
 
-int main()
+// LWG 3809. Is std::subtract_with_carry_engine<uint16_t> supposed to work?
+void
+test_lwg3809()
 {
   // It should be possible to construct this engine with a 16-bit result_type:
   std::subtract_with_carry_engine<uint16_t, 12, 5, 12> s16;
@@ -24,3 +25,17 @@  int main()
   for (int i = 0; i < 10; ++i)
     VERIFY( s16() == s32() );
 }
+
+// LWG 4014. LWG 3809 changes behavior of some existing code
+void
+test_lwg4014()
+{
+  std::ranlux48_base g(-1U + 1LL);
+  VERIFY( g() == 22575453646312 );
+}
+
+int main()
+{
+  test_lwg3809();
+  test_lwg4014();
+}