Created attachment 676404 [details] minimal autoconf scripts for generating error (boost_test_minimal.3.tar.bz2). Description of problem: On RHEL 5.8 compiling against boost141 worked, on RHEL 5.9 compiling against boost141 fails. Version-Release number of selected component (if applicable): boost141-devel-1.41.0-2.el5 boost141-thread-1.41.0-2.el5 basically boost141*-1.41.0-2.el5 How reproducible: Reliably broken on RHEL 5.9. Steps to Reproduce: 1. 2. 3. Actual results: configure:20226: g++ -c -g -O2 -pthread -I/usr/include/boost141 conftest.cc >&5 /usr/include/boost141/boost/exception_ptr.hpp:43: error: looser throw specifier for 'virtual boost::exception_ptr::~exception_ptr()' /usr/include/boost141/boost/exception/detail/exception_ptr_base.hpp:26: error: overriding 'virtual boost::exception_detail::exception_ptr_base::~exception_ptr_base() throw ()' Expected results: No errors on compile, configure completes. Additional info: https://svn.boost.org/trac/boost/ticket/3641 The following change set fixes the issue https://svn.boost.org/trac/boost/changeset/57764
It took my return on adding the attachment as Submit... Steps to Reproduce: 1. tar -xjvf boost_test_minimal.3.tar.bz2 2. cd boost_test 3. ./autogen.sh ; ./configure 4. see "checking boost/thread.hpp usability... no checking boost/thread.hpp presence... yes configure: WARNING:..." 5. grep -n -5 exception_ptr config.log Note: from the boost r57764, only the exception.hpp and exception_ptr.hpp need fixed. i.e. (for testing purposes) cd /usr/include/boost141 patch -p 3 < ~/RHEL5.9issue/changeset_57764_RHEL5.diff and the issue is fixed.
Created attachment 676407 [details] changeset_57764_RHEL5.diff
Todd, does this issue also exist on RHEL 6 - which this package is based on?
(In reply to comment #3) > Todd, does this issue also exist on RHEL 6 - which this package is based on? Unfortunately I don't run any RHEL 6 systems, so I can't say. It should be possible to replace the boost.m4 file, because the one In my tarball has some stuff to make it work on RHEL5, with a current one from http://github.com/tsuna/boost.m4 and try the set on an RHEL 6 system too.
I am assuming that the breakage was cause either by an include file change between RHEL5.8 and RHEL5.9 or by one of the compiler changes that occurred at the same time.
I have been working with a second 5.8 machine, and slowly updating it. results: Of all the things that got updated between last 5.8 fix through 5.9 release, the only thing not yet updated and the system still work as expected was gcc (and the redhat-release packages), that is going from gcc 4.1.2-52.el5_8.1 to gcc 4.1.2-54.el5 updated the following list of packages: gcc cpp gcc-c++ gcc-gfortran gcc-gnat gcc-objc libgcc libgfortran libgnat libmudflap libmudflap-devel libobjc libstdc++ libstdc++-devel My quick delta of the include directory files from each rpm did not show any differences, so I am thinking it was a processing change in one of cpp, gcc or gcc-c++. Unfortunately that 5.8 machine had to walk out the door for the day.
Looking at the 5.9 release notes for GCC https://access.redhat.com/knowledge/docs/en-US/Red_Hat_Enterprise_Linux/5/html/5.9_Technical_Notes/gcc.html we see BZ#806394 BZ#750545 and BZ#806275 which are all related to exception handling. Unfortunately all of those tickets are still sequestered (security reasons? still valid for RHEL6?) and I only see "You are not authorized to access bug #______". Perhaps there is hope that when those are fixed in RHEL6 you'll see the issue there too? :)
I'm also seeing this problem on my RHEL 5.9 machines (both i686 and x86_64). To address comment #3, it does *not* occur on my RHEL 6.3 machines (on either arch).
Thanks for the feedback. I'll will cause a boost update soon. Originally, I was waiting for bug #828856 to be addressed as well, but looks like Red Hat is not very responsive regarding this issue nor plans to fix it for RHEL 6.
Robert, Is there any chance we can get an el5 released with the fix above in it? Of course I would also like to see bug #828856 fixed in el5, but it seems a Moderate Impact CVE Local exploit is not important enough for RH to fix in el6.
boost141-1.41.0-3.el5 has been submitted as an update for Fedora EPEL 5. https://admin.fedoraproject.org/updates/boost141-1.41.0-3.el5
Package boost141-1.41.0-3.el5: * should fix your issue, * was pushed to the Fedora EPEL 5 testing repository, * should be available at your local mirror within two days. Update it with: # su -c 'yum update --enablerepo=epel-testing boost141-1.41.0-3.el5' as soon as you are able to. Please go to the following url: https://admin.fedoraproject.org/updates/FEDORA-EPEL-2013-0432/boost141-1.41.0-3.el5 then log in and leave karma (feedback).
boost141-1.41.0-3.el5 has been pushed to the Fedora EPEL 5 stable repository. If problems still persist, please make note of it in this bug report.
(In reply to comment #8) > I'm also seeing this problem on my RHEL 5.9 machines (both i686 and x86_64). > To address comment #3, it does *not* occur on my RHEL 6.3 machines (on > either arch). In our office, we saw this error start appearing in RHEL 6.4 (and was not present in RHEL 6.3).
(In reply to comment #14) > In our office, we saw this error start appearing in RHEL 6.4 (and was not > present in RHEL 6.3). See bug #921792 for RHEL 6.
boost141-1.41.0-5.el5 has been submitted as an update for Fedora EPEL 5. https://admin.fedoraproject.org/updates/boost141-1.41.0-5.el5
Package boost141-1.41.0-5.el5: * should fix your issue, * was pushed to the Fedora EPEL 5 testing repository, * should be available at your local mirror within two days. Update it with: # su -c 'yum update --enablerepo=epel-testing boost141-1.41.0-5.el5' as soon as you are able to. Please go to the following url: https://admin.fedoraproject.org/updates/FEDORA-EPEL-2013-11777/boost141-1.41.0-5.el5 then log in and leave karma (feedback).
boost141-1.41.0-5.el5 has been pushed to the Fedora EPEL 5 stable repository. If problems still persist, please make note of it in this bug report.