December 2007 - Boost-users - lists.preview.boost.org

Re: [Boost-users] [multi_index] passing 'cons t TYPE' as 'this' argument discards qualifiers
by damny＠web.de 28 Dec '07

28 Dec '07

hi joaquín, now it works. thank you. > > are there possibilities to avoid the 'std::string' for the 'name' > > key? > > I'm not getting this question. Could you please elaborate? i would like to know if it is possible to use char[] or char* as an alphanumeric key value and not std::string. improper tests didn't work. daniel > ----- Mensaje original ----- > De: damny(a)web.de > Fecha: Martes, Diciembre 25, 2007 5:17 pm > Asunto: [Boost-users] [multi_index] passing 'cons t TYPE' as 'this' > argument discards qualifiers > Para: boost-users(a)lists.boost.org > > > merry xmas everyone, > > Merry Xmas :) > > [...] > > class cClass { > > private: > > int att1; > > public: > > std::string nme; // name > > int idf; // identifier > > cClass(); > > cClass(std::string nme, int idf); > > int get_attribute(); > > }; > > > [...] > > mMapByNme::iterator itrNme = locMsg.get<0>().find("name1"); > > mMapByIdf::iterator itrIdf = locMsg.get<1>().find(456); > > printf( "att1 = %d\n", itrNme->get_attribute() ); > [...] > > main.cpp: In function `int main()': > > main.cpp:33: error: passing 'const cClass' as 'this' argument of > > 'int cClass::get_attribute()' discards qualifiers > > Process terminated with status 1 (0 minutes, 20 seconds) > > > > any hints how to solve the problem? > > The problem here is that elements of a multi_index_container > are immutable, and the iterators ony grant const access > to them. In this regard, the behavior is the same as that > of a regular std::set. To solve your particular problem you can > const-qualify the offending member function: > > int get_attribute() const; > > > how to encapsulate this special multi_index into a class, that > > fits the above description of the needed 'map'. > > You should probably consider whether Boost.Bimap could > be a better approach, but I see you're already discussing > this issue with Matías. > > > are there possibilities to avoid the 'std::string' for the 'name' > > key? > > I'm not getting this question. Could you please elaborate? > > > where are the more easier examples for C++-templates/boost > > newbies like me :o) ? > > HTH, > > Joaquín M López Muñoz > Telefónica, Investiación y Desarrollo > _______________________________________________ > Boost-users mailing list > Boost-users(a)lists.boost.org > http://lists.boost.org/mailman/listinfo.cgi/boost-users > __________________________________________________________________________ Erweitern Sie FreeMail zu einem noch leistungsstärkeren E-Mail-Postfach! Mehr Infos unter http://produkte.web.de/club/?mc=021131

2 1

Re: [Boost-users] [multi_index] passing 'const TYPE' as 'this' argument discards qualifiers
by damny＠web.de 28 Dec '07

28 Dec '07

hello matias, i hope i can test boost.bimap today. therefor the last days were a little bit to stressful. although i thank you very much for your help. see you soon, daniel _____________________________________________________________________ Der WEB.DE SmartSurfer hilft bis zu 70% Ihrer Onlinekosten zu sparen! http://smartsurfer.web.de/?mc=100071&distributionid=000000000066

1 0

Unable to build python/example/tutorial from 1_34_1 in Windows XP with MSVC 8.0
by Andres Corrada-Emmanuel 28 Dec '07

28 Dec '07

Hello, I've been scouring the internet and boost.org documentation to do what I thought was a relatively simple task on Windows XP with MSVC 8.0: build and import into python the hello.cpp example in the 1_34_1 distribution. Warts: - The Jamroot file in the tutorial folder lacks the "import python ;" command. - No matter what I tried, bjam.exe, could not find the boost_python-vc80-mt-1_34_1.lib file. I finally gave up and copied it directly into the tutorial directory. - bjam compiled the hello.pyd but no "boost_python.dll" as the documentation for the Hello World example said it would. Has this changed? - Importing this hello.pyd fails even when I am in the same directory as the file fails. Do I need to have the boost_python.dll in the same directory? How do I create it with bjam? Is the tutorial directory broken? my bjam.exe? Any help would be appreciated. Particularly any pointers on how to debug my problems (bjam.exe --debug-configuration was not helpful in solving any of the boost_python lib import problems). -- Andres Corrada-Emmanuel Research Fellow Aerial Imaging and Remote Sensing Lab Computer Science Department University of Massachusetts at Amherst Blog: www.corrada.com/blog

1 0

Re: [Boost-users] [multi_index] passing 'const TYPE' as 'this' argument discards qualifiers
by damny＠web.de 27 Dec '07

27 Dec '07

ahh, very nice. please excuse my thick-wittedness. but ... what if, if i would like to add an additional, unique key to my structure; a last update time, maybe. name-identifer-time >> value; valu1:=map[nme1], valu1:=map[idf1], valu1:=map[time1] in this case bimap will not solve my problem and multiindex will be the best solution. is that right? and for interest, how to solve the original problem in this case? thnx daniel _____________________________________________________________________ Der WEB.DE SmartSurfer hilft bis zu 70% Ihrer Onlinekosten zu sparen! http://smartsurfer.web.de/?mc=100071&distributionid=000000000066

2 1

Re: [Boost-users] [review] Fast track formal review of Forward library begins today
by Tobias Schwinger 27 Dec '07

27 Dec '07

shunsuke wrote: > "A FunctionObject that can be used with boost::bind" is not good. :-) It's much better than using cryptic Concept names in tutorial-style documentation, IMO. We have a bunch of Function object Concepts defined in Fusion - and I have to say I'm waiting for the day I can remove them IOW result type computation crutches are not needed anymore... Regards, Tobias

2 1

time date newbie problem
by Michał Nowotka 27 Dec '07

27 Dec '07

Hi, I'm a date time library newbie and currenty working on simple web server implementation in C++. I want to generate string containing actual date in following format: Wed, 26 Dec 2007 17:33:47 GMT Date time documentation is so complicated and I can't find clear answer how to do that. Can you help me? -- Cheers, Michał Nowotka

4 6

Re: [Boost-users] [multi_index] passing 'cons t TYPE' as 'this' argument discards qualifiers
by "JOAQUIN LOPEZ MU?Z" 27 Dec '07

27 Dec '07

----- Mensaje original ----- De: damny(a)web.de Fecha: Martes, Diciembre 25, 2007 5:17 pm Asunto: [Boost-users] [multi_index] passing 'cons t TYPE' as 'this' argument discards qualifiers Para: boost-users(a)lists.boost.org > merry xmas everyone, Merry Xmas :) [...] > class cClass { > private: > int att1; > public: > std::string nme; // name > int idf; // identifier > cClass(); > cClass(std::string nme, int idf); > int get_attribute(); > }; > [...] > mMapByNme::iterator itrNme = locMsg.get<0>().find("name1"); > mMapByIdf::iterator itrIdf = locMsg.get<1>().find(456); > printf( "att1 = %d\n", itrNme->get_attribute() ); [...] > main.cpp: In function `int main()': > main.cpp:33: error: passing 'const cClass' as 'this' argument of > 'int cClass::get_attribute()' discards qualifiers > Process terminated with status 1 (0 minutes, 20 seconds) > > any hints how to solve the problem? The problem here is that elements of a multi_index_container are immutable, and the iterators ony grant const access to them. In this regard, the behavior is the same as that of a regular std::set. To solve your particular problem you can const-qualify the offending member function: int get_attribute() const; > how to encapsulate this special multi_index into a class, that > fits the above description of the needed 'map'. You should probably consider whether Boost.Bimap could be a better approach, but I see you're already discussing this issue with Matías. > are there possibilities to avoid the 'std::string' for the 'name' > key? I'm not getting this question. Could you please elaborate? > where are the more easier examples for C++-templates/boost > newbies like me :o) ? HTH, Joaquín M López Muñoz Telefónica, Investiación y Desarrollo

1 0

[lambda] problem with if_then_else_return
by Zeljko Vrba 27 Dec '07

27 Dec '07

I have a problem with the following statement: return std::accumulate( qrun_.begin(), qrun_.end(), svt, _1 < if_then_else_return( _2 != (acct*)0, bind(&acct::get_avt, _2), (unsigned)-1)); Basically, I want to find the minimal value of get_avt() over all non-null members of qrun_ which is declared as std::vector<acct*>. I get the following compiler error with SunCC 5.9: Error, badbinaryop: The operation "unsigned<const boost::lambda::detail::return_type_deduction_failure< boost::lambda::detail::return_type_2_ifthenelsereturn<1, 1, 1, 0, unsigned, const unsigned>>" is illegal. The get_avt method is declared as: unsigned get_avt() const Removing the const qualifier doesn't help. Replacing (unsigned)-1 with _1 generates the same error, but with "const unsigned" replaced by "unsigned&". Replacing (unsigned)-1 with bare 0 compiles without any error, but does not compute what I want. Any suggestions about what am I doing wrong?

1 0

Re: [Boost-users] [multi_index] passing 'const TYPE' as 'this' argument discards qualifiers
by damny＠web.de 26 Dec '07

26 Dec '07

hi mathias, thanks for your response. but i think there is a misapprehension. i want something like <http://www.boost.org/libs/multi_index/doc/tutorial/index.html>. "indexed by shape" should meet my "index by name" and "indexed by number" should meet my "index by identifier", so i can get the associated element value either by locMap[nme] or by locMap[idf], or something like that :o) there are 3 sets, a name key set, a identifier key set and an element value set; and there mapping. in a bimap are only 2 sets, a key set and an element set, which can swapped. thats not what i want. best regards, daniel _____________________________________________________________________ Der WEB.DE SmartSurfer hilft bis zu 70% Ihrer Onlinekosten zu sparen! http://smartsurfer.web.de/?mc=100071&distributionid=000000000066

2 1

[Serialization] BOOST_CLASS_EXPORT_GUID_1 macro issue (SVN Trunk)
by Richard Turrin 26 Dec '07

26 Dec '07

Hi, I am in the midst of migrating some projects into VS 2008, and have downloaded and built the boost library from the SVN Trunk (retrieved on 18 December, 2007 @ 4:09 pm) which seems to satisfy support for VS 2008. However, I was having a problem compiling some code that calls the macro BOOST_SHARED_POINTER_EXPORT (the code uses Boost serialization on some shared_ptr objects). I was running into the following error: MyOuterClass.h(11) : error C2371: '`anonymous-namespace'::boost_serialization_guid_initializer___LINE___1' : redefinition; different basic types MyOuterClass.h(10) : see declaration of '`anonymous-namespace'::boost_serialization_guid_initializer___LINE___1' The issue seems to be located in the BOOST_CLASS_EXPORT_GUID_1 macro, line 160 in serialization/export.hpp. The following macro is not expanding as desired: #define BOOST_CLASS_EXPORT_GUID_1(T, K) \ ... BOOST_PP_CAT(boost_serialization_guid_initializer_, __LINE__ ## _1) \ I believe the author of this line intended that __LINE__ would be replaced with a line number where the macro is implicitly called from, and THEN concatenated with the _1 token. Unfortunately, it seems that the two tokens are concatenated first creating __LINE___1 in the process. I have a trivial workaround: #define EXPAND(x) x #define BOOST_CLASS_EXPORT_GUID_1(T, K) \ ... BOOST_PP_CAT(boost_serialization_guid_initializer_, EXPAND(__LINE__) ## _1) \ This allows me to compile without the error (under VS 2008). Please advise if there is a better or more appropriate solution. Also, I wasn't sure whether _1 was a token literal or an mpl (Boost Metaprogramming library) placeholder. I realize that 1.35 is not yet ready for release. So, I'm not expecting everything to work correctly. So far, this is the only issue I've run into. Let me know if you need more information. Thank you for your time! Best Regards, Rich Turrin

3 4