August 2008 - Boost-users - lists.preview.boost.org

[Math] errno on WCE
by Paul Heil 02 Sep '08

02 Sep '08

I'm compiling the Boost 1.36.0 Math library for Windows Mobile 5 using Visual Studio 2008. I have STLPort-5.1.5 installed. When I get to anything requiring errno support, I get compiler errors as below: compile-c-c++ bin.v2\libs\math\build\msvc-WM5\debug\link-static\runtime-link-static\stdlib-stlport\threading-multi\acosh.obj acosh.cpp Unknown EVC++ compiler version - please run the configure tests and report the results eMbedded Visual C++ 3 and .NET don't have a errno.h header; STLport won't include native errno.h here C:\boost\boost_1_36_0\boost/math/policies/error_handling.hpp(149) : error C2065: 'errno_t' : undeclared identifier [...] C:\boost\boost_1_36_0\boost/math/policies/error_handling.hpp(149) : error C2065: 'EDOM' : undeclared identifier [..] C:\boost\boost_1_36_0\boost/math/policies/error_handling.hpp(233) : error C2065: 'ERANGE' : undeclared identifier In the STLPort errno.h header, there's a section: #ifdef _STLP_WCE // only show message when directly including this file in a non-library build # if !defined(__BUILDING_STLPORT) && (_STLP_OUTERMOST_HEADER_ID == 0x205) # pragma message("eMbedded Visual C++ 3 and .NET don't have a errno.h header; STLport won't include native errno.h here") #define errno errno_t I can add an "#include<crtdefs.h>" to get errno_t support, but that just declares it as "typedef int errno_t;". Which, of course, gives me " boost/math/policies/error_handling.hpp(149) : error C2513: 'int' : no variable declared before '=' " errors as that isn't the way Boost expects to be able to use errno. What can I do to either get errno support under CE or remove the dependency of that feature without breaking Boost? Thanks, PaulH

2 2

[Serialization] Compatibility of text archives in 1.36 and 1.34
by Glenn Ramsey 02 Sep '08

02 Sep '08

Hi, I recently upgraded from boost 1.34 to boost 1.36. A version of my application built with 1.36 can read the text archives produced by 1.34, but a version of my application built with 1.34 cannot read the text archives produced by 1.36. Is this the behaviour that I should expect? Glenn -- Glenn Ramsey http://www.componic.co.nz

3 5

Problem with Boost.Test (1.36.0) and Visual Studio 2008 SP1
by Matthieu Brucher 02 Sep '08

02 Sep '08

Hi, I'm trying to build a test suite that worked with Linux and an older Boost version (1.34.1). I tried to compile it with Visual Studio 2008 SP1 with Boost 1.36 and I ran into this error : link /nologo /OUT:IRT_test_suit.exe /LIBPATH:D:\Travail\boost_1_36_0\lib "/LIBPA TH:D:\Python 2.5\libs" /LIBPATH:. IRT.lib boost_unit_test_framework-vc90-mt-1_36 .lib tools\boost_main_test.obj test_light.obj boost_unit_test_framework-vc90-mt-1_36.lib(boost_unit_test_framework-vc90-mt-1_3 6.dll) : error LNK2005: "public: static class boost::unit_test::unit_test_log_t & __cdecl boost::unit_test::singleton<class boost::unit_test::unit_test_log_t>:: instance(void)" (?instance@?$singleton@Vunit_test_log_t@unit_test@boost@@@unit_t est@boost@@SAAAVunit_test_log_t@23@XZ) déjà défini(e) dans test_light.obj IRT_test_suit.exe : fatal error LNK1169: un ou plusieurs symboles définis à diff érentes reprises ont été rencontrés boost_main_test.cpp consists of: #define BOOST_TEST_DYN_LINK #define BOOST_TEST_MODULE "IRT test suit" #include <boost/test/unit_test.hpp> And test_light.cpp: #include <boost/test/unit_test.hpp> Both files were compiled with the arguments cl /nologo /EHsc /Ox /MD -D_CRT_SECURE_NO_DEPRECATE -DWIN32 /TP /ID:\Travail\boo st_1_36_0 /I. /c test_light.cpp /Fotest_light.obj I didn't find a thing on the Internet with this issue, so I hope someone will have a clue... Matthieu -- French PhD student Website : http://matthieu-brucher.developpez.com/ Blogs : http://matt.eifelle.com and http://blog.developpez.com/?blog=92 LinkedIn : http://www.linkedin.com/in/matthieubrucher

3 8

[thread] wrong assertion following call to pthread_mutex_timedlock()?
by Wolfgang Fertsak 02 Sep '08

02 Sep '08

In boost 1.35 and 1.36 the implementation of mutex::timed_lock() and recursive_mutex::timed_lock() for posix-platforms with BOOST_PTHREAD_HAS_TIMEDLOCK defined looks like this: bool timed_lock(system_time const & abs_time) { struct timespec const timeout=detail::get_timespec(abs_time); int const res=pthread_mutex_timedlock(&m,&timeout); BOOST_ASSERT(!res || res==EBUSY); return !res; } The assertion following pthread_mutex_timedlock() always fails in case of a timeout since the system call returns ETIMEDOUT when timed out and not EBUSY (according to the open group specification http://www.opengroup.org/onlinepubs/009695399/functions/pthread_mutex_timed…) Best regards, Wolfgang

1 1

[filesystem][xpressive] Hotfix patches available for 1.36.0
by Beman Dawes 02 Sep '08

02 Sep '08

Hotfix patches are available to fix Boost.Filesystem and Boost.Xpressive [1.36.0] problems. See https://svn.boost.org/trac/boost/wiki/ReleasePractices/HotFixes Providing hotfix patches is something new and experimental for Boost. Please let us know if you find the patches useful or have other comments. Thanks, --Beman

9 11

unit test XML ouput to HTML XSL?
by Ákos Maróy 01 Sep '08

01 Sep '08

Hi, I'm wondering if someone has already created an XSL stylesheet that transforms the XML output of boost::test to human-friendly HTML? It would be even better if this XSL could combine the output of several test output XML files. I'm also wondering on the state of integration between boost test and continous integration tools, like Hudson, Bitten, etc.. Thanks, Akos

3 2

Serializing a char *
by Diane 01 Sep '08

01 Sep '08

Based on tutorial and reference material, and some previous posts here, it seems like I should be able to serialize a char* by using a wrapper like tracked_char: BOOST_STRONG_TYPEDEF(char, tracked_char) template<class Archive> void serialize(Archive &ar, tracked_char & tc){ ar & tc; } But when I try to serialize my struct that contains char*: typedef struct MYSTRUCT { char chtest; char *chstar; } MYSTRUCT; template<class Archive> void serialize(Archive & ar, MYSTRUCT & it, const unsigned int version) { ar & it.chtest; tracked_char *tcp = it.chstar; //****error on this line****// ar & *tcp; } I get this compile error: cannot convert `char*' to `tracked_char*' in initialization What is the compiler complaining about here? How do I fix it? I'm not sure if I have the correct usage of BOOST_STRONG_TYPEDEF, so I also tried: struct tracked_char { char m_char; // the real data template<class Archive> void serialize(Archive &ar, unsigned int version){ ar & m_char; } // casting operators - not compiled char operator char () const { //****error here****// return m_char; } char & operator char() { //****error here****// return m_char; } }; I need someone to point me in the right direction. Thanks, Diane

4 5

Changing from std::map to boost::bimap
by Sean Farrow 01 Sep '08

01 Sep '08

Hi: I am looking to change from std::map to boost::bimap. I have the following defined: std::bimap<std::bitset<6>, int, BitSetComp> DotsList; And boost::bimap<std::bitset<6>, int, BitSetComp>::iterator DotsIterator; If it's needed BitsetComp is defined as: struct BitSetComp { bool operator()(const std::bitset<6>& lhs, const std::bitset<6>& rhs) const { return lhs.to_ulong() < rhs.to_ulong(); } }; How do I change these definitions to use boost::bimap. Any help apreciated. Chers Sean. __________ Information from ESET NOD32 Antivirus, version of virus signature database 3401 (20080829) __________ The message was checked by ESET NOD32 Antivirus. http://www.eset.com

2 1

[Serialization] Are the serialized char* different when code runs in 32/64 bits machines ?
by jotalo jotalo 01 Sep '08

01 Sep '08

Hi, I use the Bost serialization library to serialize some classes that I put in a char* variable ... I have a Server which publish this char* variable and Clients get this variable and de-serialize it to obtain the original class... Everything work fine when servers and clients are running in machines with the same architecture (64 or 32 bits)... But when I run the server in a 64 bits machine, the char* variable is not recognized in a 32 bits client. In this code: std::stringstream is; is.rdbuf()->pubsetbuf(c, m_StringSize); boost::archive::binary_iarchive *ia = new boost::archive::binary_iarchive(is); The last line returns the std::exception: "invalid signature" To be sure that I am not doing something wrong I printed the first 80 characters of this variable when it is created in a 64 bits server and a 32 bits server: In a 64 bits machine server, the print out is: "........serialization::archive.................MonH1F............Efficiency.(cou" (the "." are not printable ascii characters) in Hexa format it is : 16 00 00 00 00 00 00 00 73 65 72 69 61 6C 69 7A 61 74 69 6F 6E 3A 3A 61 72 63 68 69 76 65 04 04 08 04 08 01 00 00 00 06 00 00 00 00 00 00 00 4D 6F 6E 48 31 46 00 00 00 00 1E 00 00 00 00 00 00 00 45 66 66 69 63 69 65 6E 63 79 20 28 63 6F 75 And ind the case of a 32 bits machine server, it is: "....serialization::archive.............MonH1F........Efficiency.(counter2/counte" 16 00 00 00 73 65 72 69 61 6C 69 7A 61 74 69 6F 6E 3A 3A 61 72 63 68 69 76 65 04 04 04 04 08 01 00 00 00 06 00 00 00 4D 6F 6E 48 31 46 00 00 00 00 1E 00 00 00 45 66 66 69 63 69 65 6E 63 79 20 28 63 6F 75 6E 74 65 72 32 2F 63 6F 75 6E 74 65 They are similar but not equal... Thank you, very much Juan Otalora _________________________________________________________________ Explore the seven wonders of the world http://search.msn.com/results.aspx?q=7+wonders+world&mkt=en-US&form=QBRE

4 4

problem with boost io_service
by lakers fan 01 Sep '08

01 Sep '08

Hello, I have an application in which I created an io_service object in the main thread of the application. Then I created a listener object (listener socket class), this class creates a new worker thread and waits for incoming messages. First I created listener socket in the context of the main thread and called async_receive_from() so that io_service has something to work, and then I called io_service.run() in the worker thread. But the run() returned immediately. So I tried creating the socket in the worker thread and called async_receive_from() and then called io_service.run(). Again run() returned immeidately with boost::system::error_code's m_val = 0. Can someone point me in the right direction? i m kind of clueless.. Thanks, Bharani _________________________________________________________________ Talk to your Yahoo! Friends via Windows Live Messenger. Find out how. http://www.windowslive.com/explore/messenger?ocid=TXT_TAGLM_WL_messenger_ya…

3 3