June 2006 - Boost-users - lists.preview.boost.org

gcc + stlport + boost 1_33_0
by William Deegan 22 Jun '06

22 Jun '06

Greetings, We're using gcc 3.2.3 stlport 4.61 and boost 1_33_0. When compiling some code which uses scoped_ptr, it was necessary to remove the explicit call to std::auto_ptr and change it to use just plain auto_ptr. I've attached the diffs below. Is there a boost configuration or other change I would need to make to allow this combination to compile without the below changes. So std::auto_ptr<> would be resolvable? Thanks, Bill These are the files I had to modify: M boost/intrusive_ptr.hpp diff -r1.2 intrusive_ptr.hpp 210c210 < return std::less<T *>()(a.get(), b.get()); --- > return less<T *>()(a.get(), b.get()); 257c257,258 < template<class E, class T, class Y> std::basic_ostream<E, T> & operator<< (std::basic_ostream<E, T> & os, intrusive_ptr<Y> const & p) --- > using _STL::basic_ostream; > template<class E, class T, class Y> basic_ostream<E, T> & operator<< (basic_ostream<E, T> & os, intrusive_ptr<Y> const & p) M boost/scoped_ptr.hpp diff -r1.2 scoped_ptr.hpp 63c63 < explicit scoped_ptr(std::auto_ptr<T> p): ptr(p.release()) // never throws --- > explicit scoped_ptr(_STL::auto_ptr<T> p): ptr(p.release()) // never throws M boost/shared_array.hpp diff -r1.2 shared_array.hpp 160a161 > using _STL::less; 163c164 < return std::less<T*>()(a.get(), b.get()); --- > return less<T*>()(a.get(), b.get()); M boost/shared_ptr.hpp diff -r1.2 shared_ptr.hpp 197c197 < explicit shared_ptr(std::auto_ptr<Y> & r): px(r.get()), pn() --- > explicit shared_ptr(_STL::auto_ptr<Y> & r): px(r.get()), pn() 221c221 < shared_ptr & operator=(std::auto_ptr<Y> & r) --- > shared_ptr & operator=(_STL::auto_ptr<Y> & r) 437c437,438 < template<class E, class T, class Y> std::basic_ostream<E, T> & operator<< (std::basic_ostream<E, T> & os, shared_ptr<Y> const & p) --- > using _STL::basic_ostream; > template<class E, class T, class Y> basic_ostream<E, T> & operator<< (basic_ostream<E, T> & os, shared_ptr<Y> const & p) M boost/detail/shared_count.hpp diff -r1.2 shared_count.hpp 138c138 < explicit shared_count( std::auto_ptr<Y> & r ): pi_( new sp_counted_impl_p<Y>( r.get() ) ) --- > explicit shared_count( _STL::auto_ptr<Y> & r ): pi_( new sp_counted_impl_p<Y>( r.get() ) ) 213c213 < return std::less<sp_counted_base *>()( a.pi_, b.pi_ ); --- > return _STL::less<sp_counted_base *>()( a.pi_, b.pi_ ); 307c307 < return std::less<sp_counted_base *>()(a.pi_, b.pi_); --- > return _STL::less<sp_counted_base *>()(a.pi_, b.pi_);

1 0

[Review results][PQS]
by fred bertsch 22 Jun '06

22 Jun '06

The Physical Quantities System review has ended, and, unfortunately, we are going to have to turn it down at this time. However, it is clear from the reviews that there is an enormous amount of interest in the library. I'm really hoping that that Andy Little will continue his work on it and resubmit it for review. Andy has been following the large discussion of PQS, so I will not enumerate the more minor criticisms included there. Some of the major criticisms that influenced my difficult decision include the following: * Many of the reviews liked the documentation, but improvements there would help out for a large number of reviewers. * If the library were more flexible about the type system used, several reviewers would have been happier. However, I would suggest that as long as a plan is in place for improving the flexibility, this should not prevent an initial version of the library from being accepted. * Since the library includes a large number of compile-time type manipulations, it would be great if the library included standard metafunctions to query these types. In addition, following other template metaprogramming conventions established within Boost such as documenting the concepts used would be a great help. I'd like to thank Andy Little for the work done so far on PQS and also for responding rapidly to the prolific discussion on the developer's list. -Fred Bertsch

4 5

Re: [Boost-users] 'standard' free functions for operators, etc
by Gottlob Frege 22 Jun '06

22 Jun '06

> From: "Peter Dimov" <pdimov(a)mmltd.net> > intrusive_ptr_add_ref is passed a pointer to a class with an embedded > reference count. It doesn't matter how many typedefs resolve to X*, X > still > has its reference count that needs to be incremented. > > You only run into the problem when "assign" is underspecified and > overloaded > to mean different things in different contexts. I understand your point, and maybe it is not a problem for intrusive_ptr in all pratical cases, but intrusive_ptr_add_ref is passed a T *. T may or may NOT be a class. ie T could be an int or an HWND (which is often just void *, depending on #define STRICT) or some other handle, or whatever. Maybe all those examples are a bit ugly - ie the int would have to be something like an index into a global array, or something like that. But it is not necessarily a distinct 'class' that holds its refcount intrinsicly. Or maybe it is necessarily that, but only because of how intrusive_ptr_add_ref was done, not for any other reason. So intrusive_ptr is ever so slightly more restrictive than it need be. Probably not a big deal. Note also, that for those few strange cases, you could also just wrap the HWND or whatever into a class, but then of course, intrusive_ptr could have just required the class to have a specific add_ref member function (and then you might need to wrap all classes). So maybe intrusive_ptr is just an example and not necessarily the best one, but I still wonder 1. if policy based could have been the same for 95% of the cases, and better for the last 5%, 2. there are any other techniques to consider. Either for intrusive_ptr or boost iterators or other places where I want to use the same techniques. I think that's the real question - if I am going to borrow some idioms for boost or wherever, I'd like to borrow the right ones for the right situations. Tony

2 1

linking
by Tamas Hegedus 22 Jun '06

22 Jun '06

Hi, I am newbie to c/cpp and I have linking and/or loading shared library problem. I can not link with dynamic libraries. Thanks for your help in advance, Tamas *** 1 *** [violin]$ ldconfig -p | grep libboost_program_options libboost_program_options.so.1 (libc6) => /usr/lib/libboost_program_options.so.1 *** 2 *** As I do not have the file /usr/lib/libboost_program_options.so.1 I created a link: libboost_program_options.so.1 -> libboost_program_options-gcc-1_33_1.so.1.33.1 *** 3 *** Compiling my src file [violin]$ g++ -Wall -I /usr/local/include/boost-1_33_1 -c test.cpp *** 4A *** [violin]$ g++ -Wall -o test test.o /usr/local/lib/libboost_program_options-gcc-1_33_1.so.1.33.1 [violin]$ ./test ./test: error while loading shared libraries: libboost_program_options-gcc-1_33_1.so.1.33.1: cannot open shared object file: No such file or directory *** 4B *** [violin]$ g++ -Wall -o test test.o /usr/local/lib/libboost_program_options.so.1 [violin]$ ./test ./test: error while loading shared libraries: libboost_program_options-gcc-1_33_1.so.1.33.1: cannot open shared object file: No such file or directory *** 4C *** [violin]$ g++ -Wall -o test test.o /usr/local/lib/libboost_program_options-gcc-1_33_1.a [violin]$ ./test Compression level was not set. -- Tamas Hegedus, PhD | phone: (1) 919-966 0329 UNC - Biochem & Biophys | fax: (1) 919-966 5178 6107 Thurston-Bowles Bldg | mailto:hegedus@med.unc.edu Chapel Hill, NC, 27599-7248 | http://biohegedus.org

3 2

Re: [Boost-users] Trap in regex library (help requested)
by Marty Eaves 22 Jun '06

22 Jun '06

Hi John, When I attempted to isolate the code and make a simple VC7 project, I did not have the problem any longer either. I think I have found a solution to the problem I was trying to solve, so I will continue to work in that direction and get back with you if I have any issues... Thanks for the update on the regression test. Marty ________________________________ Marty Eaves STAFF SOFTWARE ENGINEER CALL PROCESSING & MEDIA SERVICES RESEARCH AND DEVELOPMENT INTERVOICE, INC. P: (972) 454-8348 marty.eaves(a)intervoice.com ________________________________ Intervoice: Connecting People and Information. -----Original Message----- From: boost-users-bounces(a)lists.boost.org [mailto:boost-users-bounces@lists.boost.org] On Behalf Of John Maddock Sent: Friday, June 16, 2006 6:58 PM To: boost-users(a)lists.boost.org Subject: Re: [Boost-users] Trap in regex library (help requested) Marty Eaves wrote: > John, > > Thanks, I'll look into the STL compatibility. We were using STL port > with VC6, but I think we are using the Microsoft STL with VC7. I'll > double check. > The test case failing is using the above grammar with the value of > into > the "inputValue" parameter set to "2" in the code snippet below: If you want me to look at that further you'll need to break it down into a nice easy test case, that I can just compile and test right off. BTW I've just run the main regression test programs with VC7.1 and managed code, and captures enabled, and everything checks out OK. John. _______________________________________________ Boost-users mailing list Boost-users(a)lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users This e-mail transmission may contain information that is proprietary, privileged and/or confidential and is intended exclusively for the person(s) to whom it is addressed. Any use, copying, retention or disclosure by any person other than the intended recipient or the intended recipient's designees is strictly prohibited. If you are the intended recipient, you must treat the information in confidence and in accordance with all laws related to the privacy and confidentiality of such information. If you are not the intended recipient or their designee, please notify the sender immediately by return e-mail and delete all copies of this email, including all attachments.

1 0

Access violation in boost::condition
by Eric Colleu 22 Jun '06

22 Jun '06

Hi, I'm using boost 1.33.1 on multi-processor plateform (2 Pentium4 3.2 GHz) and Windows XP SP2 (compiled with MinGW). I wrote a MessageQueue class which use boost::mutex and boost::condition to wait for message and notify message reception (see code below) I made a little test that create 2 boost::thread sending and receiving messages in 2 queues. The access violation is produced in the "WaitForSingleObject" method called by "condition_impl::do_wait" Have you ever meet this problem? Eric Colleu ------messageQueue.cpp-----START #include <iostream> #include "messageQueue.hpp" namespace util { MessageQueue::MessageQueue(void) : m_mutex() , m_condition() , m_queue() , m_cancel(false) { } MessageQueue::~MessageQueue(void) throw() { } void MessageQueue::push(MessageHandle message) { boost::mutex::scoped_lock lock(m_mutex); m_queue.push(message); m_condition.notify_all(); } MessageHandle MessageQueue::pop(const boost::xtime& delay) { boost::mutex::scoped_lock lock(m_mutex); while (m_queue.size() == 0) { if (!m_condition.timed_wait(lock, delay)) { throw MsgTimeout("timeout in pop message"); } } if(m_cancel) { m_cancel = false; throw MsgWaitCancelled("Wait Event cancelled"); } MessageHandle msg = m_queue.front(); m_queue.pop(); return msg; } MessageHandle MessageQueue::pop(void) { boost::mutex::scoped_lock lock(m_mutex); while (m_queue.size() == 0 && ! m_cancel) { m_condition.wait(lock); } if(m_cancel) { m_cancel = false; throw MsgWaitCancelled("Wait Event cancelled"); } MessageHandle msg(m_queue.front()); m_queue.pop(); return msg; } int MessageQueue::size(void) { boost::mutex::scoped_lock lock(m_mutex); return m_queue.size(); } void MessageQueue::cancel_wait(void) { boost::mutex::scoped_lock lock(m_mutex); m_cancel = true; m_condition.notify_all(); } } ------messageQueue.cpp-----END ------messageQueue.hpp-----START #if defined(CORE_UTIL_QUEUE_MESSAGEQUEUE_CYCLE) #error Header cyclic inclusion detected. #endif // defined(CORE_UTIL_QUEUE_MESSAGEQUEUE_CYCLE) #define CORE_UTIL_QUEUE_MESSAGEQUEUE_CYCLE #if !defined(CORE_UTIL_QUEUE_MESSAGEQUEUE_HPP) #define CORE_UTIL_QUEUE_MESSAGEQUEUE_HPP // =================================================================== #include <queue> #include <boost/shared_ptr.hpp> #include <boost/thread.hpp> #include <boost/thread/condition.hpp> #include "core/exceptions.hpp" #include "message.hpp" namespace util { /** File d'attente de messages. * Cette classe fonctionne comme une FIFO, et permet de * faire une attente bloquante d'un message. */ class MessageQueue { public: MessageQueue(void); virtual ~MessageQueue(void) throw(); /** Empile un message en fin de liste. * @param message Le message a ajouter. */ void push(boost::shared_ptr<Message> message); /** Attend un message pendant un temps déterminé. * @param delay Temps d'attente du message. * @return Le message si un message a été reçu * @throws MsgTimeout Si aucun message n'a été empilé pendant le temps défini en paramètre. * @throws MsgWaitCancelled Si l'attente a été interrompue par un cancel_wait. */ MessageHandle pop(const boost::xtime& delay); /** Attend un message infiniment. * @return Le message si un message a été reçu * @throws MsgWaitCancelled Si l'attente a été interrompue par un cancel_wait. */ MessageHandle pop(void); /** Retourne la taille de la file d'attente. * @return le nombre d'éléments restant dans la file d'attente. */ int size(void); /** Interrompt une attente de message. */ void cancel_wait(void); private: boost::mutex m_mutex; boost::condition m_condition; std::queue<MessageHandle > m_queue; bool m_cancel; }; } // =================================================================== #endif // !defined CORE_UTIL_QUEUE_MESSAGEQUEUE_HPP #undef CORE_UTIL_QUEUE_MESSAGEQUEUE_CYCLE ------messageQueue.hpp-----END ------message.hpp-----START #if defined(CORE_UTIL_QUEUE_MESSAGE_CYCLE) #error Header cyclic inclusion detected. #endif // defined(CORE_UTIL_QUEUE_MESSAGE_CYCLE) #define CORE_UTIL_QUEUE_MESSAGE_CYCLE #if !defined(CORE_UTIL_QUEUE_MESSAGE_HPP) #define CORE_UTIL_QUEUE_MESSAGE_HPP // =================================================================== #include <string> #include <boost/shared_ptr.hpp> namespace util { /** * La classe Message représente les message transittant dans les MessageQueue, * entre le gestionnaire (device::AbstractDeviceManager) et la pile de communication * (stack_manager::SimpleStack). * Ces messages peuvent être des requêtes, réponses ou événements. */ class Message { public: Message(std::string id, bool is_event) : m_id(id), m_isEvent(is_event) {}; virtual ~Message(void) throw() {}; /** Indique si le message est un événement. * @return * - true : s'il s'agit d'un événement * - false sinon. */ bool isEvent(void) {return m_isEvent;}; /** Retourne l'identifiant du message. * @return une chaîne de caractère représentant l'identifiant du message. */ std::string getId(void) {return m_id;}; /** Encode les données à transmettre à la pile de communication sous forme d'une chaîne. */ virtual std::string encode(void) = 0; virtual void decode(void) = 0; protected: std::string m_id; private: bool m_isEvent; }; /** Pointeur sur un Message. */ typedef boost::shared_ptr<Message> MessageHandle; } // =================================================================== #endif // !defined CORE_UTIL_QUEUE_MESSAGE_HPP #undef CORE_UTIL_QUEUE_MESSAGE_CYCLE ------message.hpp-----END

2 2

[serialization] XML archive is order-dependent
by Stan Vasilyev 22 Jun '06

22 Jun '06

I'm using boost::archive::xml_iarchive to read data from an XML archive created by hand. It works as long as I keep all the tags in the same order that they are serialized in my code. As soon as I put in an extra XML tag somewhere, de-serialization fails. Since XML is order-independent, is it possible to make boost::serialization also order-independent? Stan

5 6

'standard' free functions for operators, etc
by Gottlob Frege 22 Jun '06

22 Jun '06

In many parts of boost, free functions are used as customization points for various things, better explained with examples: intrusive_ptr_add_ref() advance/increment/etc in boost iterator library etc Are there other examples? Now consider my example: template<T, U> myclass_assign(T *t, U const & u) { t->operator=(u); } That forms the default implementation, but clients can define their own specializations (for cases where T is a simple type and not an object at all). At first I thought that 'myclass_assign' should just be 'assign' since the assignment should be the same, just based on type (not on the circumstances of why 'myclass' is doing the assign), but (just while writing the email) I saw a problem there: in the case of, say: typedef int RecordID; assign(RecordID id, SomeValue const & value) { // do the assignment } But of course, RecordID is just a typedef, not a 'strong typedef', so it could be just one example of many assign(int, SomeValue), completely unrelated to each other. So how should that be handled? Specific naming (using 'myclass_') lowers the chances of a problem, but doesn't remove it. Which means that, in theory at least, intrusive_ptr has the same problem: intrusive_ptr_add_ref(some_typedef_that_is_really_an_int) { ... } intrusive_ptr_add_ref(some_OTHER_typedef_that_is_also_an_int) { ... } Should myclass, and also intrusive_ptr instead use a policy-based approach, similar to boost iterator? intrusive_ptr<T, policy = legacy_policy> { ... } Is there some other approach that I'm not considering (ie runtime boost::functions instead of compile time policy functions)? Tony

2 1

Re: [Boost-users] Boost serializer: incomplete type error reported
by Diego Molla 22 Jun '06

22 Jun '06

Thanks for your advice, Gennaro. I had to move the following includes to the beginning of the file: #include <boost/archive/text_oarchive.hpp> #include <boost/archive/text_iarchive.hpp> I also realised that the following gives the same error of incomplete type: void SetOfRules::save(const string &fileName) // Save into a file { ofstream fout(fileName.c_str()); boost::archive::text_oarchive oa(fout); oa << *this; fout.close(); } I had to use a helper function like this: void _save(const SetOfRules &sor, const string &fileName); void SetOfRules::save(const string &fileName) // Save into a file { _save(*this,fileName); } void _save(const SetOfRules &sor, const string &fileName) { std::ofstream fout(fileName.c_str()); boost::archive::text_oarchive oa(fout); oa << sor; fout.close(); } // save And similarly to load a class. Diego >>Hello, >> >>I have an error that is puzzling me. I have defined serialization of >>several custom classes, and everything worked perfectly until we've >>changed to a new Linux OS and had to reinstall Boost. Now the program >>doesn't compile and the error given is: >> >> error: incomplete type >>'boost::serialization::extended_type_info_null<SetOfRules>' used in >>nested name specifier >> >>I don't really understand why it says there is an incomplete type when >>there used to be no errors. The same error happens with all the classes >>where I've defined serialization. > > I'm not too much into the latest gcc developments (certainly others > will give a more accurate reply) but it looks like a two-phase lookup > issue: at the time your function is parsed SetOfRules is actually > incomplete, but the old gcc version you had didn't diagnose it and > delayed analisys to instantiation time. > > This is just my feeling. You might try moving SetOfRules' #include > before. > > --Gennaro. -- Do you want to know what is going on in Language Technology in Australia and New Zealand? Join ALTA -- http://www.alta.asn.au/ This message is intended for the addressee named and may contain confidential information. If you are not the intended recipient, please delete it and notify the sender. Views expressed in this message are those of the individual sender, and are not necessarily the views of Macquarie University. --------------------------------------------------------------------- Diego MOLLA ALIOD diego(a)ics.mq.edu.au Department of Computing http://www.ics.mq.edu.au/~diego Macquarie University

1 0

Re: [Boost-users] problem in multithreaded regex_replace on linux
by Ashish Varman 22 Jun '06

22 Jun '06

Dear John Apparantly, I am doing something wrong that is very elementary. I do not have any -mt version of the regex library. I even dispensed with STLport and tried to build boost first for default system libraries only. After giving the command "bjam -sTOOLS=gcc" in the directories boost_1_33_1/libs/regex/build and boost_1_33_1/libs/thread/build, I have the following libraries: $ find | grep "\.a$" ./libs/thread/build/bin-stage/libboost_thread-gcc-mt-d-1_33_1.a ./stage/lib/libboost_regex-gcc-d-1_33_1.a ./stage/lib/libboost_regex-gcc-1_33_1.a ./bin/boost/libs/regex/build/libboost_regex.a ./bin/boost/libs/regex/build/libboost_regex.a/gcc/debug/libboost_regex-gcc-d-1_33_1.a ./bin/boost/libs/regex/build/libboost_regex.a/gcc/release/libboost_regex-gcc-1_33_1.a ./bin/boost/libs/thread/build/libboost_thread.a ./bin/boost/libs/thread/build/libboost_thread.a/gcc/debug/threading-multi/libboost_thread-gcc-mt-d-1_33_1.a I have not been able to find how to build -mt versions of the regex library. Apparantly, the threads library is built so by default. Also, while trying to compile static_mutex_test.cpp, a lot of undefined references crop up even when I link with both boost::regex and boost::thread. The command was $ g++ -I ~/Software/boost_1_33_1 -L ~/Software/boost_1_33_1/libs/thread/build/bin-stage/ -L ~/Software/boost_1_33_1/stage/lib/ -lboost_thread-gcc-mt-d-1_33_1 -lboost_regex-gcc-1_33_1 -pthread static_mutex_test.cpp /tmp/ccvkfjk2.o(.text+0x16): In function `failed_test(char const*, int, char const*)': : undefined reference to `boost::scoped_static_mutex_lock::scoped_static_mutex_lock[in-charge](boost::static_mutex&, bool)' /tmp/ccvkfjk2.o(.text+0xa8): In function `failed_test(char const*, int, char const*)': : undefined reference to `boost::scoped_static_mutex_lock::~scoped_static_mutex_lock [in-charge]()' ----------- snip -------------- /tmp/ccvkfjk2.o(.text+0x5da): In function `t3()': : undefined reference to `boost::scoped_static_mutex_lock::~scoped_static_mutex_lock [in-charge]()' /tmp/ccvkfjk2.o(.text+0x5f4): In function `t3()': : undefined reference to `boost::scoped_static_mutex_lock::~scoped_static_mutex_lock [in-charge]()' collect2: ld returned 1 exit status Since the scoped_static_mutex_lock is defined only in the regex sources, I think the errors are due to the -mt versions not being built. I have verified that -mt versions are being built and used on windows. I tried using the same technique (building using makefiles) but no luck $ make -f gcc.mak $ find | grep "\.a$" ./gcc/libboost_regex-gcc-1_33_1.a ./gcc/libboost_regex-gcc-d-1_33_1.a So now my main problem becomes "How do I build -mt versions of regex on linux". I hope the answer does not make me look too dumb. Thanks in advance Yours Sincerely Ashish Varman

2 1