public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
From: "frank.marschner at cognitec dot com" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug libstdc++/105502] New: std::normal_distribution deserialization issue
Date: Fri, 06 May 2022 08:28:01 +0000	[thread overview]
Message-ID: <bug-105502-4@http.gcc.gnu.org/bugzilla/> (raw)

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105502

            Bug ID: 105502
           Summary: std::normal_distribution deserialization issue
           Product: gcc
           Version: 9.4.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: libstdc++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: frank.marschner at cognitec dot com
  Target Milestone: ---

Found an issue deserializing an serialized instance of
std::normal_distribution.
Depending on the amount of random numbers generated deserialization
works or not.

The issue was introduced with commit
160e95dc3d73329d2367fb405ef9f0e12bd2cc7b on file bits/random.tcc from Jan 9,
2020.

The issue is in the implementation of

template< class CharT, class Traits >
friend std::basic_istream<CharT,Traits>&
    operator>>( std::basic_istream<CharT,Traits>& ist,
                normal_distribution& d );

The commit introduces wrong initialization of members param (mean and
stddev) and also _M_saved_available depending on __saved_avail.


The following example demonstrates the issue
---8<----------------------------------------------------------------
#include <iostream>
#include <sstream>
#include <random>

void test_serialization( const std::normal_distribution<>& nd) 
{
   std::stringstream sss;
   sss << nd;

   std::normal_distribution<> newND;
   sss >> newND;

   if( nd == newND) {
       std::cout << "Serialization successful." << std::endl;
   } else {
       std::cout << "Serialization not successful." << std::endl;
   }
}

int main( int argc, char** argv) 
{
   std::mt19937 gen{ 42};

   std::normal_distribution<> d{ 4, 2};

   test_serialization( d);
   const auto result0 = d(gen);
   test_serialization( d);
   const auto result1 = d(gen);
   test_serialization( d);
   const auto result2 = d(gen);
   test_serialization( d);
}
---8<----------------------------------------------------------------

The output running the code shows

Serialization not successful.
Serialization successful.
Serialization not successful.
Serialization successful.

The issues covers gcc versions since 8.4 up to recent.

             reply	other threads:[~2022-05-06  8:28 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-05-06  8:28 frank.marschner at cognitec dot com [this message]
2022-05-06 14:35 ` [Bug libstdc++/105502] [9/12/11/12/13 Regression] " redi at gcc dot gnu.org
2022-05-06 20:51 ` [Bug libstdc++/105502] [9/10/11/12/13 " redi at gcc dot gnu.org
2022-05-06 22:55 ` cvs-commit at gcc dot gnu.org
2022-05-06 22:59 ` cvs-commit at gcc dot gnu.org
2022-05-06 22:59 ` [Bug libstdc++/105502] [9/10/11 " redi at gcc dot gnu.org
2022-05-09 15:09 ` cvs-commit at gcc dot gnu.org
2022-05-09 16:35 ` cvs-commit at gcc dot gnu.org
2022-05-09 16:44 ` cvs-commit at gcc dot gnu.org
2022-05-09 16:44 ` redi at gcc dot gnu.org

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=bug-105502-4@http.gcc.gnu.org/bugzilla/ \
    --to=gcc-bugzilla@gcc.gnu.org \
    --cc=gcc-bugs@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).