September 2013 - Boost-users - lists.preview.boost.org

Regex on Windows2012
by Michael Xanadu 11 Sep '13

11 Sep '13

Hi guys, I developed a little application called MyProgram.exe (Win7, VS2008 SP1, Qt 4.8.4) and a little DLL called MyLibrary.dll. The DLL uses BOOST's regex engine 1.5. The executable calls the DLL via couple of functions, one function inside the DLL uses the regex engine. I've tested the environment on Windows7 and all is OK. BUT: If I copy the application and the DLL to Windows2012 I see that the executable crashes if I use it. I took a closer look to the issue and found out that the cause for the crash seems to be the regex engine inside the DLL. If regex functions will be called inside the DLL, it crashes and also the application does. AND: The strange thing is, that this problem seems to occur only on Windows2012 and only if the regex engine is inside a DLL. If I use Windows7 instead or uses the regex engine inside an executable on Windows2012 - everything works fine. I've already tried a lot to solve the problem, including using different linking approaches (BOOST as shared DLL etc.). I wonder if anybody had similar issues or can give me a hint of what went wrong? Regards, Michael

1 0

syntax error reporting from imported python modules
by Richi Lists 11 Sep '13

11 Sep '13

Hi, when I execute a python script that contains errors, I get nice and complete information from boost::python::error_already_set about the error and it's location. If the script that contains a syntax error is not executed directly, but imported from the main script, I don't get that exception. Instead it returns from the function as if everything was successful. When I check the result, the tasks from the other files were performed, just the stuff from the erroneous file is missing. Do I need to somehow attach the error reporting to the imported files? Rgds Richard

1 0

[BOOST]BOOST_AUTO_TEST_CASE_TEMPLATE usage
by Hemanth Choudary M V (RBEI/ETA1) 11 Sep '13

11 Sep '13

Hi, In the code snippet below, The 3rd parameter in the BOOST_AUTO_TEST_CASE_TEMPLATE is a boost::mpl::list. Is the 3rd parameter restricted only to boost::mpl::list or anything else can also be used? #define BOOST_TEST_MODULE example #include <boost/test/included/unit_test.hpp> #include <boost/test/test_case_template.hpp> #include <boost/mpl/list.hpp> //____________________________________________________________________________// typedef boost::mpl::list<long long,unsigned char> test_types; BOOST_AUTO_TEST_CASE_TEMPLATE( my_test, T, test_types ) { BOOST_CHECK_EQUAL( sizeof(T), (unsigned)4 ); } //_ Mit freundlichen Grüßen / Best Regards, Hemanth Choudary MV RBEI/ETA1

2 1

shared_ptr to this
by Gonzalo Garramuno 10 Sep '13

10 Sep '13

I have an application where I need to pass "this" in a callback belonging to a class to a function using a shared_ptr in another thread In code: struct Data { Data( void* p ) : ptr( p ) { } boost::shared_ptr<void> ptr; }; void threadA( Data* d ) { boost::shared_ptr<void> ptr( d->ptr ); delete d; // use ptr } class A { void init_thread() { Data* data = new Data( this ); boost.thread t( boost::bind( threadA, data ) ); } }; The problem is obvious. Once threadA closes, the shared ptr is destroyed as its count is 0 and as it points to the this of class A this gets deleted too. However my class A is still active and should not get deleted. How can I work around this?

5 5

[wave] Wave macro expansion problem
by Edward Diener 10 Sep '13

10 Sep '13

The code: // test_macros_o.cpp #define f(a) f(x + (a)) #define x 2 #define g f #define h g(~ int main() { #pragma wave trace(enable) h 5); #pragma wave trace(disable) return 0; } In the latest C++ standard at 16.3.5 paragraph 5 the expansion of 'h 5)' is given as: f(2 * (~5)) but the wave.exe driver gives: C:\Programming\VersionControl\boost\libs\vmd\test\test_macros_o.cpp:11:3: h C:\Programming\VersionControl\boost\libs\vmd\test\test_macros_o.cpp:4:9: see macro definition: h [ g(~ rescanning [ C:\Programming\VersionControl\boost\libs\vmd\test\test_macros_o.cpp:3:9: see macro definition: g [ f rescanning [ f ] ]

2 2

[wave] Wave driver --c++0x option
by Edward Diener 10 Sep '13

10 Sep '13

After building the Wave driver tool from the latest Boost source on trunk, if I pass the --c++0x option to the Wave.exe driver I get: wave: exception caught: unrecognised option '--c++0x' Do I have to build the wave driver with some option to support the --c++0x option when running it ?

2 2

Re: [Boost-users] [BOOST.PREPROCESSOR] Use BOOST_PP_REPEAT in macros?
by Sergey Zhuravlev 10 Sep '13

10 Sep '13

I foolishly undef ARGS_AMOUNT after using. It is habit, because not wrote global macros earlier. Thanks, Edward Diener.

1 0

[BOOST.PREPROCESSOR] Use BOOST_PP_REPEAT in macros?
by Sergey Zhuravlev 10 Sep '13

10 Sep '13

Write code: ////////////////// #define ARGS_AMOUNT 9 #define INTERNAL_GENERATE_MEMBER_APPLY_CLASS(_, ARG_COUNTER, MemberName)\ template<class TClassWithMember BOOST_PP_COMMA_IF(ARG_COUNTER) BOOST_PP_ENUM_PARAMS(ARG_COUNTER, class TArg)>\ result_type operator()(\ TClassWithMember& classWithMember BOOST_PP_COMMA_IF(ARG_COUNTER)\ BOOST_PP_ENUM_BINARY_PARAMS(ARG_COUNTER, TArg, &arg)) const\ {\ classWithMember.MemberName(BOOST_PP_ENUM_PARAMS(ARG_COUNTER, arg));\ } #define DECLARE_MEMBER_APPLY_CLASS(TMemberApplyClassName, MemberName)\ template<class TResultType>\ struct TMemberApplyClassName\ {\ typedef TResultType result_type;\ BOOST_PP_REPEAT(ARGS_AMOUNT, INTERNAL_GENERATE_MEMBER_APPLY_CLASS, MemberName)\ }; #undef ARGS_AMOUNT ////////////////// DECLARE_MEMBER_APPLY_CLASS(TClassName, MemberName) declare class template closer to boost::apply, but my declared class call member MemberName: f.MemberName(args...); opposite f.operator()(args...) at boost::apply. Code not work correctly, because ARGS_AMOUNT not replaced to number 9 internally. ARGS_AMOUNT preserved after all macros substitution. If manualy replace text "ARGS_AMOUNT" at "call" for BOOST_PP_REPEAT in macros DECLARE_MEMBER_APPLY_CLASS to 9, then code generated and work correctly. Why it does? How preserve valuable name of const ARGS_AMOUNT? Now temporarily maualy replace ARGS_AMOUNT to 9 and write comment about this hardcoded const 9: BOOST_PP_REPEAT(9 /*max amount of arguments for MemberName*/, INTERNAL_GENERATE_MEMBER_APPLY_CLASS, MemberName)\

2 1

[thread] release lock before cond.var. notified?
by Thomas M 09 Sep '13

09 Sep '13

Hi, Consider a boolean variable, a mutex and a condition variable: // simplified code snippet begin bool Continue = false; boost::mutex someMutex; boost::condition_variable someCondVar; // simplified code snippet end A number of (boost-)threads check if the boolean variable is set to true in order to continue execution, or otherwise wait for this to happen: // simplified code snippet begin boost::unique_lock<boost::mutex> threadLock(someMutex); while (Continue == false) { someCondVar.wait(threadLock); } // simplified code snippet end and a single controller thread is responsible for setting Continue to true (see snippets below). Now the question: Is there a strong point pro or against releasing controllerLock before the notify_all-signal is sent? E.g. is there preference for the first or second variant as below (see placement of curly brackets!), or even a need for one to ensure correctness? // simplified code snippet begin - Variant 1 // notify_all is issued while controllerLock is still being held { boost::unique_lock<boost::mutex> controllerLock(someMutex); Continue = true; someCondVar.notify_all(); } // simplified code snippet end // simplified code snippet begin - Variant 2 // notify_all is issued when controllerLock has already been released { boost::unique_lock<boost::mutex> controllerLock(someMutex); Continue = true; } someCondVar.notify_all(); // simplified code snippet end Thanks for your input and help, Thomas

4 5

Memory Leak with flat_map.insert(Iterator,Iterator)
by Adam Baxter 09 Sep '13

09 Sep '13

Hello, I'm experiencing a memory leak using the flat_map.insert(Iterator, Iterator) function. The leak goes away if I do regular for loop and use flat_map.insert(*Iterator). Is this a known issue? If not, what would be helpful for a bug report? Adam

2 1