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

RE: [Boost-Users] bind: I just keep trying...
by Tom Matelich 26 Sep '02

26 Sep '02

I think the best solution for you would be to put boost::function<short>'s in your map. Look at the example in the bind documentation at http://www.boost.org/libs/bind/bind.html#with_boost_function Tom > -----Original Message----- > From: dick.bridges(a)tais.com [mailto:dick.bridges@tais.com] > Sent: Thursday, September 26, 2002 7:57 AM > To: Boost-Users(a)yahoogroups.com > Subject: RE: [Boost-Users] bind: I just keep trying... > > > > I'm trying to store member function pointers. For purposes of this > discussion, the key is an int. > > <context> > X is the virtual base class for polymorphic message classes. > all messages have a session-d attribute > > Ys are instances of classes derived from a virtual message hanler base > class. > foo() is the message handler function. > > Z is the 'message pump' that dispatches messages to Ys based upon the > session-id. > > Y instances insert make_pair(session-id, message_handler) > pairs into the > map. > (I think _1 is 'this' during the insert.) > > Z calls the Y instance message-handler using the session-id > key to the > map. > </context> > > > > > > > Mark Storer > > > <mstorer@cardiff To: > "'Boost-Users(a)yahoogroups.com'" <Boost-Users(a)yahoogroups.com> > > .com> cc: > > > Subject: RE: > [Boost-Users] bind: I just keep trying... > > 09/25/2002 04:47 > > > PM > > > Please respond > > > to Boost-Users > > > > > > > > > > > > > I'm a little confused as to what you're aiming for here... a > map of ints to > function pointers (that can point to Y::foo)? > > This should do the trick, provided I'm recalling my function pointer > declarations correctly > std::map<int, short()(*X)> > > Further boost::bind creates an INSTANCE of a functor. Giving > the compiler > an instance when it wants a type is not a good way to keep it happy. > > Now if what you're actually after is calling Y::foo on a > colleciton of Y's, > then you don't need a collection of function pointers for > each instance in > the first place, which is where bind comes in. According to my flimsy > understanding of bind (I haven't used it yet), this should do > the trick: > > X someX; > std::vector<Y> &wise = BuildYs(); // wise == plural of Y > > std::for_each( > wise.begin(), > wise.end(), > boost::bind<short>(&Y::foo, &someX) > ); > > If you want the X parameter to vary from one call of foo to > the next, I > suggest rolling your own loop... or incorporating the X* into > Y, at which > point you can get rid of the X* parameter, and use > boost::bind(&Y::foo), or > std::mem_fun(&Y::foo). > > --Mark Storer > Software Engineer > Cardiff Software > > #include <disclaimer> > typedef std::disclaimer<Cardiff> Discard; > > > -----Original Message----- > From: dick.bridges(a)tais.com [mailto:dick.bridges@tais.com] > Sent: Wednesday, September 25, 2002 4:08 PM > To: boost-users(a)yahoogroups.com > Subject: [Boost-Users] bind: I just keep trying... > > > Everytime I think I have a clue, I'm reminded how little I REALLY > understand. > > <error compiler='VC6SP4'> > 'map' : invalid template argument for '_Ty', type expected > </error> > > <snippet> > struct X > { > } > > struct Y > { > short foo(X*); > } > > struct Z > { > map<int, bind(&Y::foo, _1, _2) > myMap; > } > </snippet> > > I thought I saw how to do this somewhere on the website but I > sure can't > find it now. Any help GREATFULLY appreciated. > > > > Info: <http://www.boost.org> > Wiki: <http://www.crystalclearsoftware.com/cgi-bin/boost_wiki/wiki.pl> > Unsubscribe: <mailto:boost-users-unsubscribe@yahoogroups.com> > > > Your use of Yahoo! Groups is subject to > http://docs.yahoo.com/info/terms/ > > > > Info: <http://www.boost.org> > Wiki: <http://www.crystalclearsoftware.com/cgi-bin/boost_wiki/wiki.pl> > Unsubscribe: <mailto:boost-users-unsubscribe@yahoogroups.com> > > > Your use of Yahoo! Groups is subject to > http://docs.yahoo.com/info/terms/ > > > > > > > > ------------------------ Yahoo! Groups Sponsor > ---------------------~--> > Home Selling? Try Us! > http://us.click.yahoo.com/QrPZMC/iTmEAA/MVfIAA/EbFolB/TM > -------------------------------------------------------------- > -------~-> > > Info: <http://www.boost.org> > Wiki: <http://www.crystalclearsoftware.com/cgi-bin/boost_wiki/wiki.pl> > Unsubscribe: <mailto:boost-users-unsubscribe@yahoogroups.com> > > > Your use of Yahoo! Groups is subject to > http://docs.yahoo.com/info/terms/ > > ----------------------------------------------------------------------- DISCLAIMER: Information contained in this message and/or attachment(s) may contain confidential information of Zetec, Inc. If you have received this transmission in error, please notify the sender by return email. -----------------------------------------------------------------------

1 0

intrusive_ptr
by nobody 26 Sep '02

26 Sep '02

Hi! There is a intrusive_ptr.cpp in boost 1.28, but I did not find any doc for that... exept for in the timing . Why? Guessing from the name, intrusive_ptr requires the pointed-to object to store the reference count or something like that, doesn't it? And what does anybody think of typedefing std::auto_ptr to boost::single_ptr? Finally I have a suggestion for a new type of smart pointer one could call "notifying_ptr": Sometimes all weak pointing objects need to be notified when the object pointed to is destroyed: e.g. in a game some units are set to attack some building, so they store a weak_ptr to the building, but when that building is destroyed, it would be easier to implement and faster performing, if the units got a message so they can look for a new target, instead of having strong pointers, marking the building as dead, and polling for that state every time the unit does something. It could be quite some time until all units have realized that the building is gone and it really can be released from memory. Regards

2 2

RE: [Boost-Users] bind: I just keep trying...
by dick.bridges＠tais.com 26 Sep '02

26 Sep '02

I'm trying to store member function pointers. For purposes of this discussion, the key is an int. <context> X is the virtual base class for polymorphic message classes. all messages have a session-d attribute Ys are instances of classes derived from a virtual message hanler base class. foo() is the message handler function. Z is the 'message pump' that dispatches messages to Ys based upon the session-id. Y instances insert make_pair(session-id, message_handler) pairs into the map. (I think _1 is 'this' during the insert.) Z calls the Y instance message-handler using the session-id key to the map. </context> Mark Storer <mstorer@cardiff To: "'Boost-Users(a)yahoogroups.com'" <Boost-Users(a)yahoogroups.com> .com> cc: Subject: RE: [Boost-Users] bind: I just keep trying... 09/25/2002 04:47 PM Please respond to Boost-Users I'm a little confused as to what you're aiming for here... a map of ints to function pointers (that can point to Y::foo)? This should do the trick, provided I'm recalling my function pointer declarations correctly std::map<int, short()(*X)> Further boost::bind creates an INSTANCE of a functor. Giving the compiler an instance when it wants a type is not a good way to keep it happy. Now if what you're actually after is calling Y::foo on a colleciton of Y's, then you don't need a collection of function pointers for each instance in the first place, which is where bind comes in. According to my flimsy understanding of bind (I haven't used it yet), this should do the trick: X someX; std::vector<Y> &wise = BuildYs(); // wise == plural of Y std::for_each( wise.begin(), wise.end(), boost::bind<short>(&Y::foo, &someX) ); If you want the X parameter to vary from one call of foo to the next, I suggest rolling your own loop... or incorporating the X* into Y, at which point you can get rid of the X* parameter, and use boost::bind(&Y::foo), or std::mem_fun(&Y::foo). --Mark Storer Software Engineer Cardiff Software #include <disclaimer> typedef std::disclaimer<Cardiff> Discard; -----Original Message----- From: dick.bridges(a)tais.com [mailto:dick.bridges@tais.com] Sent: Wednesday, September 25, 2002 4:08 PM To: boost-users(a)yahoogroups.com Subject: [Boost-Users] bind: I just keep trying... Everytime I think I have a clue, I'm reminded how little I REALLY understand. <error compiler='VC6SP4'> 'map' : invalid template argument for '_Ty', type expected </error> <snippet> struct X { } struct Y { short foo(X*); } struct Z { map<int, bind(&Y::foo, _1, _2) > myMap; } </snippet> I thought I saw how to do this somewhere on the website but I sure can't find it now. Any help GREATFULLY appreciated. Info: <http://www.boost.org> Wiki: <http://www.crystalclearsoftware.com/cgi-bin/boost_wiki/wiki.pl> Unsubscribe: <mailto:boost-users-unsubscribe@yahoogroups.com> Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/ Info: <http://www.boost.org> Wiki: <http://www.crystalclearsoftware.com/cgi-bin/boost_wiki/wiki.pl> Unsubscribe: <mailto:boost-users-unsubscribe@yahoogroups.com> Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/

1 0

std::swap() vs boost::swap()
by Galante Michele 26 Sep '02

26 Sep '02

Boost classes that define a swap() member function usually also provide a global swap() function: namespace boost { class X { void swap(X &rhs); }; template<class T> void swap(X &lhs, X &rhs) { lhs.swap(rhs); } } Why is the global swap() function defined in namespace boost and not as a specialization of std::swap<>() in namespace std? Suppose I have a template function like the following: #include <algorithms> template <typename T> void do_something_and_swap(T &a, T &b) { // do something ... std::swap(a, b); } I want this template function to work with every type that is "swappable", but if I try to use it with boost::scoped_ptr<> the compiler fails because std::swap<>() is not specilized for boost::scoped_ptr<> and the generic definition of std::swap<>() requires the argument type to be copyable (and boost::scoped_ptr<> is noncopyable). I think I have read somewhere (but perhaps I'm wrong) that specialization of std::swap<>() for user defined types is allowed by the standard, so is there any other reason for not defining the global swap functions in namespace std? Michele Galante Zucchetti Centro Sistemi s.p.a. m.galante(a)centrosistemi.it

2 1

RE: [Boost-Users] std::swap() vs boost::swap()
by scleary＠jerviswebb.com 26 Sep '02

26 Sep '02

> From: Galante Michele [mailto:M.Galante@centrosistemi.it] > > Suppose I have a template function like the following: > > #include <algorithms> > > template <typename T> > void do_something_and_swap(T &a, T &b) > { > // do something ... > std::swap(a, b); > } > > I want this template function to work with every type that is "swappable", > but if I try to use it with boost::scoped_ptr<> the compiler fails because > std::swap<>() is not specilized for boost::scoped_ptr<> and the generic > definition of std::swap<>() requires the argument type to be copyable (and > boost::scoped_ptr<> is noncopyable). The best way to provide this solution is to use an unqualified swap; Koenig lookup will ensure that the correct swap() is found for std:: and boost:: components. > I think I have read somewhere (but perhaps I'm wrong) that specialization of > std::swap<>() for user defined types is allowed by the standard, so is there > any other reason for not defining the global swap functions in namespace > std? Technically, specialization of std::swap is allowed. But overloading std::swap is not. These are often confused, but here's a concrete example: template <> void std::swap(boost::scoped_ptr<int> &, boost::scoped_ptr<int> &) is a specialization, and legal; but: template <typename T> void std::swap(boost::scoped_ptr<T> &, boost::scoped_ptr<T> &) is an overload, and not legal (currently). There's been a ton of discussion of this issue on the Boost.Developer's list; you can search the list archives if you're interested. -Steve

1 0

Forte compiler template error
by Tanmay Patwardhan 26 Sep '02

26 Sep '02

Hi, I was trying to use the tokeniser lib and encountered this error: "/home/patwarta/boost_1_28_0/boost/iterator_adaptors.hpp", line 819: Error: The base "boost::detail::iterator_adaptor_traits_gen<boost::detail::token_ iterator_base<const char*>, std::basic_string<char, std::char_traits<char>, std::allocator<char>>, const std::basic_string<char, std::char_traits<char>, std::allocator<char>>&, const std::basic_string<char, std::char_traits<char>, std::allocator<char>>*, std::forward_iterator_tag, int>::type" must be a previously defined class or struct. "XMLbParserWrapper.cxx", line 11: Where: While specializing "boost::iterator_adaptor<boost::detail::token_iterator_ba se<const char*>, boost::detail::tokenizer_policy<std::basic_string<char, std::char_traits<char>, std::allocator<char>>, boost::char_delimiters_separator<char, std::char_traits<char>>>, std::basic_string<char, std::char_traits<char>, std::allocator<char>>, const std::basic_string<char, std::char_traits<char>, std::allocator<char>>&, const std::basic_string<char, std::char_traits<char>, std::allocator<char>>*, std::forward_iterator_tag, int> Any ideas? Thanks in advance, Tanmay

2 1

RE: [Boost-Users] bind: I just keep trying...
by Mark Storer 25 Sep '02

25 Sep '02

I'm a little confused as to what you're aiming for here... a map of ints to function pointers (that can point to Y::foo)? This should do the trick, provided I'm recalling my function pointer declarations correctly std::map<int, short()(*X)> Further boost::bind creates an INSTANCE of a functor. Giving the compiler an instance when it wants a type is not a good way to keep it happy. Now if what you're actually after is calling Y::foo on a colleciton of Y's, then you don't need a collection of function pointers for each instance in the first place, which is where bind comes in. According to my flimsy understanding of bind (I haven't used it yet), this should do the trick: X someX; std::vector<Y> &wise = BuildYs(); // wise == plural of Y std::for_each( wise.begin(), wise.end(), boost::bind<short>(&Y::foo, &someX) ); If you want the X parameter to vary from one call of foo to the next, I suggest rolling your own loop... or incorporating the X* into Y, at which point you can get rid of the X* parameter, and use boost::bind(&Y::foo), or std::mem_fun(&Y::foo). --Mark Storer Software Engineer Cardiff Software #include <disclaimer> typedef std::disclaimer<Cardiff> Discard; -----Original Message----- From: dick.bridges(a)tais.com [mailto:dick.bridges@tais.com] Sent: Wednesday, September 25, 2002 4:08 PM To: boost-users(a)yahoogroups.com Subject: [Boost-Users] bind: I just keep trying... Everytime I think I have a clue, I'm reminded how little I REALLY understand. <error compiler='VC6SP4'> 'map' : invalid template argument for '_Ty', type expected </error> <snippet> struct X { } struct Y { short foo(X*); } struct Z { map<int, bind(&Y::foo, _1, _2) > myMap; } </snippet> I thought I saw how to do this somewhere on the website but I sure can't find it now. Any help GREATFULLY appreciated. Info: <http://www.boost.org> Wiki: <http://www.crystalclearsoftware.com/cgi-bin/boost_wiki/wiki.pl> Unsubscribe: <mailto:boost-users-unsubscribe@yahoogroups.com> Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/

1 0

bind: I just keep trying...
by dick.bridges＠tais.com 25 Sep '02

25 Sep '02

Everytime I think I have a clue, I'm reminded how little I REALLY understand. <error compiler='VC6SP4'> 'map' : invalid template argument for '_Ty', type expected </error> <snippet> struct X { } struct Y { short foo(X*); } struct Z { map<int, bind(&Y::foo, _1, _2) > myMap; } </snippet> I thought I saw how to do this somewhere on the website but I sure can't find it now. Any help GREATFULLY appreciated.

1 0

RE: [Boost-Users] pointer traints ?
by Mark Storer 25 Sep '02

25 Sep '02

Couldn't you do something along the lines of: template < typename PointerT, typeName RefereceT = &(*PointerT) > Only I think that's an address rather than a reference... How about: template < typename T > void f( T *ptr ) { T &ref = *ptr; // ... do something with ref ... } And what do you have against "->"? I don't think a reference's "." is any faster than a pointer's "->", so that shouldn't be a concern... just trying to save some knuckle grease? I don't get it. (I'm getting a lot of practice with that phrase since I started digging into STL & Boost) --Mark Storer Software Engineer Cardiff Software #include <disclaimer> typedef std::disclaimer<Cardiff> Discard; -----Original Message----- From: Bohdan [mailto:yg-boost-users@m.gmane.org] Sent: Monday, September 23, 2002 7:38 AM To: boost-users(a)yahoogroups.com Subject: [Boost-Users] pointer traints ? I have following code: template < typename PointerT > void f( PointerT ptr ) { ReferenceT ref = *ptr; // ... do something with ref ... } The problem is that ReferenceT is unknown inside f( ... ). If i had similar code for InteratorT instead of PointerT i was using iterator_traits. AFAIK there is no pointer traits for a raw and smart pointers and i can not use type_traits for smart pointers. So, the question is: "Is there some simple solution for this problem or this is just lack of pointer_traits in boost (std?) ?" regards, bohdan Info: <http://www.boost.org> Wiki: <http://www.crystalclearsoftware.com/cgi-bin/boost_wiki/wiki.pl> Unsubscribe: <mailto:boost-users-unsubscribe@yahoogroups.com> Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/

2 1

BoostPython and VC 7.1 Everett ( again )
by Bernhard Glueck 25 Sep '02

25 Sep '02

Hi again. John, i checked out the RC_1_29_0 branch a few hours ago, and tried to compile everything. Using bjam everything builds fine, EXCEPT Boost Python... all other libs work as expected using MSVC1310.... Here is the error i get for all the targets of boost python: vc-C++ libs\python\build\bin\boost_python.dll\vc7\debug\runtime-link-dynamic\typ es.obj types.cpp E:\sdk\boost\boost\boost\python\reference.hpp(213) : error C2888: 'boost::shared _ptr' : symbol cannot be defined within namespace 'python' E:\sdk\boost\boost\boost\python\reference.hpp(222) : see reference to cl ass template instantiation 'boost::python::reference<T>' being compiled Maybe this just is not fixed yet... or its an error of mine Till next time Bernhard [Non-text portions of this message have been removed]

1 0