- Boost - lists.boost.org

interest in thread-pool library
by Oliver Kowalke 12 Jun '08

12 Jun '08

Hi, I don't know why but my previous email from yesterday did not occur on the mailing list? I've uploaded two versions of my threadpool library - maybe you are interressted in. It provides thread management, queueing and scheduling policies as well chaining of tasks and lazy processing of submited tasks (== processed only if requested). version 1: submit returens a future (Braddock future library) whic his cancelable (remove pending task from queue) version 2: submit returns a task object which provides a interrupt memebr function (task can be interrupted if its is running and cooperative) Both libraries are available at Boost.Vault / Concurrent Programming. so long, Oliver -- Psssst! Schon vom neuen GMX MultiMessenger gehört? Der kann`s mit allen: http://www.gmx.net/de/go/multimessenger

1 0

Boost build mailing list moderation?
by Emil Dotchevski 12 Jun '08

12 Jun '08

Recently I subscribed and made 2 posts to the boost build mailing list, and got the standard reply that they need to be reviewed by the moderators and that they will either be posted or I'll get a notification that there is a problem. I got no notification but I don't see my questions appearing in the boost build mailing list. My 2 posts have the following subject lines: - Using boost build on linux - Testing problem with Boost Build Any ideas? Emil Dotchevski Reverge Studios, Inc. http://www.revergestudios.com/reblog/index.php?n=ReCode

2 1

Boost exceptions under Windows
by Zack S 11 Jun '08

11 Jun '08

Hi folks, Why doesn't the Boost code, in the cases where it needs to throw an exception under Windows, offer the option to display a dialog window? Something like a #define'able option could be useful. Zack

10 17

Re: [boost] [circular_buffer] why is there no reserve(size_type)?
by Jan Gaspar 11 Jun '08

11 Jun '08

because: 1) it's behaviour is slightly different from the standard - you can increase but also decrease the capaciry of the container; decreasing of the capacity is not supported by reserve() 2) the word capacity is used all over the documentation of the circular buffer. I think set_capacity expresses better what the method does. Jan ----- Original Message ---- From: Steven Watanabe <watanabesj(a)gmail.com> To: boost(a)lists.boost.org Sent: Wednesday, 11 June, 2008 3:49:21 PM Subject: Re: [boost] [circular_buffer] why is there no reserve(size_type)? AMDG Jan Gaspar wrote: > it is called set_capacity(capacity_type) > Why is this different from the standard library name? In Christ, Steven Watanabe _______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost __________________________________________________________ Sent from Yahoo! Mail. A Smarter Email http://uk.docs.yahoo.com/nowyoucan.html

2 1

Building boost 1.35 with boost.serialization 1.36
by Chris Hamilton 11 Jun '08

11 Jun '08

I require a few of the new features introduced in Boost.Serialization 1.36, namely the support of serialization across multiple shared libraries and multi-threaded support. The Serialization library itself doesn't build cleanly, but adding a missing #include in one of the headers solves that for me. However, I now get errors when building the MPI library: gcc.compile.c++ bin.v2/libs/mpi/build/gcc-3.4.6/release/threading-multi/packed_iarchive.o In file included from libs/mpi/src/packed_iarchive.cpp:11: ./boost/archive/impl/archive_pointer_oserializer.ipp:27: error: redefinition of `class boost::archive::detail::<unnamed>::serializer_map<Archive>' ./boost/archive/impl/archive_pointer_iserializer.ipp:27: error: previous definition of `class boost::archive::detail::<unnamed>::serializer_map<Archive>' "g++" -ftemplate-depth-128 -O3 -finline-functions -Wno-inline -Wall -fPIC -pthread -DBOOST_ALL_NO_LIB=1 -DBOOST_MPI_DYN_LINK=1 -DBOOST_MPI_SOURCE=1 -DNDEBUG -I"." -I"/opt/lam/gnu/include" -c -o "bin.v2/libs/mpi/build/gcc-3.4.6/release/threading-multi/packed_iarchive.o" "libs/mpi/src/packed_iarchive.cpp" ...failed gcc.compile.c++ bin.v2/libs/mpi/build/gcc-3.4.6/release/threading-multi/packed_iarchive.o... ...removing bin.v2/libs/mpi/build/gcc-3.4.6/release/threading-multi/packed_iarchive.o gcc.compile.c++ bin.v2/libs/mpi/build/gcc-3.4.6/release/threading-multi/packed_oarchive.o gcc.compile.c++ bin.v2/libs/mpi/build/gcc-3.4.6/release/threading-multi/packed_skeleton_iarchive.o In file included from libs/mpi/src/packed_skeleton_iarchive.cpp:11: ./boost/archive/impl/archive_pointer_oserializer.ipp:27: error: redefinition of `class boost::archive::detail::<unnamed>::serializer_map<Archive>' ./boost/archive/impl/archive_pointer_iserializer.ipp:27: error: previous definition of `class boost::archive::detail::<unnamed>::serializer_map<Archive>' "g++" -ftemplate-depth-128 -O3 -finline-functions -Wno-inline -Wall -fPIC -pthread -DBOOST_ALL_NO_LIB=1 -DBOOST_MPI_DYN_LINK=1 -DBOOST_MPI_SOURCE=1 -DNDEBUG -I"." -I"/opt/lam/gnu/include" -c -o "bin.v2/libs/mpi/build/gcc-3.4.6/release/threading-multi/packed_skeleton_iarchive.o" "libs/mpi/src/packed_skeleton_iarchive.cpp" ...failed gcc.compile.c++ bin.v2/libs/mpi/build/gcc-3.4.6/release/threading-multi/packed_skeleton_iarchive.o... ...removing bin.v2/libs/mpi/build/gcc-3.4.6/release/threading-multi/packed_skeleton_iarchive.o gcc.compile.c++ bin.v2/libs/mpi/build/gcc-3.4.6/release/threading-multi/packed_skeleton_oarchive.o ...skipped <pbin.v2/libs/mpi/build/gcc-3.4.6/release/threading-multi>libboost_mpi-gcc34-mt-1_35.so.1.35.0 for lack of <pbin.v2/libs/mpi/build/gcc-3.4.6/release/threading-multi>packed_iarchive.o... Any idea on how to solve these? It seems that Boost.MPI is rather 'chummy' with the internals of Boost.Serialization, and that things have changed rather significantly there in 1.36. Regards, Chris Hamilton

2 1

[serialization] ambiguous symbol if include "binary_iarchive.hpp"
by Wang Yun 11 Jun '08

11 Jun '08

Hi, I meet a compile error of boost::function only when I add this line #include <boost/archive/binary_iarchive.hpp> I'm using VC2005, boost 1.35.0 error message: error C2872: 'function' : ambiguous symbol 1> could be 'e:\libraries\boost_1_35_0\boost\function\function_base.hpp(106) : boost::function' 1> or 'boost::detail::function' 1> or 'e:\libraries\boost_1_35_0\boost\function\function_base.hpp(106) : boost::function' // ======= the code =============================== #include "stdafx.h" #include <boost/function.hpp> using boost::function; #include <boost/archive/binary_iarchive.hpp> //////////////// when add this line, error will happen int _tmain(int argc, _TCHAR* argv[]) { function<void()> fn; return 0; } //================================================== Thanks, Wang Yun

3 2

[python] can I stop iteration without exception being thrown?
by Gennadiy Rozental 11 Jun '08

11 Jun '08

Hi, I an using vector_indexing_suite for both standard containers export and for our in-house ones. I've noticed that boost::python::error_already_set is bwing thrown every time iteration in python over container instances is about to finish. The exception is ignored somewhere inside boost::python but it's still quite inconvenient for debugging: I want to catch actual error, but can't skip 100s of fake ones. Is there any reason why it was designed this way? Is there any way to avoid it? Gennadiy

6 14

Re: [boost] [circular_buffer] why is there no reserve(size_type)?
by Jan Gaspar 11 Jun '08

11 Jun '08

it is called set_capacity(capacity_type) ----- Original Message ---- From: Thorsten Ottosen <thorsten.ottosen(a)dezide.com> To: "boost(a)lists.boost.org" <boost(a)lists.boost.org> Sent: Wednesday, 11 June, 2008 2:05:14 PM Subject: [boost] [circular_buffer] why is there no reserve(size_type)? Hi Jan, The headline pretty much says it all. -Thorsten _______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost __________________________________________________________ Sent from Yahoo! Mail. A Smarter Email http://uk.docs.yahoo.com/nowyoucan.html

2 1

[circular_buffer] why is there no reserve(size_type)?
by Thorsten Ottosen 11 Jun '08

11 Jun '08

Hi Jan, The headline pretty much says it all. -Thorsten

1 0

Bug in iterator_range when NDEBUG not defined but boost asserts are enabled
by John Reid 11 Jun '08

11 Jun '08

iterator_range.hpp defines is_singular() if NDEBUG is not defined but it calls is_singular through the BOOST_ASSERT macro. Unfortunately it is possible to enable BOOST_ASSERT and have NDEBUG defined which leads to a compilation error. To see this compile the iterator_range.cpp example included with the library with the following defines: gcc -DBOOST_ENABLE_ASSERT_HANDLER -I/usr/local/src/boost-trunk/ -DNDEBUG -c iterator_range.cpp John.

4 3