- Boost - lists.boost.org

[random] Update for better TR1 compliance
by Jens Maurer 20 May '05

20 May '05

I've checked in changes to the Boost.Random library that add constructors and seeding for both template<class Generator> and unsigned long arguments, thereby moving Boost.Random one step further towards TR1 compliance. The random_test.cpp program now has basic tests for this feature. In order to guide overload resolution, I've used boost::enable_if<> and type traits to make sure that integral arguments don't try to call the "template<class Generator>" overloads. This could mean decreased compiler support, sorry. It works for me with gcc 3.4.3. Jens Maurer

5 7

Interest in C function pointer utility?
by Rene Rivera 20 May '05

20 May '05

In rewriting a plugin interface I use, which for language portability uses C function pointers as the entry points, I came up with a "make_c_function" utility. I use it to translate C++ functions, any of objects, binds, function, members, et. to _real_ C function pointers. For example: class P { public: int foo(int); }; P p; extern "C" { void * get_plugin_call(int n) { if (1 == n) { return make_c_function<P,int(int)>( boost::bind(&p,p::foo,_1)); } else { return 0; } } } [example much abbreviated from real life use] -- -- Grafik - Don't Assume Anything -- Redshift Software, Inc. - http://redshift-software.com -- rrivera/acm.org - grafik/redshift-software.com -- 102708583/icq - grafikrobot/aim - Grafik/jabber.org

3 4

[iostreams] patch for examples/dictionary_filter.hpp for bcc32
by Russell Hind 20 May '05

20 May '05

Please find attached a patch for libs/iostreams/example/dictionary_filter.hpp to help get the tests to build under bcc32 5.6.x and STLPort Thanks Russell Index: libs/iostreams/example/dictionary_filter.hpp =================================================================== RCS file: /cvsroot/boost/boost/libs/iostreams/example/dictionary_filter.hpp,v retrieving revision 1.3 diff -u -r1.3 dictionary_filter.hpp --- libs/iostreams/example/dictionary_filter.hpp 20 May 2005 03:59:33 -0000 1.3 +++ libs/iostreams/example/dictionary_filter.hpp 20 May 2005 12:22:17 -0000 @@ -42,7 +42,11 @@ using namespace std; while (true) { int c = std::cin.get(); +#if defined(__BORLANDC__) && (__BORLANDC__ >= 0x560) && (__BORLANDC__ <= 0x564) && !defined(_USE_OLD_RW_STL) + if (c == EOF || !std::isalpha(c)) { +#else if (c == EOF || !std::isalpha(c, dictionary_.getloc())) { +#endif dictionary_.replace(current_word_); cout.write( current_word_.data(), static_cast<streamsize>(current_word_.size()) ); @@ -84,7 +88,11 @@ if ((c = iostreams::get(src)) == WOULD_BLOCK) return WOULD_BLOCK; +#if defined(__BORLANDC__) && (__BORLANDC__ >= 0x560) && (__BORLANDC__ <= 0x564) && !defined(_USE_OLD_RW_STL) + if (c == EOF || !std::isalpha(c)) { +#else if (c == EOF || !std::isalpha(c, dictionary_.getloc())) { +#endif dictionary_.replace(current_word_); off_ = 0; if (c == EOF) @@ -127,7 +135,11 @@ if (off_ != std::string::npos && !write_current_word(dest)) return false; +#if defined(__BORLANDC__) && (__BORLANDC__ >= 0x560) && (__BORLANDC__ <= 0x564) && !defined(_USE_OLD_RW_STL) + if (!std::isalpha(c)) { +#else if (!std::isalpha(c, dictionary_.getloc())) { +#endif dictionary_.replace(current_word_); off_ = 0; } @@ -190,8 +202,13 @@ if (it == map_.end()) return false; string& value = it->second; +#if defined(__BORLANDC__) && (__BORLANDC__ >= 0x560) && (__BORLANDC__ <= 0x564) && !defined(_USE_OLD_RW_STL) + if (!value.empty() && !key.empty() && std::isupper(key[0])) + value[0] = std::toupper(value[0]); +#else if (!value.empty() && !key.empty() && std::isupper(key[0], loc_)) value[0] = std::toupper(value[0], loc_); +#endif key = value; return true; } @@ -201,7 +218,11 @@ inline void dictionary::tolower(std::string& str) { for (std::string::size_type z = 0, len = str.size(); z < len; ++z) +#if defined(__BORLANDC__) && (__BORLANDC__ >= 0x560) && (__BORLANDC__ <= 0x564) && !defined(_USE_OLD_RW_STL) + str[z] = std::tolower(str[z]); +#else str[z] = std::tolower(str[z], loc_); +#endif } } } } // End namespaces example, iostreams, boost.

2 1

RE: Non-intrusive serialization with VC2003 (7.1) and building from CVS
by Marcin Tustin 20 May '05

20 May '05

It looks a lot like the problem I was having was caused by constness. To be clear, I was doing: template<class Archive> void serialize(Archive & ar, const Jrs::SmartPtr<D2::Message>& g, const unsigned int version) { //... } Now I'm doing: template<class Archive> void serialize(Archive & ar, Jrs::SmartPtr<D2::Message>& g, const unsigned int version) { // ... } The new version does compile, but as I am having problems building boost (from CVS), it doesn't link because the libraries don't exist. It compiles with the CVS version, but still does not compile under 1.32. As to building, I am trying to use the version of bjam I built to build 1.32, running it from the place it was built to: mt@penfold /cygdrive/c/new_boost/boost $ /cygdrive/c/boost_1_32_0/tools/build/jam_src/bin.ntx86/bjam.exe But this cannot create files or directories. Any suggestions?

1 0

[serialization][ptr_container] serialization of ptr_containers
by Felipe Magno de Almeida 20 May '05

20 May '05

Hi, I was trying to serialiaze ptr_containers and I couldnt find a way, so I copied some headers and modified them to serialize ptr_list(the only one I need to serialize in the moment), and it seems to be working. So If someone is interested, I could send it to someone. Im not an expert in the serialize library, so I dont even know if it is really right... But it worked to me(the objects are all serialized as pointers, so if there's two objects then it just serialize one...). I think that extending it to serialize others sequence containers would be very easy, but I didnt touch in the associative containers... The files that I created: ptr_collections_load_imp.hpp // from collections_load_imp.hpp ptr_collections_save_imp.hpp // from collections_save_imp.hpp ptr_list.hpp // from list.hpp Thank you. -- Felipe Magno de Almeida UIN: 2113442 email: felipe.m.almeida+spam at gmail com felipe.almeida+spam at ic unicamp br, felipe.m.almeida at gmail com, felipe+spam at synergy com (Remove the +spam from the email) I am a C, modern C++, MFC, ODBC, Windows Services, MAPI developer from synergy, and Computer Science student from State University of Campinas(UNICAMP). To know more about: Unicamp: http://www.ic.unicamp.br Synergy: http://www.synergy.com.br current work: http://www.mintercept.com "There is no dark side of the moon really. Matter of fact it's all dark."

2 1

Re: [boost] Re: Re: char[] support in Boost.Range
by brangdon＠cix.compulink.co.uk 20 May '05

20 May '05

In-Reply-To: <d6dhiu$hai$1(a)sea.gmane.org> technews(a)kangaroologic.com (Jonathan Turkanis) wrote (abridged): > > Are we going to generalise it to allow any (compile-time constant) > > terminator? > > Why does the terminator have to be a compile-time constant? I was hesitant about suggesting a generalisation which added run-time overhead over only supporting '\0'. Also, the need is not very common in my experience. -- Dave Harris, Nottingham, UK.

3 4

RE: [boost] Logging Library -- Formal Review Request
by Spencer.Collyer＠barclayscapital.com 19 May '05

19 May '05

> >Thanks!!! > > > >Also, an application should have multiple levels at the same > time (i.e., > >an app may want to have a different level of debug logging per > >component, library, class, whatever). > >_______________________________________________ > > > > > This is an interesting request, and I assume it'll change the way I > implement multiple levels. > > Is this a desireable feature? What do you all think? > > In other words, if I understand correctly, you want to be able to set > the level of each log (independently of other logs). > In this case, instead of enabling/disabling logs, I can > simply set the > level of the logs -- and have a level called DISABLED (which is > equivalent to disabling one log) Just de-lurking for a moment here, but I'd like to add my vote for a form of this. At my work we have a logging library that has concepts of log topics and log levels. These are orthogonal - any particular log message is associated with one topic and has one level, but the messages for a given topic can have any level specified. The topics are arranged in a tree-like hierarchy, and if a minimum logging level is specified for a topic it also affects any subtopics of that topic which have not had a logging level specifically set. This allows us to easily tune the amount of logging information produced so that only the items we are specifically working on get logged at a high debugging level, without interference from components we are not interested in. This makes tracking problems considerably easier, as only the relevant information is output. Just my twopen'orth. S> ------------------------------------------------------------------------ For more information about Barclays Capital, please visit our web site at http://www.barcap.com. Internet communications are not secure and therefore the Barclays Group does not accept legal responsibility for the contents of this message. Although the Barclays Group operates anti-virus programmes, it does not accept responsibility for any damage whatsoever that is caused by viruses being passed. Any views or opinions presented are solely those of the author and do not necessarily represent those of the Barclays Group. Replies to this email may be monitored by the Barclays Group for operational or business reasons. ------------------------------------------------------------------------

2 1

RE: [boost] Re: Re: Non-intrusive serialization with VC2003 (7.1)
by Marcin Tustin 19 May '05

19 May '05

> -----Original Message----- > From: Robert Ramey [mailto:ramey@rrsd.com] > Sent: 19 May 2005 16:34 > To: boost(a)lists.boost.org > Subject: [boost] Re: Re: Non-intrusive serialization with VC2003 (7.1) > > Marcin Tustin wrote: > > c:\new_boost\boost\boost\archive\detail\oserializer.hpp(152) : > > > The choice is made here. Yes, and for the moment my understanding of function selection in C++ is not enough to understand why the choice is made. > >>>> namespace boost { > >>>> namespace serialization { > >>>> template<class Archive > > >>>> static void serialize(Archive & ar, const > >>> D2::Message& g, const > >>>> unsigned int version) > >>>> { > >>>> //... > >>>> } > >>>> > >>>> } }; > >>>> > > Note that your serialize template above uses datatype > D2::Message while one is trying to serialize D2::MessagePtr. I've changed the function to use D2::MessagePtr already. The move to D2::MessagePtr (Which is a typedef for Jrs::SmartPtr<D2::Message>) is for other reasons. In any case, the problem still obtains. The new code is: template<class Archive> void serialize(Archive & ar, const Jrs::SmartPtr<D2::Message>& g, const unsigned int version) { //ar & g.stdString(); ar & g.px; // contained pointer ar & g.pn; // reference counter } And yes, I am aware that there is a whole load of extra complexity in correctly serialising smart pointers, and I will deal with that once I can get any non-intrusive serialisation to work. Thanks again for your help and attention.

1 0

RE: [boost] Re: Non-intrusive serialization with VC2003 (7.1)
by Marcin Tustin 19 May '05

19 May '05

> -----Original Message----- > From: Jason Earl [mailto:Jason@hybd.net] > Sent: 19 May 2005 16:20 > To: boost(a)lists.boost.org > Subject: Re: [boost] Re: Non-intrusive serialization with VC2003 (7.1) > > I'm probably completely wrong, but but am I right in thinking > the problem is due Jrs::SmartPtr<T> and not the fact that > VS.NET or the serialization library is not find the right > serialize method. If so I'm surprised no one else has pointed > this out before! The big deal is that there is a non-intrusive serialize function (i.e. an external function) rather than a method on SmartPtr. The problem is that this function is not being detected, so the template library falls through to trying to use a serialize method, which does not exist. Because I'm building on another library, there is no possibility of going over to any other kind of smart pointer.

1 0

RE: [boost] Logging Library -- Formal Review Request
by Spencer.Collyer＠barclayscapital.com 19 May '05

19 May '05

> > At my work we have a logging library that has concepts of > log topics and > > log levels. These are orthogonal - any particular log message is > > associated with one topic and has one level, but the messages for a > > given topic can have any level specified. The topics are > arranged in a > > tree-like hierarchy, and if a minimum logging level is > specified for a > > topic it also affects any subtopics of that topic which > have not had a > > logging level specifically set. > > I would assume what you call "topic" here, is what I've done by using > loggers/hierarchy. Have you read the docs? John, what we call topics are indeed analogous to your loggers. I have indeed read the docs and found a lot to like in the proposed library, but it does need the ability to specify different levels for different messages on a particular log (IMHO, of course :) ) Spencer ------------------------------------------------------------------------ For more information about Barclays Capital, please visit our web site at http://www.barcap.com. Internet communications are not secure and therefore the Barclays Group does not accept legal responsibility for the contents of this message. Although the Barclays Group operates anti-virus programmes, it does not accept responsibility for any damage whatsoever that is caused by viruses being passed. Any views or opinions presented are solely those of the author and do not necessarily represent those of the Barclays Group. Replies to this email may be monitored by the Barclays Group for operational or business reasons. ------------------------------------------------------------------------

1 0