Re: libstdc++/6718: Boost1.28 run-time problem

Re: libstdc++/6718: Boost1.28 run-time problem

[paolo]

Synopsis: Boost1.28 run-time problem

State-Changed-From-To: analyzed->closed
State-Changed-By: paolo
State-Changed-When: Thu Oct 31 09:10:31 2002
State-Changed-Why:
    I'm closing this. In fact current mainline is not affected
    by this specific problem (another PR will be probably filed
    for a different abort, happening later at run time)
    and this is not a regression from 3.0 and 3.1. However, I have confirmed when the test is compiled by 2.95.x the
    early abort does not happen: something may be wrong with
    either v3 or the boost/random, we don't know, right now.
    I will try to isolate the problem in a small testcase.
    Ciao, Paolo.

http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=6718

Re: libstdc++/6718: Boost1.28 run-time problem

[Paolo Carlini]

The following reply was made to PR libstdc++/6718; it has been noted by GNATS.

From: Paolo Carlini <pcarlini@unitus.it>
To: gdr@gcc.gnu.org,  gcc-gnats@gcc.gnu.org,  gcc-prs@gcc.gnu.org, 
 pcarlini@unitus.it,  gcc-bugs@gcc.gnu.org
Cc:  
Subject: Re: libstdc++/6718: Boost1.28 run-time problem
Date: Mon, 28 Oct 2002 20:50:38 +0100

 For the time being I only reduce the priority to "medium".
 
 Before completely closing the PR I'd like to be
 100% sure that:
 1- The issue is known to the Boosters.
 2- Current mainline is not affected.
 
 Paolo.
 
 http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=6718
 

Re: libstdc++/6718: Boost1.28 run-time problem

[Gabriel Dos Reis]

The following reply was made to PR libstdc++/6718; it has been noted by GNATS.

From: Gabriel Dos Reis <gdr@integrable-solutions.net>
To: paolo@gcc.gnu.org
Cc: gcc-gnats@gcc.gnu.org
Subject: Re: libstdc++/6718: Boost1.28 run-time problem
Date: 28 Oct 2002 02:41:54 +0100

 paolo@gcc.gnu.org writes:
 
 | Synopsis: Boost1.28 run-time problem
 | 
 | Responsible-Changed-From-To: unassigned->gdr
 | Responsible-Changed-By: paolo
 | Responsible-Changed-When: Sun Oct 27 13:55:21 2002
 | Responsible-Changed-Why:
 |     Hi Gaby. Nathan ascertained that this serious run-time
 |     problem is in fact due to a <limits> issue.
 
 std::numeric_limits<> is known to provide only partial support for
 floating point types since the beginning.  Note that
 std::numeric_limits<>::has_infinity() is false. 
 
 My trying to provide better support is what prompted RTH to rewrite
 real.c.  I have no plan to back-port that rewrite to branch.
 
 This is not a regression.
 
 As far as I'm concerned, this PR should be closed.
 
 -- Gaby

Re: libstdc++/6718: Boost1.28 run-time problem

[paolo]

Synopsis: Boost1.28 run-time problem

Responsible-Changed-From-To: unassigned->gdr
Responsible-Changed-By: paolo
Responsible-Changed-When: Sun Oct 27 13:55:21 2002
Responsible-Changed-Why:
    Hi Gaby. Nathan ascertained that this serious run-time
    problem is in fact due to a <limits> issue. Could you
    possibly look into it?
    Thanks in advance, Paolo.

http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=6718

Re: libstdc++/6718: Boost1.28 run-time problem

[Nathan Sidwell]

The following reply was made to PR libstdc++/6718; it has been noted by GNATS.

From: Nathan Sidwell <nathan@codesourcery.com>
To: gcc-gnats@gcc.gnu.org, gcc-prs@gcc.gnu.org, pcarlini@unitus.it,
   gcc-bugs@gcc.gnu.org
Cc:  
Subject: Re: libstdc++/6718: Boost1.28 run-time problem
Date: Fri, 25 Oct 2002 22:36:05 +0100

 http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=6718
 
 this is a libstdc++ issue. <limits> appears to specialize
 template<> struct numeric_limits<double> to
 
      static double infinity() throw()
        { return __glibcpp_double_infinity; }
 
 and earlier,
 #ifndef __glibcpp_double_infinity
 #define __glibcpp_double_infinity 0.0
 #endif
 
 on the mainline this is significantly different, using __builtin_huge_val.
 
 nathan