- Boost - lists.preview.boost.org

Compile boost python extension using IDE on windows
by Raul Durand 13 Jan '11

13 Jan '11

Hi guys, I have a extension on Linux that uses boost python with no problems. Now I want to compile my extension on Windows, so I downloaded the precompiled boost libraries from: http://www.boostpro.com/download/boost_1_44_setup.exe I tried to compile my extension on windows using Visual Studio c++ 2008 express edition. I followed all the instructions from: http://boost.cppll.jp/HEAD/libs/python/doc/building.html to use an IDE to compile my own project and obtain a dll. I included all sources and libs and turned on RTTI and changed the debug and release builds to Multi-thread dll debug and Multi-thread dll release. All compilation process is successful. The problem is during linking appears *some warnings as*: 1> Creating library C:\My Dropbox\codes\fem\pyfem\Debug\pyfem.lib and object C:\My Dropbox\codes\fem\pyfem\Debug\pyfem.exp 1>libfembase.obj : warning LNK4217: locally defined symbol ??0str_base@detail@python@boost@@IAE@PBD@Z (protected: __thiscall boost::python::detail::str_base::str_base(char const *)) imported in function "public: __thiscall boost::python::str::str(char const *)" (??0str@python@boost@@QAE@PBD@Z) 1>libfembase.obj : warning LNK4217: locally defined symbol ?? 1str_base@detail@python@boost@@QAE@XZ (public: __thiscall boost::python::detail::str_base::~str_base(void)) imported in function "public: __thiscall boost::python::str::~str(void)" (??1str@python@boost@ @QAE@XZ) 1>libfembase.obj : warning LNK4217: locally defined symbol ?throw_error_already_set@python@boost@@YAXXZ (void __cdecl boost::python::throw_error_already_set(void)) imported in function "int __cdecl boost::python::len(class boost::python::api::object const &)" (?len@python@boost@@YAHABVobject@api@12@@Z) *and some errors as:* 1>boost_python-vc90-mt-gd-1_44.lib(dict.obj) : error LNK2019: unresolved external symbol __imp__PyDict_Items referenced in function "public: class boost::python::list __thiscall boost::python::detail::dict_base::items(void)const " (?items@dict_base @detail@python@boost@@QBE?AVlist@34@XZ) 1>boost_python-vc90-mt-gd-1_44.lib(dict.obj) : error LNK2019: unresolved external symbol __imp__PyDict_Keys referenced in function "public: class boost::python::list __thiscall boost::python::detail::dict_base::keys(void)const " (?keys@dict_base @detail@python@boost@@QBE?AVlist@34@XZ) 1>boost_python-vc90-mt-gd-1_44.lib(dict.obj) : error LNK2019: unresolved external symbol __imp__PyDict_Update referenced in function "public: void __thiscall boost::python::detail::dict_base::update(class boost::python::api::object const &)" (?update@dict_base@detail@python@boost@ @QAEXABVobject@api@34@@Z) 1>boost_python-vc90-mt-gd-1_44.lib(dict.obj) : error LNK2019: unresolved external symbol __imp__PyDict_Values referenced in function "public: class boost::python::list __thiscall boost::python::detail::dict_base::values(void)const " (?values@dict_base @detail@python@boost@@QBE?AVlist@34@XZ) 1>C:\My Dropbox\codes\fem\pyfem\Debug\pyfem.dll : fatal error LNK1120: 133 unresolved externals 1>Build log was saved at "file://c:\My Dropbox\codes\fem\pyfem\Debug\BuildLog.htm" 1>pyfem - 230 error(s), 56 warning(s) *#############* I got the same warnings and errors using bjam. I have been successful compiling and linking as a static library pyfem.lib using VS, but it does not work as a python extension even renaming to .pyd. - What I am missing to obtain the dll for my extension using VS? - Is it possible to create a boost python extension using VS C++ 2008 express edition? - I am using Win7 64 bits and Python 2.6 64 bits, so using boost 64 bit binaries is mandatory? Thanks. Raul Durand

2 1

[server] Continuing problems with svn operations
by Beman Dawes 13 Jan '11

13 Jan '11

I'm using svn a lot as part of the run-up to 1.46.0, but am having trouble. svn operations are very slow, and often fail. A typical error message is: Error: REPORT of '/svn/boost/!svn/vcc/default': Could not read chunk size: An existing Error: connection was forcibly closed by the remote host. Error: (https://svn.boost.org) Can anything be done about svn problems? I know Rene has already done a lot of work trying to improve the situation, but there are still serious performance problems. --Beman

3 2

[1.46.0][Date_Time] Could someone please commit some of the simple uncontroversial patches?
by KTC 13 Jan '11

13 Jan '11

Hi, Release time code freeze is coming again, please could anyone have a quick glance at <https://svn.boost.org/trac/boost/query?status=assigned&status=new&status=re…> and commit some of the simple uncontroversial patches? Some of them have been around for quite a while..... Pretty please? :) KTC

2 1

Development Practices (Proposal Progress)
by Dean Michael Berris 13 Jan '11

13 Jan '11

Hi Everyone, I was looking around at the Trac wiki and saw that there's a page called "ImprovingPractices". It's https://svn.boost.org/trac/boost/wiki/ImprovingPractices if you want to look. The Objectives stated on the page are definitely the same ones I am looking to address with a proposal I was thinking of writing from scratch. A quick read through definitely seems to me that there is significant overlap between what I had in mind and what's already in this document. One thing that doesn't seem to be in this document, is how to get new libraries into a Boost distribution. I'm not sure if this part should be in the same document either, but I really do think that somehow the continuity between the development of a library for inclusion into Boost and the release procedure should be in-tune. The document also assumes that the development of a library happens in the same repository -- mainly because it assumes that subversion is going to be used exclusively. That's understandable, but in the case of a proposal to support Git-based libraries as well I wonder whether it's alright if I edit this document for the proposed enhancements. So my question would be: - What would be the best way for me to solicit feedback on a proposal that tweaks the current process? Is editing this wiki page the way or should I create a new one? - Because there a number of parts of the proposal I'm going to write that deal with things on the front-end of the process (the interest-gathering, development (collaborative or otherwise), and review/inclusion part) as well as the back-end of the process (distribution maintenance, bug fixing / patching, and release management / packaging) is it better if these areas are described in different Wiki pages or should one page with different sections suffice? - Is the Wiki the best place to put this? Would others be more comfortable in a different means of looking at and commenting on a complete proposal? Thanks in advance and I definitely look forward to feedback and guidance. -- Dean Michael Berris about.me/deanberris

3 3

[type_traits] common_type fails when typeof is not used
by vicente.botet 13 Jan '11

13 Jan '11

Hi, The current common_type implementation doesn't works when BOOST_COMMON_TYPE_DONT_USE_TYPEOF is defined. This is needed on platforms that don't support Boost.Typeof correctly. It seems that the the implementation proposed by Jeffrey Hellrung has not been applied completly. I have attached a complementaty test on builting types that proves the current code does't works and also the implementation proposed by Jeffrey Hellrung to solve these issues some time ago. John, let me know if this works for you and if it can be included on trunk (I could do it myself if you prefer). Best, _____________________ Vicente Juan Botet Escribá http://viboes.blogspot.com/

3 5

Boost.Filesystem v3 as default in Boost 1.46.0?
by Jeremiah Willcock 12 Jan '11

12 Jan '11

Will the default version of Boost.Filesystem be v3 in the Boost 1.46.0 release? The current release branch still has v2 as the default (at least in <boost/filesystem.hpp>). Also, if I want to write code that works for both v2 and v3, should I test the BOOST_FILESYSTEM_VERSION macro directly? -- Jeremiah Willcock

4 3

[spirit] semantic action for mismatches?
by caustik 12 Jan '11

12 Jan '11

I've come to notice that there seems to be a missing bit of symmetry in spirit with regards to semantic actions. If a rule matches, and thus executes it's semantic action(s), but a rule which includes that rule mismatches, there seems to be no way to "unwind" the code executed down the chain. For example, if one of your semantic actions allocates memory or increments a reference count, how do you free / release that reference in the mismatch scenario? I've thought about using something like a shared_ptr, but it seems like that gets pretty sloppy and unnatural. Is there something you can think of that would work? caustik

5 25

[move] abiguious ctor (msvc-10.0)
by Oliver Kowalke 11 Jan '11

11 Jan '11

I get error C2668 with msvc-10.0 - (gcc.4.x works). Any ideas? 1>test.cpp(44): error C2668: 'X::X': Mehrdeutiger Aufruf einer überladenen Funktion 1> test.cpp(19): kann 'X::X<boost::_bi::bind_t<R,F,L>>(Fn &&,size_t)' sein 1> with 1> [ 1> R=void, 1> F=void (__cdecl *)(int), 1> L=boost::_bi::list1<boost::_bi::value<int>>, 1> Fn=boost::_bi::bind_t<void,void (__cdecl *)(int),boost::_bi::list1<boost::_bi::value<int>>> 1> ] 1> test.cpp(16): oder "X::X<boost::_bi::bind_t<R,F,L>>(Fn,size_t)" 1> with 1> [ 1> R=void, 1> F=void (__cdecl *)(int), 1> L=boost::_bi::list1<boost::_bi::value<int>>, 1> Fn=boost::_bi::bind_t<void,void (__cdecl *)(int),boost::_bi::list1<boost::_bi::value<int>>> 1> ] 1> bei Anpassung der Argumentliste '(boost::_bi::bind_t<R,F,L>, int)' 1> with 1> [ 1> R=void, 1> F=void (__cdecl *)(int), 1> L=boost::_bi::list1<boost::_bi::value<int>> 1> ] test app: #include <cstdlib> #include <iostream> #include <boost/bind.hpp> #include <boost/move/move.hpp> class X { private: BOOST_COPYABLE_AND_MOVABLE( X); public: X() {} template< typename Fn > X( Fn fn, std::size_t i) {} template< typename Fn > X( BOOST_RV_REF( Fn) fn, std::size_t i) {} X( X const& other) {} X & operator=( BOOST_COPY_ASSIGN_REF( X) other) { X tmp( other); return * this; } X( BOOST_RV_REF( X) other) {} X & operator=( BOOST_RV_REF( X) other) { X tmp( boost::move( other) ); return * this; } }; void f( int) {} int main() { try { X x( boost::bind( f, 1), 1); return EXIT_SUCCESS; } catch ( std::exception const& e) { std::cerr << "exception: " << e.what() << std::endl; } catch (...) { std::cerr << "unhandled exception" << std::endl; } return EXIT_FAILURE; } -- Neu: GMX De-Mail - Einfach wie E-Mail, sicher wie ein Brief! Jetzt De-Mail-Adresse reservieren: http://portal.gmx.net/de/go/demail

2 3

[local] local functions ready for review
by Lorenzo Caminiti 11 Jan '11

11 Jan '11

Hello all, A few months ago, I probed for interest in a library for local functions with Boost.ScopeExit's variable binding. Based on that discussion, I have now implemented the Boost.Local library. I have uploaded the source files boost-local_0_1_0-818.zip in the Boost Vault under Function Objects: http://www.boostpro.com/vault/index.php?&direction=0&order=&directory=Funct… The documentation on the web: http://contractpp.sourceforge.net/boost-local_0_1_0-818/libs/local/doc/html… Any comment? As far as I can tell, Boost.Local is ready for a formal review so I am looking for a review manager. Any volunteer? Thanks, Lorenzo ________________________________________________ P.S. A preview from Boost.Local's documentation: The Boost Local library implements local functions, local blocks, and local exits for the C++ programming language. INTRODUCTION Local functions are a form of information hiding and are useful for dividing procedural tasks into subtasks which are only meaningful locally, avoiding cluttering other parts of the program with functions, variables, etc unrelated to those parts. Local functions therefore complement other structuring possibilities such as namespaces and classes. Local functions are a feature of many programming languages, notably Pascal and Ada, yet lacking from C++ (see also [N2511]). This library supports the following features for local functions (see the Alternatives section for a comparison between this library and features offered by C++ local class members, C++0x lambda functions, Boost.Lambda, Boost.Phoenix, and Boost.ScopeExit): * Local functions can access, or better bind, any of the variables from the enclosing scope. Furthermore, local functions defined within a member function can bind the enclosing object this. * The local function body is programmed using the usual C++ syntax. * Local functions can be passed as template parameters (so they can be conveniently passed to STL algorithms, etc). * However, local functions must be defined within a declarative context (e.g., at a point in the code where local variables can be declared) thus they cannot be defined within an expression. In addition to local functions, this library also supports the following features: * Local blocks which define blocks of code that bind variables from the enclosing scope. Local blocks allow programmers to bind variables as constants (constant binding) so to prevent local chunks of code from modifying selected variables. * Local exits which define blocks of code that are executed when the enclosing scope is exited (again with support for constant binding and binding of the object this). AN EXAMPLE The following example illustrates a simple use of this library and it starts introducing the library API (see the Tutorial section for more details on how to use this library): * A local function is passed to the STL std::for_each() algorithm to add together the values of an array. Variables in scope are bound to the local function by both constant value and non-constant reference. * A local exit is used to automatically release the array's memory at scope exit. * A local block is used to assert the correct final value of the summation in a constant-correct context (therefore preventing the assertion from mistakenly changing any of the variables in scope). #include <boost/local/function.hpp> #include <boost/local/block.hpp> #include <boost/local/exit.hpp> #include <algorithm> #include <iostream> #include <cassert> int main() { double sum = 0.0; int factor = 10; BOOST_LOCAL_FUNCTION( (void) (add)( (double)(num) (const bind)((factor)) (bind)((&sum)) ) ) { sum += factor * num; std::clog << "Summed: " << sum << std::endl; } BOOST_LOCAL_FUNCTION_END(add) add(100.0); size_t size = 2; double* nums = new double[size]; BOOST_LOCAL_EXIT( (const bind)((&size)) (bind)((nums)) ) { if (size && nums) delete[] nums; std::clog << "Freed array: " << nums << std::endl; } BOOST_LOCAL_EXIT_END nums[0] = 90.5; nums[1] = 7.0; std::for_each(nums, nums + size, add); BOOST_LOCAL_BLOCK( (const bind)((&sum)) ) { // So far: sum = 10 * 100.0 + 10 * 90.5 + 10 * 7.0 = 1975.0 assert(sum == 1975.0); std::clog << "Asserted summation: " << sum << std::endl; } BOOST_LOCAL_BLOCK_END return 0; }

2 5

[geometry] Excessive inspection report errors
by Beman Dawes 10 Jan '11

10 Jan '11

See http://boost.cowic.de/rc/inspect-trunk.html#geometry A bunch of these problems are related to documentation - Daniel James and some of the other Boost doc folks have dealt with similar problems in other libraries, so can probably help. Thanks, --Beman

2 3

[QuickBook] missing prev/next
by Lorenzo Caminiti 10 Jan '11

10 Jan '11

Hello all, I have generate the Boost.Units documentatiom using Boost.QuickBook from the latest Boost release 1.45.0 but the HTML pages are missing the prev and next page links. I only have the up and home links in the navbar. How can I get the prev and next links to show? Do I need to enable some XLS configuration in the Jamfile? Thank you very much. -- Sent from my mobile device Lorenzo

4 21

cygwin API: posix or windows
by vicente.botet 09 Jan '11

09 Jan '11

Hi, while trying to make Boost.System and Boost.Chrono configurable header only libs we need to include windows.h on some include files. To avoid the inclusion of this intrusive file, we decided sometime ago to include some files in detail/win that will present just the needed interfaces. In my humble opinion cygwin is a wrapper of posix intefaces under windows OS. So, boost libraries shouldn't use Windows interfaces when working under cygwin. What's the experience of others? Best, Vicente

10 10

[tasklets] new lib
by Oliver Kowalke 09 Jan '11

09 Jan '11

boost.tasklet is a split-up from old boost.fiber. The lib allow to execute micro-tasks (tasklets) in a thread and assigns each small task with a micro-thread/fiber - so tasklets are cooperativly scheduled. The API contains classes and functions to manage and synchronize tasklets (even if the tasklets are running on different threads). boost.tasklet is an example how boost.fiber could be used for cooperative scheduling ansd will be used by the threadpool from boost.task. Oliver

2 3

Workaround for clang-linux + GNU stdlib + Boost.Iostreams
by Bryce Lelbach 09 Jan '11

09 Jan '11

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 There's a discrepancy in the definition of wint_t between GCC and some later versions of clang (clang defines it as an int, GCC an unsigned int). This ends breaking Boost.Iostreams with a GNU GCC compiled version of the GNU stdlib and clang (I'll restrain myself from giving the lengthy technical explaination - if you have link-time errors with clang + GCC compiled stdlib + Boost.Iostreams, this is probably the issue). Unfortunately, I cannot patch this issue with a workaround in Boost easily. The right combination of clang flags can work around this by not using the clang intrinsic/builtin headers, but this has a good chance of causing other problems. The simplest solution here is to rebuild clang (building the GNU stdlib with clang is tricky at best). I think newer version of the mainstream clang trunk may have fixed this, but if not, you need to change the assignment to WIntType in the clang::TargetInfo ctor ($CLANG_ROOT/lib/Basic/TargetInfo.cpp, line 40-50ish). It should be UnsignedInt. You also need to hack $CLANG_ROOT/lib/Headers/stdint.h, around 630 - the macro definition of WINT_MIN and WINT_MAX need to be changed to: #define WINT_MIN 0u #define WINT_MAX __UINTN_MAX(__WINT_WIDTH__) If you'd rather not recompile clang, you can try hacking your clang intrinsic headers (not advisable, might break more stuff). Find the resource/intrinsic header directory (usually something like /usr/local/clang/2.9/include/), and make the change specified above to the stdint.h header there. In the stddef.h header, around like 60, you can change the wint_t typedef. - -- Bryce Lelbach aka wash boost-spirit.com -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (GNU/Linux) iEYEARECAAYFAk0fnjQACgkQO/fqqIuE2t7NVQCgowl7TqnCRgWyjRGN+/u3fh2R NywAoK59zWsZiB7vVvtKU0IehHiMIVdz =COrS -----END PGP SIGNATURE-----

4 5

[wiki] Warning Guidelines for VC
by Michael Caisse 08 Jan '11

08 Jan '11

A question just came up on IRC concerning the use of the /Za ("disable language extensions") and it not working. While searching for a previous post from STL on the subject [1] Google offered this link: <https://svn.boost.org/trac/boost/wiki/Guidelines/WarningsGuidelines> STL's (Stephan Lavavej's) advice on the subject was to *not* use /Za because it has "an extremely nasty bug". Unfortunately our guidelines suggest it should be used. I began to edit the wiki page but considering the amount of text and an example on the subject I thought it might be best to ask permission first. Should I edit the page to be in-line with STL's suggestion or just leave it as-is? [1] <http://lists.boost.org/Archives/boost/2010/04/165531.php > michael -- Michael Caisse Object Modeling Designs www.objectmodelingdesigns.com

5 13

Boost.Bind and boost.ref propagation
by Felipe Magno de Almeida 07 Jan '11

07 Jan '11

Hi, Why this code asserts? struct foo { typedef void result_type; template <typename T> void operator()(T obj) const { BOOST_MPL_ASSERT ((boost::mpl::or_ < typename boost::mpl::not_ < typename boost::is_same<typename boost::unwrap_reference<T>::type, T>::type >::type , typename boost::is_reference<T>::type >)); } }; int main() { int i = 0; boost::bind(::foo(), boost::ref(i))(); } I would expect the T type to be boost::reference_wrapper<int>, but it is actually int. It is not int& as well. How do I get to propagate the boost::reference_wrapper through boost.bind to a template? I don't want to use T& for the parameter, because the algorithm accepts objects-by-value correctly, but I would like to be able to use boost.bind as I would call this algorithm with boost.ref for some data-types. foo()(boost::ref(i)); // Calls with boost::reference_wrapper<int> Thanks in advance, -- Felipe Magno de Almeida

2 1

serialization linking issue - mac os x 10.6 + boost 1.45
by John Clayton 07 Jan '11

07 Jan '11

Hi i've got a new problem with my code, now that I'm compiling and linking against the 1.45 boost libraries. When I compile and link, I now get these errors at the link stage: "boost::serialization::version<Model::KeyValue<QPointF> >::value", referenced from: __ZN5boost13serialization7versionIN5Model8KeyValueI7QPointFEEE5valueE$non_lazy_ptr in libDataModel.a(Project.o) __ZN5boost13serialization7versionIN5Model8KeyValueI7QPointFEEE5valueE$non_lazy_ptr in libRendering.a(EditorDocument.o) __ZN5boost13serialization7versionIN5Model8KeyValueI7QPointFEEE5valueE$non_lazy_ptr in libDataModel.a(Serialization.o) (maybe you meant: __ZN5boost13serialization7versionIN5Model8KeyValueI7QPointFEEE5valueE$non_lazy_ptr) I have version definitions for the template types using boost serialization, along the form of: namespace boost { namespace serialization { template<typename T, int scale, int defaultCurveType> struct version< Model::PropertyArrayT<T, scale, defaultCurveType> > { typedef mpl::int_<1> type; typedef mpl::integral_c_tag tag; BOOST_STATIC_CONSTANT(unsigned int, value = version::type::value); }; template<typename T> struct version< Model::KeyValue<T> > { typedef mpl::int_<1> type; typedef mpl::integral_c_tag tag; BOOST_STATIC_CONSTANT(unsigned int, value = version::type::value); }; } } I tried instantiating (in the .cpp of the header where this stuff is declared), like this: boost::serialization::version< Model::KeyValue<QPointF> > a; but the link error for this type still persists. The code is compiled into a library, and then the link step fails on the application. I hope its not the linker being extra ingenious and throwing this fake definition away. Can anyone see what I'm doing wrong? The code works OK if I use/link against boost 1.42. Thanks -- John Clayton Skype: johncclayton

1 0

Re: [boost] Are memory leaks reported by Boost.Test Microsoft Memory Leak Detection Enabling spurious or real?
by Thomas Harning Jr. 07 Jan '11

07 Jan '11

Regarding this problem, I've found more details about it on the Visual Studio Connect website: http://connect.microsoft.com/VisualStudio/feedback/details/106937/memory-le… typeinfo is used by the unit test framework if you use BOOST_AUTO_TEST_CASE_TEMPLATE. Few fixes that seem reasonable: All use the following tools: _CrtMemCheckpoint [1] and _CrtMemDumpAllObjectsSince [2] available since VS2003 to report objects in specific regions Few reasonable cases: * For the entire test path, if dumping enabled, checkpoint before all tests start, dump after all tests complete * For each test suite, checkpoint before suite starts, dump after suite ends (complicated by nested suites...) * For each test case, checkpoint before test starts, dump after test ends 1: http://msdn.microsoft.com/en-us/library/h3z85t43(v=vs.90).aspx 2: http://msdn.microsoft.com/en-us/library/hhy09244(v=vs.90).aspx -- Thomas Harning Jr.

5 5

Re: [boost] [BGL] problem w/ push_relabel_max_flow and floating point capacities
by Marc Schafer 07 Jan '11

07 Jan '11

Jeremiah, I am looking at a particular case in the debugger and I see that is_flow returns false because of comparisons that fail by an error that is many orders of magnitude smaller than the numbers being compared. For example, x==y fails because x and y are both approximately 1. and x-y is 1.e-14. I could replace the particular comparisons that are failing for my test case, but I need to know how to make the entire algorithm generically suitable for use with floating point numbers. is_flow just performs a consistency check after the calculation is complete. Does the calculation itself require any modifications to work correctly with floating point numbers? What are the appropriate tolerances to use? Do the relational tests (<, >) also need modification to properly handle floating point? -Marc "Jeremiah Willcock" <jewillco(a)osl.iu.edu> wrote in message news:<alpine.LRH.2.00.1101051355460.26649(a)flowerpot.osl.iu.edu>... > On Wed, 22 Dec 2010, Marc Schafer wrote: > > > Using the push_relabel_max_flow feature on graphs that have edge > > capacities in double or single precision often results in a failed > > assertion (line 706 push_relabel_max_flow.hpp) on debug builds although > > the answer provided is correct for builds with assertions disabled. > > > > The is_flow() method (line 564, push_relabel_max_flow.hpp) performs some > > consistency checking after the algorithm completes and there is an > > assertion that it returns true. The comparisons inside is_flow are done > > using ==, !=, >, and <. Using a locally modified version, I was able to > > see that all the cases where is_flow returns false are caused by errors > > on the order of machine precision. > > > > I believe the error is a faulty comparison mechanism inside is_flow > > rather than a problem with my graphs or the push_relabel_max_flow > > algorithm. > > I agree with what you are saying. If you just replace the == and != > operations in is_flow (and not < and >), does the code work correctly? > Boost already has floating point equality comparisons with a tolerance, > but does not appear to have inequalities. > > -- Jeremiah Willcock > _______________________________________________ > Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost >

2 1

Re: [boost] [BGL] problem w/ push_relabel_max_flow and floating point capacities
by Marc Schafer 07 Jan '11

07 Jan '11

Apologies for the double post. Outlook crashed while I was typing the first one and I didn't think it got sent. -Marc "Marc Schafer" <Marc.Schafer(a)mathworks.com> wrote in message news:<AA067D07858B794CA4E1612284A7E8DAFFF4(a)exmb-01-ah.ad.mathworks.com>... > Jeremiah, > For my test case, is_flow fails on a comparison x==y where x and y are both approximately 1. and x-y is 1.e-14. > > However, the graphs I want to process could have capacities that are any valid floating point numbers. This leads to several questions: > > is_flow is a consistency check after the algorithm is complete, does the algorithm need modifications to work correctly with floating point numbers? > > How do I determine the correct tolerances to use for comparisons? > > How do I deal with the relational operators (<, >)? > > -Marc > > "Jeremiah Willcock" <jewillco(a)osl.iu.edu> wrote in message news:<alpine.LRH.2.00.1101051355460.26649(a)flowerpot.osl.iu.edu>... > > On Wed, 22 Dec 2010, Marc Schafer wrote: > > > > > Using the push_relabel_max_flow feature on graphs that have edge > > > capacities in double or single precision often results in a failed > > > assertion (line 706 push_relabel_max_flow.hpp) on debug builds although > > > the answer provided is correct for builds with assertions disabled. > > > > > > The is_flow() method (line 564, push_relabel_max_flow.hpp) performs some > > > consistency checking after the algorithm completes and there is an > > > assertion that it returns true. The comparisons inside is_flow are done > > > using ==, !=, >, and <. Using a locally modified version, I was able to > > > see that all the cases where is_flow returns false are caused by errors > > > on the order of machine precision. > > > > > > I believe the error is a faulty comparison mechanism inside is_flow > > > rather than a problem with my graphs or the push_relabel_max_flow > > > algorithm. > > > > I agree with what you are saying. If you just replace the == and != > > operations in is_flow (and not < and >), does the code work correctly? > > Boost already has floating point equality comparisons with a tolerance, > > but does not appear to have inequalities. > > > > -- Jeremiah Willcock > > _______________________________________________ > > Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost > > > _______________________________________________ > Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost >

1 0

Re: [boost] [BGL] problem w/ push_relabel_max_flow and floating point capacities
by Marc Schafer 07 Jan '11

07 Jan '11

Jeremiah, For my test case, is_flow fails on a comparison x==y where x and y are both approximately 1. and x-y is 1.e-14. However, the graphs I want to process could have capacities that are any valid floating point numbers. This leads to several questions: is_flow is a consistency check after the algorithm is complete, does the algorithm need modifications to work correctly with floating point numbers? How do I determine the correct tolerances to use for comparisons? How do I deal with the relational operators (<, >)? -Marc "Jeremiah Willcock" <jewillco(a)osl.iu.edu> wrote in message news:<alpine.LRH.2.00.1101051355460.26649(a)flowerpot.osl.iu.edu>... > On Wed, 22 Dec 2010, Marc Schafer wrote: > > > Using the push_relabel_max_flow feature on graphs that have edge > > capacities in double or single precision often results in a failed > > assertion (line 706 push_relabel_max_flow.hpp) on debug builds although > > the answer provided is correct for builds with assertions disabled. > > > > The is_flow() method (line 564, push_relabel_max_flow.hpp) performs some > > consistency checking after the algorithm completes and there is an > > assertion that it returns true. The comparisons inside is_flow are done > > using ==, !=, >, and <. Using a locally modified version, I was able to > > see that all the cases where is_flow returns false are caused by errors > > on the order of machine precision. > > > > I believe the error is a faulty comparison mechanism inside is_flow > > rather than a problem with my graphs or the push_relabel_max_flow > > algorithm. > > I agree with what you are saying. If you just replace the == and != > operations in is_flow (and not < and >), does the code work correctly? > Boost already has floating point equality comparisons with a tolerance, > but does not appear to have inequalities. > > -- Jeremiah Willcock > _______________________________________________ > Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost >

1 0

Re: [boost] [Boost-commit] svn:boost r67745 - trunk/boost/fusion/container/list/detail
by Steven Watanabe 07 Jan '11

07 Jan '11

AMDG On 1/6/2011 5:17 PM, marshall(a)idio.com wrote: > Author: marshall > Date: 2011-01-06 20:17:40 EST (Thu, 06 Jan 2011) > New Revision: 67745 > URL: http://svn.boost.org/trac/boost/changeset/67745 > > Log: > Applied patch; refs #2823; will merge to release once tests cycle > Text files modified: > trunk/boost/fusion/container/list/detail/build_cons.hpp | 3 ++- > 1 files changed, 2 insertions(+), 1 deletions(-) > > Modified: trunk/boost/fusion/container/list/detail/build_cons.hpp > ============================================================================== > --- trunk/boost/fusion/container/list/detail/build_cons.hpp (original) > +++ trunk/boost/fusion/container/list/detail/build_cons.hpp 2011-01-06 20:17:40 EST (Thu, 06 Jan 2011) > @@ -48,7 +48,8 @@ > static type > call(First const& f, Last const& l) > { > - return type(*f, next_build_cons::call(fusion::next(f), l)); > + typename result_of::value_of<First>::type v = *f; > + return type(v, next_build_cons::call(fusion::next(f), l)); > } > }; result_of::deref would be better. There's no reason to make an extra copy. (Although this code looks pretty copy-happy to begin with). In Christ, Steven Watanabe

1 0

[chrono] isuue with accuracy test on some platforms
by vicente.botet 06 Jan '11

06 Jan '11

Hi, I'm observing some errors in Boost.Chrono related to this accuracy test whcih fails if the elapsed time is over 200ms (see http://svn.boost.org/svn/boost/trunk/libs/chrono/example/chrono_accuracy_te… ). The results are showed below. I'm tempted to remove the check on the unit tests as there is nothing working wrong with the library but with the platform. This user can run this test and see the output to have an idea of the accuracy of the std::clock function on her platform. Comments? Best, Vicente ________________________________________ accuracy test for 1 second(s)... 1000000000 timeout_in_nanoseconds 200000000 maximum_delta 1811871936 sys_dur ../libs/chrono/test/../example/chrono_accuracy_test.cpp(74): test 'sys_dur > timeout_in_nanoseconds - maximum_delta && sys_dur < timeout_in_nanoseconds + maximum_delta' failed in function 'void<unnamed>::accuracy_test(int, char**)' 1811868977 mono_dur ../libs/chrono/test/../example/chrono_accuracy_test.cpp(80): test 'mono_dur > timeout_in_nanoseconds - maximum_delta && mono_dur < timeout_in_nanoseconds + maximum_delta' failed in function 'void<unnamed>::accuracy_test(int, char**)' 1811868554 hires_dur ../libs/chrono/test/../example/chrono_accuracy_test.cpp(86): test 'hires_dur > timeout_in_nanoseconds - maximum_delta && hires_dur < timeout_in_nanoseconds + maximum_delta' failed in function 'void<unnamed>::accuracy_test(int, char**)' 3 errors detected. Or accuracy test for 1 second(s)... 1000000000 timeout_in_nanoseconds 200000000 maximum_delta 1536756457 sys_dur ../libs/chrono/test/../example/chrono_accuracy_test.cpp(74): test 'sys_dur > timeout_in_nanoseconds - maximum_delta && sys_dur < timeout_in_nanoseconds + maximum_delta' failed in function 'void<unnamed>::accuracy_test(int, char**)' 1536855050 mono_dur ../libs/chrono/test/../example/chrono_accuracy_test.cpp(80): test 'mono_dur > timeout_in_nanoseconds - maximum_delta && mono_dur < timeout_in_nanoseconds + maximum_delta' failed in function 'void<unnamed>::accuracy_test(int, char**)' 1536855104 hires_dur ../libs/chrono/test/../example/chrono_accuracy_test.cpp(86): test 'hires_dur > timeout_in_nanoseconds - maximum_delta && hires_dur < timeout_in_nanoseconds + maximum_delta' failed in function 'void<unnamed>::accuracy_test(int, char**)' 3 errors detected.

1 0

[system][chrono] header-only libs
by vicente.botet 06 Jan '11

06 Jan '11

Hi, after discussion with Ion about whether Boost.Interprocess could use Boost.Chrono and Boost.System he has found as major drawback that both are not header-only libs. I have implemented Boost.Chrono as a header only lib. If the user wants Boost.Chrono be a header-only lib s/he needs to define the BOOST_CHRONO_INLINED macro. While this could be useful, there is a problem as Boost.Chrono depends on Boost.System that is not a header-only lib. * I have added this on the configuration file #ifdef BOOST_CHRONO_INLINED #define BOOST_CHRONO_INLINE inline #define BOOST_CHRONO_DECL #else #define BOOST_CHRONO_INLINE // as before #endif * Copied all the .cpp files to the detail/inlined directory. * Renamed the .cpp file by .hpp files * Included in the old .cpp files the corresponding .hpp file in the detail/inlined directory * I've replaced the prefix suffix as follows: #ifndef BOOST_CHRONO_INLINED #include <boost/config/abi_prefix.hpp> // must be the last #include #endif #ifndef BOOST_CHRONO_INLINED #include <boost/config/abi_suffix.hpp> // pops abi_prefix.hpp pragmas #else #include <boost/chrono/detail/inlined/file_name.hpp> #endif * prefixed the definitions of the function in the .cpp files by BOOST_CHRONO_INLINE While doing this I have found a minor issue on Windows platform. I hve nedeed to sorrounded by () all the use of min and max functions on the test files to avoid the macros defined on the windows.h file. That means that a header-only lib could break user sources as it will include implicitly some platform files. Another drawback that I have no inspected yet could be the increase in size and compilation time of the user code. Beman do you mind if I try to make your library configurable as header-only, so Ion could use Boost.System and Boost.Chrono and preserve Boost.Interprocess as a header-only lib? Best, Vicente P.S. You can see the sandbox for the last modifications _____________________ Vicente Juan Botet Escribá http://viboes.blogspot.com/

8 23

library_status.cpp
by Bryce Lelbach 06 Jan '11

06 Jan '11

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 library_status.cpp (cd $BOOST_ROOT/tools/regression/build && bjam library_status toolset=gcc) does not build for me. I don't know how long this has been broken for, but who is the maintainer of this software? - -- Bryce Lelbach aka wash boost-spirit.com -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (GNU/Linux) iEYEARECAAYFAk0lOrQACgkQ9cB/V3/s9Ez94wCgpSQpaWDHsBFm1S1uvdRKrSG9 fwAAoKlJWaEaAuEKZwvZMCIwC5C//tOO =5BVP -----END PGP SIGNATURE-----

3 2