- Boost-users - lists.preview.boost.org

RE: [Boost-users] Re: shared_ptr: LWM versus atomic_count?
by Ben Hutchings 16 Jul '04

16 Jul '04

Russell Hind <rhind(a)mac.com> wrote: > Caleb Epstein wrote: > > > I've been perusing the source code of <boost/shared_ptr.hpp> and > > noticed that it uses the <boost/detail/shared_count.hpp> class to do > > reference counting and not <boost/detail/atomic_count.hpp> (unless > > BOOST_NO_MEMBER_TEMPLATES or BOOST_MSVC6_MEMBER_TEMPLATES are > > defined). > > > > Can anyone comment on the reasoning behind using a lightweight > > mutex-protected counter instead of an atomic counter? I'd > > think that an atomic counter would be faster, but I suspect that > > there is a good reason for protecting the count with a mutex. This > > then begs the question of why use the atomic counter in the > > situations mentioned above? > > If you look at shared_count, you'll see it has to keep track of 2 > values, strong_count and weak_count, you can't do this with 'atomic > count' so therefore it uses lwm. It isn't fundamentally necessary to have two counters, though; see <http://groups.google.com/groups?threadm=slrnbhbnpc.154.do-not-spam -benh%40tin.bwsint.com>.

2 1

dupe symbols linker error with indexed_set
by Jedediah Smith 16 Jul '04

16 Jul '04

scbarg.obj : error LNK2005: "struct unbounded_type::boost::detail::unbounded_type & boost::indexed_sets::`anonymous namespace'::unbounded" (?unbounded@?A0x65131fae@indexed_sets@boost@@3AAUunbounded_type@detail@23@A) already defined in scbarcf.obj scbarg.obj : error LNK2005: "struct unbounded_type::boost::detail::unbounded_type boost::indexed_sets::`anonymous namespace'::unbounded_obj" (?unbounded_obj@?A0x65131fae@indexed_sets@boost@@3Uunbounded_type@detail@23@A) already defined in scbarcf.obj Compiler is VC7.1 The two modules in the error messages don't use indexed_set at all but indexed_set.hpp is included globally. It seems as if indexed_set.hpp has a definition hiding in it somewhere. Here are the two declarations that appear in my project (they are in different namespaces so m_events_t does not collide): typedef is::indexed_set< event_t, is::index_list< is::unique <is::member<event_t,eventid_t,&event_t::id> >, is::non_unique <is::member<event_t,double, &event_t::time> >, is::non_unique <is::member<event_t,hoption<notepos_t>,&event_t::notepos>, notepos_compare> > > m_events_t; typedef is::indexed_set< event_t, is::index_list< is::unique<is::member<event_t,eventid_t,&event_t::id> >, is::non_unique<is::member<event_t,double,&event_t::time> >, is::non_unique<is::member<event_t,Score::beatref_t,&event_t::beat> > > > m_events_t; -- Using Opera's revolutionary e-mail client: http://www.opera.com/m2/

1 0

Heterogeneous binary buffers.
by Hagen Möbius 16 Jul '04

16 Jul '04

Hi Boosters. I recently started working on a class which provides the functionality of any other binary buffer and a bit more (which is why I do it ;)). Because the content of this buffer is not homogeneous regarding the type of the elements stored in it I have this more general question: Which type should I choose to be the basis for those buffers? Can anybody advise me, pls? Regards, Hagen.

1 0

Building boost under mingw: bjam invokes VC7?? (part II: typo cor rected)
by Brewer, Christopher 15 Jul '04

15 Jul '04

Oops, sorry, Ben and Heinquoi- that was a typo :P What I *meant* to say was: Starting bjam with: bjam -sTOOLS=mingw install causes bjam to invoke VC7 instead of gcc. Same thing if quotes are used (as described in the documentation: bjam "-sTOOLS=mingw" install I ran it again, making sure that I actually used -sTOOLS and not -sTOOL, and I still see the described behavior. Again, this is using boost 1.31.0, msys and mingw on Windows 2000, and a prebuilt version of bjam recently downloaded from sourceforge. Anyone have any ideas how to make bjam (properly) invoke gcc? -Chris Brewer -----Original Message----- From: Ben Hutchings [mailto:ben.hutchings@businesswebsoftware.com] Sent: Tuesday, July 13, 2004 8:03 AM To: boost-users(a)lists.boost.org Subject: RE: [Boost-users] building boost under mingw: bjam invokes VC7?? Christopher Brewer <cbrewer(a)wmsgaming.com> wrote: > I recently attempted to build and install boost 1.31.0 using > msys (i.e., mingw). > I enter the boost directory and start bjam with: > > bjam -sTOOL=mingw install > > ...and then I see bjam invoking vc7 instead of gcc. <snip> The variable is called TOOLS, not TOOL.

3 3

RE: [Boost-users] Re: Building boost under mingw: bjam invokes VC7??
by Schalk_Cronje＠NAI.com 15 Jul '04

15 Jul '04

I might be missing the point here, but I have built Boost under mingw using TOOLS=gcc and I seem to get satisfactory results. I have not actually built the regression tests, so there might be some lurking issues that I am not aware of.

2 1

TR: pb with use Random on VC++6 and Intel C++ 8
by Heinquoi 15 Jul '04

15 Jul '04

Hello, I install and test boost on my computer since two weeks. And I have much problems. Actually I test random with the code indicated on http://www.boost.org/libs/random/index.html The code is #include <iostream> #include <boost/random.hpp> int main () { boost::mt19937 rng; // produces randomness out of thin air // see pseudo-random number generators boost::uniform_int<> six(1,6); // distribution that maps to 1..6 // see random number distributions boost::variate_generator<boost::mt19937, boost::uniform_int<> > die(rng, six); // glues randomness with mapping int x = die(); // simulate rolling a die } More simple, there is not. And, the error messages on intel c++ 8 are : Compiling... C:\BOOST_1_31_0\boost/integer_traits.hpp(129): error: class "boost::integer_traits<wchar_t={unsigned short}>" has already been defined class integer_traits<unsigned short> ^ compilation aborted for D:\X Mes Documents X\Mes Programmes\VisualC++ 6\random boost 1\random-1.cpp (code 2) Error executing cl.exe. random boost 1.exe - 1 error(s), 0 warning(s) Would anybody have already seen it? errors on template class who have multiples definitions on the header "integer_traits.hpp". The compiler don't accept surcharge of template class. Perhaps there is an option for the compiler. What do you think about this ? After this problem, I test with msvc++6 sp5 + service pack processor 5 and... Compiling... random-1.cpp C:\BOOST_1_31_0\boost/random/uniform_01.hpp(34) : fatal error C1001: INTERNAL COMPILER ERROR (compiler file 'msc1.cpp', line 1794) Please choose the Technical Support command on the Visual C++ Help menu, or open the Technical Support help file for more information Error executing cl.exe. random boost 1.exe - 1 error(s), 1 warning(s) it gives me an error fatal with no information on Technical Support of VC6. I ask me if the problem is linked at the service pack 5 for msvc and service pack 5 for processor ( http://msdn.microsoft.com/vstudio/downloads/tools/ppack/download.aspx ),I have install. Someone have installed this pack ? Thanks for your responses. heinquoi

1 1

Re: Building boost under mingw: bjam invokes VC7?? (part III: dat a)
by Brewer, Christopher 14 Jul '04

14 Jul '04

David- When you asked me to run the following lines: bjam -f- "-sTOOLS=mingw" install ECHO $(TOOLS) ; ^Z I see the following: bjam -f- "-sTOOLS=mingw" install (hundreds of lines of error text go by here) ECHO $(TOOLS) ; reply is: sh: ./TOOLS: is a directory ...and nothing is displayed after entering ctrl-Z :) The reply to ECHO $(TOOLS) ; seems odd. I doubt that this is what you were expecting. There *is* a 'tools' directory under boost_1_31_0, but not a 'TOOLS' directory. Case-sensitivity issues? I have also run the following: bjam -d+5 "-sTOOLS=mingw" install > bug.out Note that not everything is captured in 'bug.out'. I see hundreds of lines of the following on the console: The system cannot find the path specified. This is not surprising as I don't have VC7 installed. Unfortunately I don't have a place to publicly park the compressed 'bug.out' file so I will email it directly to you. However, here's the first few lines if that's of any help. (BEGIN LISTING===============) --------------------------------------------------------------------- skipping Boost.Python library build due to missing or incorrect configuration couldn't find Python.h in "c:/tools/python/include" You can configure the location of your python installation by setting: PYTHON_ROOT - currently "c:/tools/python" PYTHON_VERSION - The 2-part python Major.Minor version number (e.g. "2.2", NOT "2.2.1") - currently "2.2" The following are automatically configured from PYTHON_ROOT if not otherwise set: PYTHON_LIB_PATH - path to Python library object; currently "c:/tools/python/libs" --------------------------------------------------------------------- ...found 5266 targets... ...updating 632 targets... vc-C++ bin\boost\libs\date_time\build\boost_date_time.dll\vc7.1\debug\greg_month.ob j CALL "C:\Program Files\Microsoft Visual Studio .NET 2003\VC7\bin\VCVARS32.BAT" >nul "C:\Program Files\Microsoft Visual Studio .NET 2003\VC7\bin\cl" /Zm800 -nologo -GX -c -DBOOST_DATE_TIME_POSIX_TIME_STD_CONFIG -DBOOST_ALL_DYN_LINK=1 /Z7 /Od /Ob0 /GX /GR /MDd /Op /Zc:forScope /Zc:wchar_t -I"bin\boost\libs\date_time\build" -I"C:\msys\1.0\share\boost_1_31_0" -Fo"bin\boost\libs\date_time\build\boost_date_time.dll\vc7.1\debug\greg_mont h.obj" -Tp"C:\msys\1.0\share\boost_1_31_0\libs\date_time\build\../src/gregorian/gre g_month.cpp" ...failed vc-C++ bin\boost\libs\date_time\build\boost_date_time.dll\vc7.1\debug\greg_month.ob j... (SNIP============== as this error apparently occurs for every single file) Best regards, Chris Brewer (Company-mandated disclaimer) This message originates from WMS Gaming Inc. It contains information which may be confidential or privileged and is intended only for the individual or entity named above. It is prohibited for anyone else to disclose, copy, distribute or use the contents of this message. All personal messages express views solely of the sender, which are not to be attributed to WMS Gaming Inc., and may not be copied or distributed without this disclaimer. If you received this message in error, please notify us immediately at (847) 785-3000.

2 1

Source of Boost Graph Library
by JONATHAN JAMES HITCHCOCK 14 Jul '04

14 Jul '04

Please can anyone tell me where to find the source files for the Boost Graph Library? I downloaded boost_1_31_0.tar.gz and expected to find the source files in boost-1.31.0/libs/graph/src, but the only files there are for something called graphviz.

4 3

shared_ptr: LWM versus atomic_count?
by Caleb Epstein 14 Jul '04

14 Jul '04

I've been perusing the source code of <boost/shared_ptr.hpp> and noticed that it uses the <boost/detail/shared_count.hpp> class to do reference counting and not <boost/detail/atomic_count.hpp> (unless BOOST_NO_MEMBER_TEMPLATES or BOOST_MSVC6_MEMBER_TEMPLATES are defined). Can anyone comment on the reasoning behind using a lightweight mutex-protected counter instead of an atomic counter? I'd think that an atomic counter would be faster, but I suspect that there is a good reason for protecting the count with a mutex. This then begs the question of why use the atomic counter in the situations mentioned above? Thanks. -- Caleb Epstein caleb.epstein(a)gmail.com

2 2

need help building on AIX 5.1
by Jim Edwards 13 Jul '04

13 Jul '04

Hi, Beginning at the top, apparently my cpp doesn't like the syntax of BOOST_TESTED_AT(0x531) : xlc-C++-action bin/boost/libs/python/build/libboost_python.so/vacpp/debug/numeric.o "/home/jedwards/cvs_src/boost/boost/iterator/detail/config_def.hpp", line 28.39: 1540-0827 (S) The constant expression on the #if directive contains a syntax error at "BOOST_TESTED_AT". -- Jim Edwards jedwards(a)ucar.edu IBM Applications Analyst NCAR SCD BOULDER CO 303-497-1842

1 0