November 2004 - Boost-users

Re: newbie: Graph library - bundled properties
by TC MA 24 Nov '04

24 Nov '04

After modifying it, still had some compile problems. using std::string; struct edgeStruct { int weight; }; struct graphStruct { string name; }; int main() { using namespace boost; typedef adjacency_list<vecS, vecS, directedS,no_property, edgeStruct, graphStruct> graph_t; graph_t g; graph_t::vertex_descriptor v = *vertices(g).first; graph_t::edge_descriptor e = *out_edges(v, g).first; g[e].weight = 1; g[e].name = "graphname"; std::cout << "name: " << g[e].name << std::endl; typedef subgraph<graph_t> subgraph_t; subgraph_t sg; get_property(sg, graph_name) = "subgraph"; std::cout << "name: " << get_property(sg, graph_name) << std::endl; return exit_success; } # compile output is: gcc-C++-action /home/tcma/cpp/boosttcma/libs/graph/graph_property.test/gcc/debug/inlining-on/graph_property.o /home/tcma/cpp/boosttcma/libs/graph/graph_property.cpp: In function `int main()': /home/tcma/cpp/boosttcma/libs/graph/graph_property.cpp:32: error: 'struct edgeStruct' has no member named 'name' /home/tcma/cpp/boosttcma/libs/graph/graph_property.cpp:34: error: 'struct edgeStruct' has no member named 'name' /home/tcma/cpp/boost_1_32_0/boost/graph/subgraph.hpp: At global scope: /home/tcma/cpp/boost_1_32_0/boost/graph/subgraph.hpp: In instantiation of `boost::subgraph<main()::graph_t>': /home/tcma/cpp/boosttcma/libs/graph/graph_property.cpp:39: instantiated from here /home/tcma/cpp/boost_1_32_0/boost/graph/subgraph.hpp:249: error: incomplete type `boost::STATIC_ASSERTION_FAILURE<0u>' used in nested name specifier /home/tcma/cpp/boost_1_32_0/boost/graph/subgraph.hpp:249: error: size of array has non-integral type `<type error>' /home/tcma/cpp/boost_1_32_0/boost/pending/detail/property.hpp: In instantiation of `boost::detail::build_property_tag_value_alist<graphStruct>': /home/tcma/cpp/boost_1_32_0/boost/pending/property.hpp:63: instantiated from `boost::property_value<graphStruct, boost::graph_name_t>' /home/tcma/cpp/boost_1_32_0/boost/graph/properties.hpp:235: instantiated from `boost::graph_property<main()::graph_t, boost::graph_name_t>' /home/tcma/cpp/boosttcma/libs/graph/graph_property.cpp:40: instantiated from here /home/tcma/cpp/boost_1_32_0/boost/pending/detail/property.hpp:94: error: no type named `next_type' in `struct graphStruct' /home/tcma/cpp/boost_1_32_0/boost/pending/detail/property.hpp:95: error: no type named `value_type' in `struct graphStruct' /home/tcma/cpp/boost_1_32_0/boost/pending/detail/property.hpp:96: error: no type named `tag_type' in `struct graphStruct' /home/tcma/cpp/boost_1_32_0/boost/pending/detail/property.hpp:97: error: no type named `next_type' in `struct graphStruct' /home/tcma/cpp/boost_1_32_0/boost/pending/detail/property.hpp:98: error: no type named `tag_type' in `struct graphStruct' /home/tcma/cpp/boost_1_32_0/boost/pending/property.hpp: In instantiation of `boost::property_value<graphStruct, boost::graph_name_t>': /home/tcma/cpp/boost_1_32_0/boost/graph/properties.hpp:235: instantiated from `boost::graph_property<main()::graph_t, boost::graph_name_t>' /home/tcma/cpp/boosttcma/libs/graph/graph_property.cpp:40: instantiated from here /home/tcma/cpp/boost_1_32_0/boost/pending/property.hpp:63: error: no type named `type' in `struct boost::detail::build_property_tag_value_alist<graphStruct>' /home/tcma/cpp/boost_1_32_0/boost/pending/property.hpp:64: error: no type named `type' in `struct boost::detail::build_property_tag_value_alist<graphStruct>' /home/tcma/cpp/boost_1_32_0/boost/graph/properties.hpp: In instantiation of `boost::graph_property<main()::graph_t, boost::graph_name_t>': /home/tcma/cpp/boosttcma/libs/graph/graph_property.cpp:40: instantiated from here /home/tcma/cpp/boost_1_32_0/boost/graph/properties.hpp:235: error: no type named `type' in `struct boost::property_value<graphStruct, boost::graph_name_t>' /home/tcma/cpp/boosttcma/libs/graph/graph_property.cpp: In function `int main()': /home/tcma/cpp/boosttcma/libs/graph/graph_property.cpp:40: error: no matching function for call to `get_property(main()::subgraph_t&, boost::graph_name_t)' /home/tcma/cpp/boosttcma/libs/graph/graph_property.cpp:42: error: no matching function for call to `get_property(main()::subgraph_t&, boost::graph_name_t)' /home/tcma/cpp/boost_1_32_0/boost/graph/subgraph.hpp: In constructor `boost::subgraph<Graph>::subgraph() [with Graph = main()::graph_t]': /home/tcma/cpp/boosttcma/libs/graph/graph_property.cpp:39: instantiated from here /home/tcma/cpp/boost_1_32_0/boost/graph/subgraph.hpp:106: error: no matching function for call to `boost::detail::error_property_not_found::error_property_not_found(int)' /home/tcma/cpp/boost_1_32_0/boost/pending/detail/property.hpp:21: note: candidates are: boost::detail::error_property_not_found::error_property_not_found() /home/tcma/cpp/boost_1_32_0/boost/pending/detail/property.hpp:21: note: boost::detail::error_property_not_found::error_property_not_found(const boost::detail::error_property_not_found&) set -e "g++" -c -Wall -ftemplate-depth-255 -g -O0 -Wno-inline -I"../../../bin/boost/libs/graph/example" -I "/home/tcma/cpp/boost_1_32_0" -o "/home/tcma/cpp/boosttcma/libs/graph/graph_property.test/gcc/debug/inlining-on/graph_property.o" "/home/tcma/cpp/boosttcma/libs/graph/graph_property.cpp" "/usr/bin/objcopy" --set-section-flags .debug_str=contents,debug "/home/tcma/cpp/boosttcma/libs/graph/graph_property.test/gcc/debug/inlining-on/graph_property.o" ...failed gcc-C++-action /home/tcma/cpp/boosttcma/libs/graph/graph_property.test/gcc/debug/inlining-on/graph_property.o... ...skipped </home/tcma/cpp/boosttcma/libs/graph/graph_property.test/gcc/debug/inlining-on>/home/tcma/cpp/boosttcma/libs/graph/graph_property for lack of </home/tcma/cpp/boosttcma/libs/graph/graph_property.test/gcc/debug/inlining-on>graph_property.o... ...skipped </home/tcma/cpp/boosttcma/libs/graph/graph_property.test/gcc/debug/inlining-on>/home/tcma/cpp/boosttcma/libs/graph/graph_property.run for lack of </home/tcma/cpp/boosttcma/libs/graph/graph_property.test/gcc/debug/inlining-on>/home/tcma/cpp/boosttcma/libs/graph/graph_property... ...failed updating 1 target... ...skipped 3 targets... > Date: Mon, 22 Nov 2004 20:36:07 -0500 > From: Doug Gregor <dgregor(a)cs.indiana.edu> > Subject: Re: [Boost-users] newbie: Graph library - bundled properties > To: boost-users(a)lists.boost.org > Message-ID: <0BEF9B5F-3CF0-11D9-84DF-000D932B7224(a)cs.indiana.edu> > Content-Type: text/plain; charset=US-ASCII; format=flowed > > > On Nov 22, 2004, at 8:25 PM, TC MA wrote: > > int > > main() > > { > > using namespace boost; > > using std::string; > > Move this struct... > > > struct edgeStruct { > > int weight; > > }; > > outside of main, because of this error message: > > > /home/tcma/cpp/boosttcma/libs/graph/graph_property.cpp:21: > > error: `main()::edgeStruct' uses local type > > `main()::edgeStruct' > > Doug ______________________________________________________________________ Post your free ad now! http://personals.yahoo.ca

3 7

Re: [Boost-users] Re: newbie: Graph library - bundled properties
by TC MA 24 Nov '04

24 Nov '04

Fixed compile error, but has run time error. Modified graph_property.cpp: int main() { using namespace boost; using std::string; typedef adjacency_list<vecS, vecS, directedS,no_property, property<edge_index_t, int>, property<graph_name_t, string> > graph_t; graph_t g; get_property(g, graph_name) = "graph"; std::cout << "name: " << get_property(g, graph_name) << std::endl; typedef subgraph<graph_t> subgraph_t; subgraph_t sg; get_property(sg, graph_name) = "subgraph"; std::cout << "name: " << get_property(sg, graph_name) << std::endl; return exit_success; } Output: $ bjam ...found 409 targets... ...updating 4 targets... gcc-C++-action /home/tcma/cpp/boosttcma/libs/graph/graph_property.test/gcc/debug/inlining-on/graph_property.o gcc-Link-action /home/tcma/cpp/boosttcma/libs/graph/graph_property Chmod1 /home/tcma/cpp/boosttcma/libs/graph/graph_property execute-test /home/tcma/cpp/boosttcma/libs/graph/graph_property.run ====== BEGIN OUTPUT ====== EXIT STATUS: 139 ====== END OUTPUT ====== /home/tcma/cpp/boosttcma/libs/graph/graph_property > /home/tcma/cpp/boosttcma/libs/graph/graph_property.output 2>&1 status=$? echo >> /home/tcma/cpp/boosttcma/libs/graph/graph_property.output echo EXIT STATUS: $status >> /home/tcma/cpp/boosttcma/libs/graph/graph_property.output if test $status -eq 0 ; then cp -f /home/tcma/cpp/boosttcma/libs/graph/graph_property.output /home/tcma/cpp/boosttcma/libs/graph/graph_property.run fi verbose=1 if test $status -ne 0 ; then verbose=0 fi if test $verbose -eq 0 ; then echo ====== BEGIN OUTPUT ====== cat /home/tcma/cpp/boosttcma/libs/graph/graph_property.output echo ====== END OUTPUT ====== fi exit $status ...failed execute-test /home/tcma/cpp/boosttcma/libs/graph/graph_property.run... ...removing /home/tcma/cpp/boosttcma/libs/graph/graph_property.run ...failed updating 1 target... ...skipped 1 target... ...updated 2 targets... $ cat graph_property.output EXIT STATUS: 139 The expected output should be: $ cat graph_property.output name: graph name: subgraph EXIT STATUS: 0 > On Nov 23, 2004, at 11:23 AM, TC MA wrote: > > struct graphStruct { > string name; > }; > > Unfortunately, bundled graph properties aren't supported (yet). You can, however, add a graph name property using property<graph_name_t, string> (see below). > > typedef adjacency_list<vecS, vecS, > directedS,no_property, edgeStruct, graphStruct> > graph_t; > > Change this to: > > typedef adjacency_list<vecS, vecS, directedS, no_property, edgeStruct, > property<graph_name_t, string> > graph_t; > > g[e].name = "graphname"; > > std::cout << "name: " << g[e].name << std::endl; > > These two won't compile because we can't get at the graph properties that way. "e" is an edge descriptor, so "g[e]" is going to give you an edgeStruct& back (that's how we were able to set the edge weight ). To get to the graph_name property we just added above, use: > > get_property(g, graph_name); > > If we can come up with a better syntax for the future, we'll definitely do it. For now, we're stuck with that mess. > > typedef subgraph<graph_t> subgraph_t; > > You're actually going to run into a minor problem here, because the subgraph adaptor requires an edge_index property that we haven't provided. The easiest way to introduce such a property is to actually mix bundled and non-bundled parameters, like so: > > typedef adjacency_list<vecS, vecS, directedS, no_property, property<edge_index_t, int, edgeStruct>, > property<graph_name_t, string> > graph_t; > > Notice how we put in the edge_index_t property, but ended with an "edgeStruct"? That allows, e.g., g[e].weight to still work, but the subgraph can use the edge index property map directly. > > subgraph_t sg; > get_property(sg, graph_name) = "subgraph"; > > std::cout << "name: " << get_property(sg, > graph_name) << std::endl; > > return exit_success; > } > > All of this is fine; it's actually what we changed the above code to. > > As you can probably tell, we're still working out some of the kinks in the bundled properties setup. It's better than what we had before, but we still have a bit of work to do to make it really slick. > > Doug ===== TingChong Ma ______________________________________________________________________ Post your free ad now! http://personals.yahoo.ca

1 0

tuple manipulators?
by Paul Elliott 24 Nov '04

24 Nov '04

Would it be possible to have certain obvious tuple manipulators? For example if A and B are 2 tuples have an easy way to form the cross product of the 2 tuples? That is if A is of length L, and B is of length M, to form the tuple C of length L+M, containing copies of the elements from both? Also if D is a tuple of length N, and f is a functor taking N arguments (of the right types) find some way to take f(D) == f(D.get(1),D.get(2),...,D.get(N) ) Of course the ... is pseudo code. Fill in the appropriate list. Of course I realize that limitations in the C++ language may prevent doing this for arbitrarily large L, M, or N. But it could be done for L, M, N, large enough so that it is not a practical limitation? -- Paul Elliott 1(512)837-1096 pelliott(a)io.com PMB 181, 11900 Metric Blvd Suite J http://www.io.com/~pelliott/pme/ Austin TX 78758-3117

2 1

Re: Boost-users Digest, Vol 378, Issue 4
by Yarden Livnat 23 Nov '04

23 Nov '04

no, the problem persist in build 1671... :-( compiling with --without-serialization temporarily solve the problem though (if you can call it 'solving' :-) Yarden On Nov 23, 2004, at 2:02 PM, boost-users-request(a)lists.boost.org wrote: > > Message: 4 > Date: Tue, 23 Nov 2004 11:07:31 -0800 > From: Thomas Costa <oohrah(a)mac.com> > Subject: Re: [Boost-users] compiling 1_32 in release mode brings the > computer to its knees (Modified by Yarden Livnat) > To: boost-users(a)lists.boost.org > Message-ID: <ECC0A694-3D82-11D9-84B9-00039303FB26(a)mac.com> > Content-Type: text/plain; charset=US-ASCII; format=flowed > > Are these problems maybe fixed with Apple's updated gcc 3.3 build 1671 > that got posted to the Apple developer site recently? > > On Nov 23, 2004, at 10:19 AM, Chris Little wrote: > >> on 11/23/04 9:32 AM, Glen Simmons at gsimmons(a)macdev.itg.ti.com wrote: >> >>> On 23 Nov, 2004, at 12:17 AM, Mail Delivery Subsystem wrote: >>> >>>> >>>> Hi, >>>> >>>> I'm trying to compile 1_32 on OSX with the latest gcc (GCC) 3.3 >>>> 20030304 (Apple Computer, Inc. build 1666) >>>> however, while the debug version compile in just a few minutes, the >>>> 'release' start trashing on : >>>> libs/serialization/src/xml_grammar.cpp >>>> libs/serialization/src/xml_wgrammar.cpp >>>> >>>> in effect the compilation takes all the memory and then pages >>>> endlessly for hours, bringing the machine down to its knees. >>>> >>>> my compile command is: >>>> >>>> sudo bjam -sTOOLS=darwin "-sBUILD=release <threading>multi >>>> <define>BOOST_SIGNALS_NAMESPACE=Signals " install >>>> >>>> where the <define> is for Qt, >>>> >>>> >>>> Anybody has a clue why this is happaning and how to get the >>>> compilation to complete (other then copy the two .o from the debug >>>> tree into the release tree ? >>> >>> The same thing happened to me. Since I'm not using the serialization >>> library, I used the --without-serialization option. Not an ideal >>> solution, especially if you need the serialization lib, but a >>> workaround if you don't. >>> >>> HTH, >>> Glen Simmons >> >> On the main list Troy Straszheim post about this problem also. Troy >> found >> that it is a compiler bug. If you remove the -O flag from the build >> options >> then everything is fine. Troy filed a bug with Apple. >> >> Chris >> >> _______________________________________________ >> Boost-users mailing list >> Boost-users(a)lists.boost.org >> http://lists.boost.org/mailman/listinfo.cgi/boost-users > > >

1 0

Re: [Boost-users] IMPORTANT: yahoogroups admin rights problem
by Daniel L Elliott 23 Nov '04

23 Nov '04

Hello all, Where can I find the uBLAS license? What does rewrite uBLAS mean? Toon: Thank you for working on these issues! Thank you, dan elliott ------------------------------ Message: 6 Date: Tue, 23 Nov 2004 12:42:02 +0100 From: Toon Knapen <toon.knapen(a)fft.be> Subject: [Boost-users] IMPORTANT: yahoogroups admin rights problem To: ublas <ublas-dev(a)yahoogroups.com>, Michael Stevens <Michael.Stevens(a)epost.de>, boost-users(a)lists.boost.org, boost <boost(a)lists.boost.org> Message-ID: <41A3220A.7060104(a)fft.be> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Hello ALL, As many of you might know, currently we are unable to add new members to the ublas-ml. The problem is that only Joerg has the rights to add members to the list but Joerg does not respond anymore. Could some of you try to contact Joerg or give me his telephone number that I talk to him about this. If we do not have a solution by end of November I will create a new project on yahoogroups (and we all have to switch). Additionally we need to contact Joerg and Mathias (again!) about applying the Boost Software License. I currently have the impression that they are not willing to change ublas to the BSL. In that case IMO, we will have to rewrite uBLAS to not get thrown out of boost. Again his phone-number would be handy if you have it available. Please contact the ml or me asap about the above issues or we will need to take drastic actions. toon ------------------------------

2 1

Re: [Boost-users] Gregorian 1.32 still has VC71 warnings like 1.31
by jkharris7＠comcast.net 23 Nov '04

23 Nov '04

> most of the warnings have been L4, but unless I'm mis-reading you are set to > L3? Also, are you sure all of these were in 1.31 -- there were changes in Yes, it's set to Level 3, and no--you're right to point it out--I'm not sure whether the warnings are exactly the same as in 1.29. They were the same code (4244), because I ended up using the same suppression pragma. Everything is a plain vanilla VC71 setup, generated from the wizard, except with "Detect 64-bit Portability Issues" set to FALSE, and I tested using no other headers, just to be sure. jh

1 0

Using program_options with other options methods
by Allen Bierbaum 23 Nov '04

23 Nov '04

I have recently started using the program_options library, and I find it very useful. Unfortunately I can't seem to figure out how to allow it to work with other option parsers/handlers at the same time. For example, right now I am writing an application that uses a library that has its own set of common options that are process internally. So I end up needing to write code similar to the following: po::options_description desc("Allowed options"); desc.add_options() ("help", "produce help message") ("my_opt", po::value<int>(), "set my option") ; po::variables_map vm; po::store(po::parse_command_line(ac, av, desc), vm); po::notify(vm); custom_lib::options opts; opts.process(ac, av); Then when I run this application: ./app --my_opt=10 --liboption=custom The problem is that I get an exception from program_options: error: unknown option --liboption Is there any way to tell the program options parser to ignore options that it does not recognize? If there is no option like this, how can program_options be used with other option parsers? -Allen

1 0

Gregorian 1.32 still has VC71 warnings like 1.31
by jkharris7＠comcast.net 23 Nov '04

23 Nov '04

Do I have some incorrect options here? I would've thought these would go away, but they were the same in 1.31. I can't find others reporting the same warnings. With this at the beginning of my source: #include <boost/date_time/gregorian/gregorian.hpp> #include <boost/date_time/posix_time/posix_time.hpp> #include <boost/date_time/date_generators.hpp> ...on VC71, boost 1.32 with all defaults during build of boost, with these compiler options: /Od /I "..\..\..\Vendor\Boost\1_32_0\include\boost-1_32" /I "..\..\..\Vendor\ace\5_4\dev" /I "..\..\..\Vendor\Apache\Xerces_C\2_6_0\dev" /I "..\\" /D "WIN32" /D "_DEBUG" /D "_LIB" /D "_MBCS" /Gm /EHsc /RTC1 /MDd /GS /Fo"Debug/" /Fd"c:\my_lib\/lib/my_libD.pdb" /W3 /c /Zi /TP .\MyFile.cpp ...and getting these warnings: ..\..\..\vendor\boost\1_32_0\include\boost-1_32\boost\date_time\time_system_coun ted.hpp(57) : warning C4244: 'initializing' : conversion from 'boost::date_time::counted_time_rep<config>::int_type' to 'boost::date_time::gregorian_calendar_base<ymd_type_,date_int_type_>::date_int_t ype', possible loss of data with [ config=boost::posix_time::millisec_posix_time_system_config ] and [ ymd_type_=boost::gregorian::greg_year_month_day, date_int_type_=boost::date_time::int_adapter<unsigned long>::int_type ] ..\..\..\vendor\boost\1_32_0\include\boost-1_32\boost\date_time\time_system_coun ted.hpp( 52) : while compiling class-template member function 'boost::date_time::counted_time_rep<config>::date_type boost::date_time::counted_time_rep<config>::date(void) const' with [ config=boost::posix_time::millisec_posix_time_system_config ] ..\..\..\vendor\boost\1_32_0\include\boost-1_32\boost\date_time\time_system_coun ted.hpp( 111) : see reference to class template instantiation 'boost::date_time::counted_time_rep<config>' being compiled with [ config=boost::posix_time::millisec_posix_time_system_config ] ..\..\..\vendor\boost\1_32_0\include\boost-1_32\boost\date_time\time.hpp(52) : see reference to class template instantiation 'boost::date_time::counted_time_system<time_rep>' being compiled with [ time_rep=boost::posix_time::int64_time_rep ] ..\..\..\vendor\boost\1_32_0\include\boost-1_32\boost\date_time\posix_time\ptime .hpp(32) : see reference to class template instantiation 'boost::date_time::base_time<T,time_system>' being compiled with [ T=boost::posix_time::ptime, time_system=boost::posix_time::posix_time_system ] ..\..\..\vendor\boost\1_32_0\include\boost-1_32\boost\date_time\date_parsing.hpp (211) : warning C4244: 'argument' : conversion from 'std::allocator<_Ty>::value_type' to 'char', possible loss of data with [ _Ty=wchar_t ] ..\..\..\vendor\boost\1_32_0\include\boost-1_32\boost\date_time\date_parsing.hpp (253) : see reference to function template instantiation 'date_type boost::date_time::from_stream_type<date_type,std::basic_string<_Elem,_Traits,_Ax >::iterator>( iterator_type &,iterator_type &,wchar_t)' being compiled with [ date_type=boost::gregorian::date, _Elem=wchar_t, _Traits=std::char_traits<wchar_t>, _Ax=std::allocator<wchar_t>, iterator_type=std::basic_string<wchar_t,std::char_traits<wchar_t>,std::allocator < wchar_t>>::iterator ] ..\..\..\vendor\boost\1_32_0\include\boost-1_32\boost\date_time\gregorian\parser s.hpp(78) : see reference to function template instantiation 'boost::date_time::period<point_rep,duration_rep> boost::date_time::from_simple_string_type<boost::gregorian::date,wchar_t>(const std::basic_string<_Elem,_Traits,_Ax> &)' being compiled with [ point_rep=boost::gregorian::date, duration_rep=boost::gregorian::date_duration, _Elem=wchar_t, _Traits=std::char_traits<wchar_t>, _Ax=std::allocator<wchar_t> ]

2 1

IMPORTANT: yahoogroups admin rights problem
by Toon Knapen 23 Nov '04

23 Nov '04

Hello ALL, As many of you might know, currently we are unable to add new members to the ublas-ml. The problem is that only Joerg has the rights to add members to the list but Joerg does not respond anymore. Could some of you try to contact Joerg or give me his telephone number that I talk to him about this. If we do not have a solution by end of November I will create a new project on yahoogroups (and we all have to switch). Additionally we need to contact Joerg and Mathias (again!) about applying the Boost Software License. I currently have the impression that they are not willing to change ublas to the BSL. In that case IMO, we will have to rewrite uBLAS to not get thrown out of boost. Again his phone-number would be handy if you have it available. Please contact the ml or me asap about the above issues or we will need to take drastic actions. toon

1 0

Memory leak reported using threads library
by gast128 23 Nov '04

23 Nov '04

Dear all, Downloading and using the 1.32 version was probably not such a good idea... Now I get a mem-leak reported by the debug crt of VC 7.1. The report is located in 'init_threadmon_mutex' where a global object is allocated but never deleted. Before we start a discussion of what is a memory leak or not, in our production code we do not accept any leak reported by VC. The past has learned us that if you accept one leak, programmers/colleaques start to ignore them and in the end you have 200 leaks, instead of this one. I guess it wouldn't be difficult to resolve this one, on process exit it can be deallocated. And yes I am aware of the (phoenix) singleton discussions in Alexandrescu/Meyers, but really this kind of reinit of singletons after atexit calls won't happen too much. Wkr, me

3 4