November 2009 - Boost-users - lists.preview.boost.org

[bind] Binding a bound function does not compile
by Eugene Yakubovich 17 Nov '09

17 Nov '09

Hello, The following code does not compile (gcc 4.3.2, boost 1.40). #include <boost/bind.hpp> using boost::bind; void foo(int) {} template <class F> void invoker(F f) { f(); } template <class F> void call(F f) { bind(invoker<F>, f)(); } int main(int argc, char* argv[]) { call(bind(foo, 1)); return 0; } I am passing a bound function as an argument to bind. The binding itself seems to work but the call fails. In other words, if I change the line in call() to bind(invoker<F>, f); it compiles just fine. Any ideas? Thanks, Eugene This e-mail and its attachments are intended only for the individual or entity to whom it is addressed and may contain information that is confidential, privileged, inside information, or subject to other restrictions on use or disclosure. Any unauthorized use, dissemination or copying of this transmission or the information in it is prohibited and may be unlawful. If you have received this transmission in error, please notify the sender immediately by return e-mail, and permanently delete or destroy this e-mail, any attachments, and all copies (digital or paper). Unless expressly stated in this e-mail, nothing in this message should be construed as a digital or electronic signature.

2 1

Losing memory connecting multiple signals to a passed slot_type
by COBryan 17 Nov '09

17 Nov '09

Hi, I've been losing memory when using Boost::Signal. I cannot find anywhere in forums and documentation if it's legal to use a passed slot_type multiple times, so I feel that may be the issue, but was wondering if someone here could either verify that or otherwise help me fix my problem. I have a PageManager that inherits from boost::signals::trackable. From this PageManager I make this call: ------------------------- texthandle[i].getPhrases ( boost::bind( &PageManagerCommon::shortcutHandler, this, i ) ); ------------------------- TextHandle also inherits boost::signals::trackable. To test the leak, I replaced TextHandle::getPhrases with the following code: ------------------------- typedef boost::signal0<void> PhraseSignalType; struct PhraseType { PhraseSignalType signal; }; void TextHandle::getPhrases ( const PhraseSignalType::slot_type& aSlot //!< The slot to connect the phrase signals to ) const { PhraseType *phrase; std::list<PhraseType*> phraseList; std::list<PhraseType*>::iterator phraseIter; for(int i = 0; i < 50; i++) { phrase = new PhraseType; phrase->signal.connect(aSlot); phraseList.push_back(phrase); } for(phraseIter = phraseList.begin(); phraseIter != phraseList.end(); phraseIter++) { delete *phraseIter; } phraseList.clear(); return; } ------------------------- Now, what I expect to happen is to create 50 phrases, push them into a list, then iterate through the list destroying these phrases. When the phrase is destroyed, the signal should disconnect and clean itself up. Instead, my memory usage simply goes up. We use an allocator that tracks 'memory handles,' and it seems that the number of unfreed handles is always the number of phrases - 1, so in this case we had 49 handles that weren't freed. I feel like this has something to do with using the passed slot_type for multiple connects, but can't find anywhere saying that I can't do that. I would appreciate any suggestions. Thanks, Chris -- View this message in context: http://old.nabble.com/Losing-memory-connecting-multiple-signals-to-a-passed… Sent from the Boost - Users mailing list archive at Nabble.com.

1 0

[boost::signals] Who deletes copied objects passed to slots?
by Ed Sutton 17 Nov '09

17 Nov '09

I am using boost::signals2 to pass copied objects from my library's worker thread to a client GUI. While this works well, the GUI must delete the copied objects when done which seems wrong. Are there any techniques I could use so that the client would not be responsible for the object deletion? Thanks in advance, -Ed Multi-threaded Red Hat Enterprise 5.x QT 4x

2 1

Boost.Statechart and CppUnit in Eclipse on Ubuntu
by Patricia Hawley 17 Nov '09

17 Nov '09

Hi, I have a simple state machine that I put together in NetBeans CDT (on a WinXP machine with MinGW) with a short main program to check the first few states. I then moved the code to a Linux box (Ubuntu) with Eclipse (gcc 4.3.3) and the CppUnit plug-in ECUT to follow our practices and procedures. The code compiles in the CppUnit test fixture; however, when the test reaches myMachine.initiate(); (the first thing that is Statechart-related) the testing stops. What can fail in the initiate() method that would stop the execution? Any suggestions on debugging this would be appreciated. Pat

2 1

boost statechart - using deep_history/shallow_history with custom reactions
by ae ss 17 Nov '09

17 Nov '09

Hi Question on using history with custom reactions I have a parent-state-A state and a parent-state-B state parent-state-A has a child-state-A1 and a child-state-A2 (nested state) The machine initializes on parent-state-A entering by default in child-state-A1 (the initial state) Then moves to child-state-A2 Then to parent-state-B Then a event occurs and I need the machine to go back to child-state-A2 (and not the initial state child-state-A2) question is how do I do it when using custom reactions that is, instead of doing a hardcoded parent-state-B::react (event& evty) { return transit<child-state-A2 > (); } I would like to do something like parent-state-B::react (event& evty) { return transit<parent-state-A > (); } and let the state machine figure out that it needs to go back to child-state-A2 rather then the initial child-state-A1 the documentation shows how to do it with sc::transition... but I couldn't find a way to do it in the case of custom reactions trial and error didn't work either... help! : ) thanks in advance silva

2 4

Re: [Boost-users] [serialization] Problems with serializing a shared_ptr
by Stefan Ulbrich 17 Nov '09

17 Nov '09

Hi, thanks a lot for you help. Unfortunately, adding a virtual dummy function to the base class (the only one I can alter directly in the library) didn't resolve the problem. So I don't know (didn't check the beta) if 1.41 will resolve this issue. I tried to track it down deep into the heap allocation but I don't have any experience at all with meta-programming and the MPL and gave up in the end. I found a quick (and dirty) fix, though, that helps in my case. Therefor, I wrote a specialization of boost::archive::details::heap_allocator for the eigen::Matrix<> classes that uses code that was reserved for compiler that didn't support the has_new_operator feature (which is obviously the case). namespace boost { namespace archive { namespace detail { template<typename Scalar, int Rows, int Cols> struct heap_allocator<Eigen::Matrix<Scalar, Rows, Cols> > { typedef Eigen::Matrix<Scalar, Rows, Cols> T; static T * invoke() { return static_cast<T *> (operator new(sizeof (T))); } }; }; }; }; Best regards, Stefan

1 0

Preprocessor Question
by Surya Kiran Gullapalli 17 Nov '09

17 Nov '09

Hello all, I'm trying to come up with a solution for the following problem. I've a macro with three arguments, but if the third is not given, it should be assumed as '0'. Something on the lines of default arguments for functions. I'm using VS2005, and g++ 4.4.x so i can use Variadic macros, but I'm unable to come up with a solution. Is there any way i can achieve this using Boost.Preprocessor ? I tried using BOOST_PP_IF but unable to proceed further. Any Ideas ? Surya

4 12

[test] Still SIGCHLD issues?
by Johan Nilsson 17 Nov '09

17 Nov '09

Hi, I'm in progress of updating some projects to use Boost 1.40 (upgrading from 1.34.1). Everything now works fine under Windows, but under Linux there are issues with running unit tests that spawns and control child processes. My tests now report fatal errors and stops responding. Searching the archives I see several references to a problem with SIGCHLD, but no fix seems to have been issued. I'm sure there is a valid reason for treating SIGCHLD as errors, but does it really need to be enabled by default? Seems like a regression to me. Is it possible to disable this without patching Boost.Test? I'd really, really, really like to not go that route. Regards / Johan

3 5

Re: [Boost-users] [serialization] Problems with serializing a shared_ptr
by Stefan Ulbrich 17 Nov '09

17 Nov '09

Thanks for your answer, Robert. Mi compiler version is: gcc version 4.0.1 (Apple Inc. build 5465) The error occurs only in the load method of the class and the serialization routines for the matrix class are non-inversive (save seems to work). Do you have any idea what is causing this error? Best, Stefan

3 2

boost_1_40_setup___.exe - failing to download mirror list
by Carlos Tapia 17 Nov '09

17 Nov '09

I'm trying to run the BoostPro 1.40.0 Installer on an XP machine but I keep getting the following message: Failed to download mirror list, all downloadable content has been disabled.

2 1