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

Regex Access violation in BCB5
by Tim Daplyn 09 Oct '02

09 Oct '02

I've used a previous version of the boost regex library before and have just upgraded to the latest version. Built the libs and dll fine, but now I get a A/V when calling any boost code. I've even set up a vanilla console app with all the Borland default settings and the following code: #include "boost/regex.hpp" int main(int argc, char* argv[]) { boost::wregex regExpr2(L"bill.*"); bool result; result = boost::regex_match(L"freddy", regExpr2); result = boost::regex_match(L"billabong", regExpr2); return 0; } And it blows up on the first regex_match!! I can't figure out what's going on - I know it's linking to the _sdi version of the lib because it complains about it being missing if I remove it, and the dlls are on the path and are loaded as soon as this mini-app starts up. It makes no difference whether I build through the command line or in BCB GUI. Please help!! Thanks, Tim D.

1 0

building smart_ptr as a separate library
by Nik Kolev 07 Oct '02

07 Oct '02

Hi everybody: Is it possible to build/integrate just the smart_ptr part of boost. If yes I would greatly apreciate your suggestions on how this is to be accomplished. Thanks, Nik

2 1

RE: [Boost-Users] Iterator adapters and maps
by scleary＠jerviswebb.com 04 Oct '02

04 Oct '02

I think the problem is in the definition of Select2ndRef; if you use: template<class pair_type> struct Select2ndRef { typedef pair_type argument_type; typedef typename pair_type::second_type result_type; result_type & operator()(pair_type & source) const { return source.second; } }; then your code will compile. This has to do with an unusual meaning of "result_type" for projection iterator adaptors. -Steve

2 1

Use a shared_ptr as a 'backpointer'?
by dick.bridges＠tais.com 04 Oct '02

04 Oct '02

I run into situations where it's useful for an object to have a 'backpointer' to the object instance that created it (example below). Is there a pattern/idiom for doing this using shared_ptr? Or maybe I need to re-think the idea of the backpointer in the first place? class Bar; class Foo { public: Foo(Bar* b) : p(b){} private: Foo() : p(0){} shared_ptr<Bar> p; } class Bar { void doit() { new Foo(???); // <<< new Foo(this) obviously won't work. } }

3 2

[BGL]: 2 desgin questions
by Björn Lindberg 03 Oct '02

03 Oct '02

I have two questions regarding the design of the BGL. I'm interested in this to get a better understanding of the design choices, which influences how I use the BGL and how I'm building on it. 1) Why is a lot of the function arguments made by value rather than reference? As an example, consider the BFS algorithm, where the documentation states that "[1] Since the visitor parameter is passed by value, if your visitor contains state then any changes to the state during the algorithm will be made to a copy of the visitor object, not the visitor object passed in. Therefore you may want the visitor to hold this state by pointer or reference." It would be much easier to use the visitor objects for keeping state throughout the algorithm if it was passed by reference instead. 2) Non-constness. It seems impossible to use BGL together with code that is "const-correct". If, for instance, I have a function implementing some kind of examining algorithm on a graph (so the graph is passed to it as const), and this algorithm uses breadth_first_search with a visitor that doesn't modify the graph, it still can't be const when passed to breadth_first_search. This only leaves the options of either (i) casting away constness in the call, which is bad because it could introduce bugs later on, or (ii) leaving out const altogether which is not good either. Maybe I've just misunderstood something and this is possible, or perhaps it is impossible to implement BGL in this way? Björn

2 2

boost build on redhat7.3
by Alkis Evlogimenos 03 Oct '02

03 Oct '02

Hi, I just compiled boost from cvs. I am using an x86 with redhat linux 7.3. I also have STLport 4.5.3 installed. I compiled against STLport using this command: bjam "-sTOOLS=gcc-stlport" "-sSTLPORT_INCLUDE_DIRECTORY=include/stlport" "-sSTLPORT_LIB_DIRECTORY=lib" "-sSTLPORT_ROOT=/usr" and against gcc using the following: bjam "-sTOOLS=gcc" Both builds finish with success (compiling with j4 leaves 2 targets behind but they succeed after a reinvocation). Now I enter status/ and compile regression using the following: g++ -I.. regression.cpp -o regression Modify compiler.cfg to look like: linux gcc-296 GCC 2.96 g++ -c -Wall -W -ftemplate-depth-30 -I%include %source g++ -o boosttmp.exe -Wall -W -ftemplate-depth-30 -I%include %source -lrt GNU <a href="http://gcc.gnu.org/">GCC</a> 2.96 linux gcc-296-stlport GCC 2.96 STLport 4.5 g++ -c -Wall -W -ftemplate-depth-30 -isystem /usr/include/stlport -I%include %source -lstlport_gcc g++ -o boosttmp.exe -Wall -W -ftemplate-depth-30 -isystem /usr/include/stlport -I%include %source -lrt -lstlport_gcc GNU <a href="http://gcc.gnu.org/">GCC</a> 2.96 <a href="http://www.stlport.org/">STLport</a> 4.5 and finally run regression as follows: ./regression -o cs-rh7.3.html Attached is a gzipped log of the regression run. The text was 429Kb so I didn't paste it to this email :-) It seems that the testing frameworks has some problems (libs/config/test/config_test.cpp fails). Am I doing something wrong here? Any help is appreciated. Thanks, -- Alkis [Non-text portions of this message have been removed]

1 1

[BGL] Daisy chaining algorithms using vistor mechanism?
by Chris Russell 02 Oct '02

02 Oct '02

Still up hacking... Without cluttering the question with a ton of detail, basically I need to perform a DFS on a forest stored in a bidirectionalS adjacency_list and then perform a subgraph traversal in response to each finish_vertex invocation in the visitor. (The idea is to push information down the branches towards the root of a tree in the forest). Calling in_edges(v,g) from within the finish_vertex method in a dfs_visitor will not compile however. This is because the graph is passed by const reference into the visitor method and in_edges requires a non-const graph reference? (this seems similar in spirit to a question I posted a week or two ago about boost::get(some_property, Graph) from within a visitor). Generally, I would like to be able to daisy chain graph algorithms - that is invoke a sub-algorithm from within a visitor method called by a controlling algorithm. This would require that the sub-algorithm not change the graph topology, trample on properties used by its controller (e.g. coloring), or otherwise invalidate vertex descriptors or iterators so some care would need to be taken in designing the sub-algorithm. But it seems like a useful mechanism for extension. Am I missing something about getting at the in_edges given that I'm working with a const graph_t& in the context of the "controlling" DFS' visitor method? Why is this so restrictive? I'm guessing it might have something to do with having multiple copies of vertex and edge iterators floating around but I'm not sure. Will you comment on this please Jeremy? Thank you. - Chris Russell

5 6

RE: [Boost-Users] MPL?
by Mark Storer 02 Oct '02

02 Oct '02

"Modern C++ Design" has a pretty good coverage so far (I'm on chapter 6 out of 11). As for impressing the ladies... well... you could always show them your stamp collection. --Mark Storer Software Engineer Cardiff Software #include <disclaimer> typedef std::disclaimer<Cardiff> Discard; -----Original Message----- From: Jeff Faust [mailto:jeff@opticalres.com] Sent: Wednesday, October 02, 2002 12:37 PM To: Boost-Users(a)yahoogroups.com Subject: RE: [Boost-Users] MPL? Well, I didn't know I needed boost until I saw it. Same for STL several years back. 'Metaprogramming' comes up a lot on the boost list, but I have no understanding about what it's for--what problems it solves. The documentation is sparse. Then this thread comes up and everybody talks about it like it's obvious. I dislike being baffled. But mostly it's curiosity. I need something to talk about at parties and impress the ladies. MPL will do that, right? Right? Hello? Jeff -----Original Message----- From: David B. Held [mailto:dheld@codelogicconsulting.com] Sent: Wednesday, October 02, 2002 12:14 PM To: Boost-Users(a)yahoogroups.com Subject: Re: [Boost-Users] MPL? ----- Original Message ----- From: "Jeff Faust" <jeff(a)opticalres.com> To: <Boost-Users(a)yahoogroups.com> Sent: Wednesday, October 02, 2002 11:32 AM Subject: [Boost-Users] MPL? > [...] > Can anybody explain in English what the following does? I believe that Aleksey posted the runtime pseudo-equivalent. > Posted by David Abrahams on boost developers list: > > Have I mentioned recently how totally cool MPL's lambda facility is? > > I can't believe I can do things like this, even with MSVC6. It makes my > > metaprograms SO much easier to manage! > > mpl::logical_and< > > is_reference_to_class<add_reference<mpl::_> > > > , mpl::logical_not< > > is_reference_to_args<add_reference<mpl::_> > [Aleksey, on 9/27] // "ideal" version, for comparison purposes: // is_reference_to_class( add_reference(_) ) // && !is_reference_to_args( add_reference(_) ) I can't say for sure what the code does myself, but if I could venture a guess, I would say that it's a condition for something else (maybe an iterative algorithm over a type container?) that checks to see if an argument is a reference type of various sorts. Basically, if you don't understand it, in this case, you probably don't need it. ;) I know I don't! Dave 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/

3 2

RE: [Boost-Users] MPL?
by Randy Bowen 02 Oct '02

02 Oct '02

Metaprogramming techniques are discussed pretty thoroughly in Andrei Alexandescru's "Modern C++ Design" and in Czarnecki and Eisenecker's "Generative Programming". --Randy -----Original Message----- From: Jeff Faust [mailto:jeff@opticalres.com] Sent: Wednesday, October 02, 2002 12:37 PM To: Boost-Users(a)yahoogroups.com Subject: RE: [Boost-Users] MPL? Well, I didn't know I needed boost until I saw it. Same for STL several years back. 'Metaprogramming' comes up a lot on the boost list, but I have no understanding about what it's for--what problems it solves. The documentation is sparse. Then this thread comes up and everybody talks about it like it's obvious. I dislike being baffled. But mostly it's curiosity. I need something to talk about at parties and impress the ladies. MPL will do that, right? Right? Hello? Jeff -----Original Message----- From: David B. Held [mailto:dheld@codelogicconsulting.com] Sent: Wednesday, October 02, 2002 12:14 PM To: Boost-Users(a)yahoogroups.com Subject: Re: [Boost-Users] MPL? ----- Original Message ----- From: "Jeff Faust" <jeff(a)opticalres.com> To: <Boost-Users(a)yahoogroups.com> Sent: Wednesday, October 02, 2002 11:32 AM Subject: [Boost-Users] MPL? > [...] > Can anybody explain in English what the following does? I believe that Aleksey posted the runtime pseudo-equivalent. > Posted by David Abrahams on boost developers list: > > Have I mentioned recently how totally cool MPL's lambda facility is? > > I can't believe I can do things like this, even with MSVC6. It makes my > > metaprograms SO much easier to manage! > > mpl::logical_and< > > is_reference_to_class<add_reference<mpl::_> > > > , mpl::logical_not< > > is_reference_to_args<add_reference<mpl::_> > [Aleksey, on 9/27] // "ideal" version, for comparison purposes: // is_reference_to_class( add_reference(_) ) // && !is_reference_to_args( add_reference(_) ) I can't say for sure what the code does myself, but if I could venture a guess, I would say that it's a condition for something else (maybe an iterative algorithm over a type container?) that checks to see if an argument is a reference type of various sorts. Basically, if you don't understand it, in this case, you probably don't need it. ;) I know I don't! Dave 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 ADVERTISEMENT <http://rd.yahoo.com/M=233351.2428261.3848243.2225242/D=egroupweb/S=17050067…> <http://us.adserver.yahoo.com/l?M=233351.2428261.3848243.2225242/D=egroupmai…> 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 the Yahoo! Terms of Service <http://docs.yahoo.com/info/terms/> . [Non-text portions of this message have been removed]

1 0

Iterator adapters and maps
by Alisdair Meredith 02 Oct '02

02 Oct '02

I have a std::map<std::string, MyObject> and want to iterate over it calling several member functions of MyObject. This seems like a good case to learn about transform/projection iterator, but I am having a hard time getting a compiling example. My most recent attempt looks like this: class MyObject { public: void CommitChanges(); void RollbackChanges(); // More implementation details would follow }; typedef std::map<std::string, MyObject> MyMap; typedef MyMap::iterator MyIterator; typedef MyMap::value_type mapped_type; typedef boost::projection_iterator_generator< Select2ndRef<mapped_type>, MyIterator >::type projection; MyMap theMap; projection begin( theMap.begin() ); projection end( theMap.begin() ); for_each( begin, end, mem_fun_ref( MyObject::CommitChanges ) ); I also have the following two templates, although I suspect their equivalent must be lurking in boost somewhere: template<class pair_type> struct Select1stRef { typedef typename pair_type::first_type & result_type; result_type operator()( pair_type &source ) const { return source.first; } }; template<class pair_type> struct Select2ndRef { typedef typename pair_type::second_type & result_type; result_type operator()( pair_type &source ) const { return source.second; } }; I am compiling using BCB5 compiler in the IDE, should be equivalent to the free 5.5.1 command line compiler. While I could simply write a forwarding functor in a fraction of the time it will take me to solve this, there is definitely a ne tool I would like to add to my toolbox here once I understand what is happenning. Unfortunately I am getting somewhat lost in reference-to-reference issues, and the occasional compiler crash does not help. Any and all hints appreciated -- AlisdairM

1 0

Loki vs. Boost
by Hamlet 02 Oct '02

02 Oct '02

Hi, I'm (desperately) looking for a serious comparative review of Loki vs. Boost libraries. Although the differences are quite clear, still it would be nice to see a serious comparison, especially in the features that exists in both places, i.e. smart pointers, small obj. allocator, type traits and template meta-programming. Does anybody knows if there exists such a paper somewhere, or alternatively have his own insights about the libraries? Thanks, Ariel

2 1

[BGL] vertex_index_t w/VertexList=listS questions
by Chris Russell 02 Oct '02

02 Oct '02

Hi Jeremy, Using adjacency_list w/VertexList=listS, I've stuffed a vertex_index_t declaration into the internal vertex property map and got my DFS invocation to compile. Taking a look at the dijkstra-example-listS.cpp in the examples directory on CVS, I see how you manually initialize this property for each of the vertices after you build up the graph. p.230 of the book discusses the auto-update of the vertex_index_t property upon vertex removal for the case of VertexList=vecS but I'll need to affect this update myself for VertexList=listS? Also, can you foresee any problems if I register vertex_index_t with a value_type of long int instead of int? I'm guessing no because I'm passing it as a template parameter but I just want to make sure. - Thanks and regards Chris Russell

2 1

BOOST_PARAM_TEST_CASE and uncaught exceptions
by Alisdair Meredith 02 Oct '02

02 Oct '02

I have a case where my test case passed to BOOST_PARAM_TEST_CASE is throwing an exception I was not expecting. This seems to still result in the correct number of calls to the test case, but the first parameter in the list is passed every time. Is this a feature of my compiler, or an issue with the library? Using Borland C++ Builder 5 (should be same as free command-line compiler) -- AlisdairM

2 1

BUG: intrusive pointer and weak_ptr
by Petr Ferschmann 01 Oct '02

01 Oct '02

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hello, I think I found BUG. Let we have this program: - -- cut here -- #include <boost/shared_ptr.hpp> #include <boost/weak_ptr.hpp> #include <boost/intrusive_ptr.hpp> #include <cstdio> using namespace std; #define INTR class MujObjekt #ifdef INTR : public boost::counted_base #endif { public: MujObjekt() { printf( "MujObjekt::MujObjekt()\n" ); } void setThis( boost::shared_ptr<MujObjekt> obj ) { m_this = obj; } virtual ~MujObjekt() { printf( "MujObjekt::~MujObjekt()\n" ); } boost::weak_ptr<MujObjekt> m_this; }; int main() { boost::shared_ptr<MujObjekt> obj( new MujObjekt() ); obj->setThis( obj ); return( 0 ); } - --- cut here --- But in case of usage of intrusive pointer it doesn't call destructor - because there is weak pointer. But when I undefine INTR and it uses non-intrusive it works as expected. This is only modelling case to simulate bug. - -- Petr "Fers" Ferschmann -=[ petr(a)ferschmann.cz ]==[ http://petr.ferschmann.cz/ ]=- - -=[ Koukni na http://www.postcard.cz/ ]==[ +420 604/781 009 ]=- GPG Fingerprint: [83B0 6378 7A9D D993 035E 60BD FEEC F665 D2C8 1B9A] -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.7 (GNU/Linux) iD8DBQE9mV9O/uz2ZdLIG5oRArF/AJ9LGfWTXfy69zNTsbiEKThTPQnWcwCeOWgj 0tX9AxdWJ8u/L1MkdZHjOPk= =7A3i -----END PGP SIGNATURE-----

3 3

RE: [Boost-Users] Re: Use a shared_ptr as a 'backpointer'?
by dick.bridges＠tais.com 30 Sep '02

30 Sep '02

1) There are circumstances where the 'creator' should stick around as long as one or more 'creations' still exist (e.g., a [network] session that's created various asynchronous conversations). 2) I haven't had time to check, but I think this is the as-yet-not-but-soon-to-be-released implementation of intrusive pointers. Interesting: search google for "intrusive pointer" (quotes not required). I got five hits - all of the URLs start with www.boost.org/. BTW: Really slick. Trying to figure out where to keep "THE POINTER" was driving me nuts. I'd completely forgotten about intrusive pointers. Thanks. A *LOT*. Mark Storer <mstorer@cardiff To: "'Boost-Users(a)yahoogroups.com'" <Boost-Users(a)yahoogroups.com> .com> cc: Subject: RE: [Boost-Users] Re: Use a shared_ptr as a 'backpointer'? 09/30/2002 09:36 AM Please respond to Boost-Users I've got a comment and a question... 1) Couldn't the "backpointer" just be a constant reference? Backpointers, in my mind, don't have anything to do with ownership, so a shared_ptr woulnd't be appropriate. 2) shared_from_this? I can't find that one in any of the headers or docs I downloaded with 1.28. I did find shared_count in (oddly enough) shared_count.hpp, which looks like it would prevent two seperate shared_ptr's from causing a double-delete... pretty slick, but I couldn't find any docs on it. --Mark Storer Software Engineer Cardiff Software #include <disclaimer> typedef std::disclaimer<Cardiff> Discard; -----Original Message----- From: David Abrahams [mailto:yg-boost-users@m.gmane.org] Sent: Sunday, September 29, 2002 4:12 PM To: boost-users(a)yahoogroups.com Subject: [Boost-Users] Re: Use a shared_ptr as a 'backpointer'? <dick.bridges(a)tais.com> wrote in message news:OF699D7956.EED22A4F-ON88256C43.00722AD0@tais.net... > I run into situations where it's useful for an object to have a > 'backpointer' to the object instance that created it (example below). Is > there a pattern/idiom for doing this using shared_ptr? Or maybe I need to > re-think the idea of the backpointer in the first place? > class Bar; class Foo { public: Foo(shared_ptr<Bar>); private: shared_ptr<Bar> p; }; class Bar : boost::shared_count { void doit() { new Foo(boost::shared_from_this(this)); // <<< new Foo(this) obviously won't work. } }; HTH, Dave -- ----------------------------------------------------------- David Abrahams * Boost Consulting dave(a)boost-consulting.com * http://www.boost-consulting.com 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

RE: [Boost-Users] Re: Use a shared_ptr as a 'backpointer'?
by Mark Storer 30 Sep '02

30 Sep '02

I've got a comment and a question... 1) Couldn't the "backpointer" just be a constant reference? Backpointers, in my mind, don't have anything to do with ownership, so a shared_ptr woulnd't be appropriate. 2) shared_from_this? I can't find that one in any of the headers or docs I downloaded with 1.28. I did find shared_count in (oddly enough) shared_count.hpp, which looks like it would prevent two seperate shared_ptr's from causing a double-delete... pretty slick, but I couldn't find any docs on it. --Mark Storer Software Engineer Cardiff Software #include <disclaimer> typedef std::disclaimer<Cardiff> Discard; -----Original Message----- From: David Abrahams [mailto:yg-boost-users@m.gmane.org] Sent: Sunday, September 29, 2002 4:12 PM To: boost-users(a)yahoogroups.com Subject: [Boost-Users] Re: Use a shared_ptr as a 'backpointer'? <dick.bridges(a)tais.com> wrote in message news:OF699D7956.EED22A4F-ON88256C43.00722AD0@tais.net... > I run into situations where it's useful for an object to have a > 'backpointer' to the object instance that created it (example below). Is > there a pattern/idiom for doing this using shared_ptr? Or maybe I need to > re-think the idea of the backpointer in the first place? > class Bar; class Foo { public: Foo(shared_ptr<Bar>); private: shared_ptr<Bar> p; }; class Bar : boost::shared_count { void doit() { new Foo(boost::shared_from_this(this)); // <<< new Foo(this) obviously won't work. } }; HTH, Dave -- ----------------------------------------------------------- David Abrahams * Boost Consulting dave(a)boost-consulting.com * http://www.boost-consulting.com 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

RE: [Boost-Users] shared_ptr and counted_base
by Hickman, Greg 30 Sep '02

30 Sep '02

> I was looking through the latest boost documentation, trying Where is the documentation to which you refer? Is this a development doc or part of the ordinary user docs (if the latter, I don't know where to find it)? -Greg

2 1

Problem using 'shared_ptr<T>::operator impl-defined-type()'
by Hickman, Greg 30 Sep '02

30 Sep '02

With Borland 5.5.1, I'm not having any luck using a shared_ptr in a boolean context. The compiler complains that the 'if' condition below contains an "Illegal structure operation". Is this a known deficiency with my compiler? -Greg Hickman #include <boost/shared_ptr.hpp> int main(int argc, char* argv[]) { boost::shared_ptr<int> intp; if (intp) // error here ; return 0; }

5 10

RE: [Boost-Users] Returning a boost::scoped_lock by value
by Darryl Green 30 Sep '02

30 Sep '02

This issue has come up before on the development list (search for Boost.Threads Locking Delima). I'm not sure that it was resolved satisfactorily - move semantics for scoped_locks seemed to be the answer, but I think there were problems actually implementing that. The intent was that if move couldn't be implemented, then the mutex lock methods would have to be exposed. > -----Original Message----- > From: Raoul Gough [mailto:yg-boost-users@m.gmane.org] > Sent: Friday, 27 September 2002 9:44 PM > To: boost-users(a)yahoogroups.com > Subject: [Boost-Users] Returning a boost::scoped_lock by value > > > > boost::scoped_lock is non-copyable, so it is impossible to > return one from a function by value. However, I would like to > be able to do something like this: > > mt::holder<std::ostream> logStream; > > logStream() << "Hello world\n"; > > where the operator() returns a proxy object that contains a > scoped_lock and a pointer to the data (the ostream). What I > currently have is this: > > (mt::proxy<std::ostream> (logStream)) << "Hello world\n"; > > So the client code has to know what type the proxy object is > and construct one itself. If scoped_lock would provide access > to its associated mutex object, this wouldn't be a necessary, > because the proxy object could then have a copy-constructor, > and operator() in the holder template would look like this: > > proxy<T> operator() () > { > proxy<T> temp (*this); > return temp; > } > > The proxy copy constructor would unlock() the original lock > and construct the copy using the original's mutex (which is > the part that can't currently work, because there is no > public access to a lock's mutex). Assuming that the compiler > implements the NRVO, the unlock/lock sequence wouldn't > actually happen, and the client code just gets its proxy > object constructed for it. If the compiler doesn't implement > the NRVO, the code still works, but is less efficient. > > I guess exposing the raw mutex from scoped_lock would break > an important level of encapsulation. Still, it would be kind > of nice in this case. Maybe if proxy<T> were part of the > library and made a friend of scoped_lock? The only other > alternative I've come up with is to allocate the lock on the > heap and return an object containing a std::auto_ptr (which > makes usage easier, but less efficient). Any comments or suggestions? > > Regards, > Raoul Gough. > > > // Returnable proxy template > > template<typename T> > class proxy > { > boost::mutex::scoped_lock mLock; > T *mPtr; > > boost::mutex &release() > { > mLock.unlock(); > return mLock.mutex(); // <<<< doesn't work > } > > public: > > proxy (proxy &other) > : mLock (other.release()) > , mPtr (other.mPtr) > { > } > > // ... > }; > > > > > > > ------------------------ 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/ > >

2 1

RE: [Boost-Users] newbie Question about the macro BOOST_JOIN
by scleary＠jerviswebb.com 28 Sep '02

28 Sep '02

> From: zhangwusheng [mailto:yg-boost-users@m.gmane.org] > > Hi,all, > In file suffix.hpp there are lines: > > #define BOOST_JOIN( X, Y ) BOOST_DO_JOIN( X, Y ) > #define BOOST_DO_JOIN( X, Y ) BOOST_DO_JOIN2(X,Y) > #define BOOST_DO_JOIN2( X, Y ) X##Y > > > why donot we define such macro instead: > #define BOOST_JOIN( X, Y ) X##Y The idea behind BOOST_JOIN is to force the preprocessor to scan the macro parameters for replacement before merging the results with ##. So, it changes the meaning if one of the parameters is a macro: #define Z_BOOST_JOIN(X, Y) X##Y Z_BOOST_JOIN(id_, __LINE__) // Expands to "id___LINE__" BOOST_JOIN(id_, __LINE__) // Expands to "id_13" if __LINE__ is 13 However, I'm not sure why BOOST_DO_JOIN2 is used; it seems to me that only one layer of indirection should work. I assume it's to work around some compiler's non-conforming preprocessor. -Steve

3 2

Boost::UBLAS problems
by pecos_outwest 27 Sep '02

27 Sep '02

Hi Folks, I have been experimenting with the new BOOST linear algebra section boost/numeric/ublas I have been running the examples and benchmark programs bench1 and test1 provided with the distribution. Both programs compile and link perfectly they also run to completion; but they only run one time ! I cannot run a second time. It looks like the executable isn't even starting. Has anyone found this problem and found the solution I am using Microsoft VC6 compiler on a PC Thanks...Pecos

1 0

shared_ptr and counted_base
by Bobby Thomale 27 Sep '02

27 Sep '02

I was looking through the latest boost documentation, trying to find a way to get a shared_ptr from a this pointer of an object that already is contained by a shared_ptr somewhere, without knowing about that other shared_ptr out there. I noticed this intriguing comment in the latest docs: > [This constructor has been changed to a template in order to remember the > actual pointer type passed. The destructor will call delete with the same > pointer, complete with its original type, even when T does not have a virtual > destructor, or is void. > > In the current implementation, if p is convertible to counted_base *, > shared_ptr will use the embedded reference count supplied by counted_base. > This is an (experimental) attempt to provide a way for shared_ptr to be > constructed from a raw pointer such as this. A free function > shared_from_this(q) performs the conversion when q is convertible to > counted_base const *. > > The optional intrusive counting employed by the current implementation allows > shared_ptr to interoperate with intrusive_ptr, an experimental generic > intrusive-counted smart pointer. > > Another possible implementation is to use a global pointer-to-count map > instead of intrusive counting. shared_from_this would no longer be O(1), which > is a concern for some users, although I do not expect any performance > problems, since the operation is rare. Maintaining a global map is difficult; > it needs to be initialized before any shared_ptr instances are constructed, > and the initialization needs to be thread safe. In addition, under the Windows > dynamic library model, it is possible for several maps to exist. > > It is not yet clear which implementation should be used, or whether the > specification should allow both; nevertheless, the ability to make a > shared_ptr from this is considered essential by experienced smart pointer > users.] That looks like EXACTLY what I am looking for. But a couple things aren't clear to me: 1. Usage - The docs tell me its there but don't really say explicitly how to use it. It looks like, to use this, all I have to do is subclass boost::counted_base in my class that is being shared. Then I can just call shared_from_this to get another shared_ptr out of myself and it should just work. If that is all I have to do, then awesome! Is it really that easy or is there another step I am missing? 2. The word "experimental" above scares me. :-) If I use the counted_base subclass and shared_from_this function, is that code going to break later on if you guys change your minds? Anyway, I just want to reiterate what the comment at the end says, that the ability to make a shared_ptr from *this is essential to use these things. :-) -- Bobby --------------------------------------------------------------------- Bobby Thomale Senior Software Developer Inoveon Corporation http://www.inoveon.com/ ---------------------------------------------------------------------

2 2

Returning a boost::scoped_lock by value
by Raoul Gough 27 Sep '02

27 Sep '02

boost::scoped_lock is non-copyable, so it is impossible to return one from a function by value. However, I would like to be able to do something like this: mt::holder<std::ostream> logStream; logStream() << "Hello world\n"; where the operator() returns a proxy object that contains a scoped_lock and a pointer to the data (the ostream). What I currently have is this: (mt::proxy<std::ostream> (logStream)) << "Hello world\n"; So the client code has to know what type the proxy object is and construct one itself. If scoped_lock would provide access to its associated mutex object, this wouldn't be a necessary, because the proxy object could then have a copy-constructor, and operator() in the holder template would look like this: proxy<T> operator() () { proxy<T> temp (*this); return temp; } The proxy copy constructor would unlock() the original lock and construct the copy using the original's mutex (which is the part that can't currently work, because there is no public access to a lock's mutex). Assuming that the compiler implements the NRVO, the unlock/lock sequence wouldn't actually happen, and the client code just gets its proxy object constructed for it. If the compiler doesn't implement the NRVO, the code still works, but is less efficient. I guess exposing the raw mutex from scoped_lock would break an important level of encapsulation. Still, it would be kind of nice in this case. Maybe if proxy<T> were part of the library and made a friend of scoped_lock? The only other alternative I've come up with is to allocate the lock on the heap and return an object containing a std::auto_ptr (which makes usage easier, but less efficient). Any comments or suggestions? Regards, Raoul Gough. // Returnable proxy template template<typename T> class proxy { boost::mutex::scoped_lock mLock; T *mPtr; boost::mutex &release() { mLock.unlock(); return mLock.mutex(); // <<<< doesn't work } public: proxy (proxy &other) : mLock (other.release()) , mPtr (other.mPtr) { } // ... };

1 0

[BGL]: Iterating through internal properties of specific value
by Björn Lindberg 27 Sep '02

27 Sep '02

I have a BGL graph (adjacency_list). I use it to represent a tree, and so I have an internal vertex property of type enum representing the node type, ie ROOT, LEAF, INNER. Now, I would like a way to access only the leaves of a tree, preferrably through iterators. I can't think of any good way to accomplish this, except possibly by writing new iterator classes which would internally keep track of a property map and thus hide traversal of the non-leaf nodes from the code using the iterators. This seems like a cumbersome approach though, and I'm hoping there is a simpler way? Björn

3 5

RE: [Boost-Users] metaprogramming and MPL
by Aleksey Gurtovoy 27 Sep '02

27 Sep '02

Ihsan Ali Al Darhi wrote: > I am interested in metaprogramming and Boost mpl. But I can't find a > complete doc about the two subjects together. Where on the > net I can such info? It's not exactly a "complete doc" yet, but it's the most of that is publicly available to the moment - http://www.mywikinet.com/mpl/. Note that MPL sources in the CVS main trunk are more up-to-date than the .zip archive available for download from the library's page. HTH, Aleksey

1 0