January 2008 - Boost-users - lists.preview.boost.org

[review] Review of Flyweight library started January 21 and still running!
by Ion Gaztañaga 26 Jan '08

26 Jan '08

Hi all, The formal review of Flyweight library started January 21 and will end January 30, we have a few nice reviews, but I'm sure that those who haven't reviewed it are also interested in the library ;-). The library solves quite a common programming pattern, is small and very well documented. So don't be afraid to try it! Information for reviewers: *Description:* Flyweights are small-sized handle classes granting constant access to shared common data, thus allowing for the management of large amounts of entities within reasonable memory limits. Boost.Flyweight makes it easy to use this common programming idiom by providing the class template flyweight<T>, which acts as a drop-in replacement for const T. *Online docs:* http://tinyurl.com/2sstyr http://svn.boost.org/svn/boost/sandbox/flyweight/libs/flyweight/index.html *Download:* http://tinyurl.com/hrdm6 http://www.boost-consulting.com/vault/index.php?&direction=0&order=&directo… *Notes:* 1) We've seen some suggestions in the mailing list for Flyweight. Joaquín has nicely explained a couple of issues that we'd like to address/discuss in the review: http://tinyurl.com/33ghtf http://svn.boost.org/svn/boost/sandbox/flyweight/libs/flyweight/doc/review_… 2) Flyweight needs Boost 1.35 elements because the library depends on libraries like Interprocess for some features/tests. Since SVN snapshot tarballs seem to be missing these days, those who want to try flyweight can download a working SVN-HEAD snapshot here: http://igaztanaga.drivehq.com/boost_trunk.tar.bz2 3) Serialization tests won't work. This feature is expected to work when some new features (discussed in the mailing list between Joaquín and Robert Ramey) are added in Boost.Serialization. Those are expected for Boost 1.36. What to include in Review Comments ================================== Your comments may be brief or lengthy, but basically the Review Manager needs your evaluation of the library. If you identify problems along the way, please note if they are minor, serious, or showstoppers. Here are some questions you might want to answer in your review: * What is your evaluation of the design? * What is your evaluation of the implementation? * What is your evaluation of the documentation? * What is your evaluation of the potential usefulness of the library? * Did you try to use the library? With what compiler? Did you have any problems? * How much effort did you put into your evaluation? A glance? A quick reading? In-depth study? * Are you knowledgeable about the problem domain? And finally, every review should answer this question: * Do you think the library should be accepted as a Boost library? Be sure to say this explicitly so that your other comments don't obscure your overall opinion. Ion Gaztañaga - Review Manager -

1 0

[serialization] hash_map declaration problem
by Jan Boehme 26 Jan '08

26 Jan '08

Hi, I encountered problems while trying to serialize hash_maps which use pool allocators ( gcc: no member 'serialize' ). I had a look into "boost/serialization/hash_map.hpp" and realized a problem in the declaration of hash_map in here: BOOST_STD_EXTENSION_NAMESPACE::hash_map<Key, HashFcn, EqualKey, Allocator> should be BOOST_STD_EXTENSION_NAMESPACE::hash_map<Key, Value, HashFcn, EqualKey, Allocator> It seems that nobody used the hash_map with more than 4 template parameters together with boost::serialization before. It would be nice, if the following changes for "boost/serialization/hash_map.hpp" would be adopted into the repos. Cheers, Jan. --- /boost-trunk/boost/serialization/hash_map.hpp 2007-08-01 09:03: 26.000000000 +0200 +++ /boost/serialization/hash_map.hpp 2007-12-17 11:36:27.000000000 +0100 @@ -33,3 +33,4 @@ class Archive, - class Key, + class Key, + class Value, class HashFcn, @@ -41,3 +42,3 @@ const BOOST_STD_EXTENSION_NAMESPACE::hash_map< - Key, HashFcn, EqualKey, Allocator + Key, Value, HashFcn, EqualKey, Allocator > &t, @@ -48,3 +49,3 @@ BOOST_STD_EXTENSION_NAMESPACE::hash_map< - Key, HashFcn, EqualKey, Allocator + Key, Value, HashFcn, EqualKey, Allocator > @@ -55,3 +56,4 @@ class Archive, - class Key, + class Key, + class Value, class HashFcn, @@ -63,3 +65,3 @@ BOOST_STD_EXTENSION_NAMESPACE::hash_map< - Key, HashFcn, EqualKey, Allocator + Key, Value, HashFcn, EqualKey, Allocator > &t, @@ -70,3 +72,3 @@ BOOST_STD_EXTENSION_NAMESPACE::hash_map< - Key, HashFcn, EqualKey, Allocator + Key, Value, HashFcn, EqualKey, Allocator >, @@ -75,3 +77,3 @@ BOOST_STD_EXTENSION_NAMESPACE::hash_map< - Key, HashFcn, EqualKey, Allocator + Key, Value, HashFcn, EqualKey, Allocator > @@ -85,3 +87,4 @@ class Archive, - class Key, + class Key, + class Value, class HashFcn, @@ -93,3 +96,3 @@ BOOST_STD_EXTENSION_NAMESPACE::hash_map< - Key, HashFcn, EqualKey, Allocator + Key, Value, HashFcn, EqualKey, Allocator > &t,

3 4

Multiple templates singletons
by Hicham Mouline 25 Jan '08

25 Jan '08

Hi, I have a template class with 6 template arguments as : template <typename contractType, typename treeType, int n, enum1 rtype, enum2 qtype, enum3 vtype> class TreeImpl {}; I need to write partial specializations for the cases n=2, n=3, for 2 cases of rtype, 2 cases of qtype and 3 cases of vtype, therefore I will write 2*2*2*3 = 24 partial specializations. Any fully specialized template instance should be a singleton. I would like to find a way to store any instanciated full specialization in some common list, so that in other parts of the code, i can deal with the various singletons without having to write actual arguments n,rtype,qtype and vtype. I thought of having all these templates inherit from a template base class template <typename contractType, typename treeType> class TreeImplBase; Each specialization of these holds a static std::map< boost::tuple<int,enum1,enum2,enum3>, const TreeImplBase* >; which contains the list of the derived template classes. Any comments are appreciated, Perhaps this is a known design pattern, perhaps there is an easier way of doing this, perhaps there is a boost library which helps with this, regards,

3 6

local_date_time as class member
by Fabien CARMAGNAC 25 Jan '08

25 Jan '08

Dear all, I am using a boost::date_time::local_date_time and I have a problem when a local_date_time is declared as class member. Case 1 class toto { protected: boost::local_time::local_date_time tt ; public: toto(boost::local_time::local_date_time t) : tt(t) {} ~toto(){} }; => compil ok =========== Case 2 (using '=' operator) class toto { protected: boost::local_time::local_date_time tt ; public: toto(boost::local_time::local_date_time t) { tt=t; } ~toto(){} }; => does not compiles ! Errors: make all Building file: ../src/toto.cpp Invoking: GCC C++ Compiler g++ -I"/home/fabien/dev/eclipse_workspace" -O0 -g3 -Wall -c -fmessage-length=0 -MMD -MP -MF"src/toto.d" -MT"src/toto.d" -o"src/toto.o" "../src/toto.cpp" ../src/toto.cpp: In constructor ‘toto::toto(boost::local_time::local_date_time)’: ../src/toto.cpp:16: error: no matching function for call to ‘boost::local_time::local_date_time_base<boost::posix_time::ptime, boost::date_time::time_zone_base<boost::posix_time::ptime, char> >::local_date_time_base()’ /usr/include/boost/date_time/local_time/local_date_time.hpp:218: note: candidates are: boost::local_time::local_date_time_base<utc_time_, tz_type>::local_date_time_base(boost::date_time::special_values, boost::shared_ptr<U>) [with utc_time_ = boost::posix_time::ptime, tz_type = boost::date_time::time_zone_base<boost::posix_time::ptime, char>] /usr/include/boost/date_time/local_time/local_date_time.hpp:211: note: boost::local_time::local_date_time_base<utc_time_, tz_type>::local_date_time_base(const boost::local_time::local_date_time_base<utc_time_, tz_type>&) [with utc_time_ = boost::posix_time::ptime, tz_type = boost::date_time::time_zone_base<boost::posix_time::ptime, char>] /usr/include/boost/date_time/local_time/local_date_time.hpp:141: note: boost::local_time::local_date_time_base<utc_time_, tz_type>::local_date_time_base(typename utc_time_::date_type, typename utc_time_::time_duration_type, boost::shared_ptr<U>, boost::local_time::local_date_time_base<utc_time_, tz_type>::DST_CALC_OPTIONS) [with utc_time_ = boost::posix_time::ptime, tz_type = boost::date_time::time_zone_base<boost::posix_time::ptime, char>] /usr/include/boost/date_time/local_time/local_date_time.hpp:100: note: boost::local_time::local_date_time_base<utc_time_, tz_type>::local_date_time_base(typename utc_time_::date_type, typename utc_time_::time_duration_type, boost::shared_ptr<U>, bool) [with utc_time_ = boost::posix_time::ptime, tz_type = boost::date_time::time_zone_base<boost::posix_time::ptime, char>] /usr/include/boost/date_time/local_time/local_date_time.hpp:79: note: boost::local_time::local_date_time_base<utc_time_, tz_type>::local_date_time_base(utc_time_, boost::shared_ptr<U>) [with utc_time_ = boost::posix_time::ptime, tz_type = boost::date_time::time_zone_base<boost::posix_time::ptime, char>] make: *** [src/toto.o] Error 1 I am using: Linux 2.6.23.8-34.fc7 boost-1.33.1-15.fc7 boost-devel-1.33.1-15.fc7 g++ (GCC) 4.1.2 20070925 (Red Hat 4.1.2-27) Thanks for your help ! Regards, Fabien CARMAGNAC _____________________________________________________________________________ Ne gardez plus qu'une seule adresse mail ! Copiez vos mails vers Yahoo! Mail http://mail.yahoo.fr

2 1

Where to download boost 1.33.0-1?
by Bo Xie 25 Jan '08

25 Jan '08

http://sourceforge.net/project/showfiles.php?group_id=7586&package_id=8041 only provides boost 1.33.0 download, not 1.33.0-1. Thanks!

2 1

BGL: Problem accessing property_map members
by James Jackson 25 Jan '08

25 Jan '08

Hi, I am trying to use the BGL to model a simple directed graph with internal edge properties. Following the example given in the documentation[1], I have done the following (with question-irrelevant padding removed): ------------------------------- #include <boost/graph/adjacency_list.hpp> #include <boost/property_map.hpp> typedef boost::property<boost::edge_weight_t, double> EdgeProperties; typedef boost::adjacency_list<boost::vecS, boost::vecS, boost::directedS, boost::no_property, EdgeProperties> NetworkGraph; typedef std::pair<int, int> Edge; NetworkGraph g(5); boost::add_edge(0, 1, g); boost::add_edge(0, 2, g); boost::add_edge(1, 3, g); boost::add_edge(2, 3, g); boost::add_edge(3, 4, g); boost::property_map<NetworkGraph, boost::edge_weight_t>::type weights = boost::get(boost::edge_weight_t(), g); weights[0] = 1.0; weights[1] = 1.0; weights[2] = 1.0; weights[3] = 1.0; weights[4] = 1.0; ------------------------------- Which all looks fairly kosher. However, attempting to compile this results in a error on each of the weights assignment lines: "error: no match for 'operator[]' in 'weights[0]' Using the syntax boost::put(weights, 0, 1.0) just moves this error inside a parameter_set template definition. Can anybody shed any light on this? For the record this is on MacOS X 10.5.1, with GCC 4.0.1 (i686-apple-darwin9-gcc-4.0.1). Regards, James Jackson. [1] interior_property_map.cpp

2 2

BoostCon 2008 Now Open For Registration
by David Abrahams 25 Jan '08

25 Jan '08

Hi All, The 2nd annual Boost Conference has been scheduled for May 4-9 2008. Please see http://www.boostcon.com/registration for registration information. If you want to stay at the Aspen Meadows (highly recommended as it is within walking distance of the conference venue and was the site of great after-hours discussions in '07), it's especially important this year that you book your rooms early. Since, unlike last year, Aspen's airport will be open this May, demand for rooms has been heavy. In fact, the following week, which we tried to get, was snapped up by another group before we even had a chance to catch our breath, just weeks after BoostCon'07 ended. We'll be working hard on getting the program together for 2008 over the coming weeks. We hope to include many more Boost authors and contributors, and build in several more intensive library-specific sessions, to make 2008 better than ever. Please send any questions to the BoostCon planning committee at: organizers(a)boostcon.com. Those of you who didn't make it to BoostCon'07 may want to have a look at the photo gallery at http://boostcon.com/community/photos, peruse the 2007 program at http://boostcon.com/program, and check out the two trip reports mentioned here: http://boostcon.com/news/2007/Jul/16/yet-another-trip-report. See you in Aspen! -- The BoostCon Planning Committee

1 0

Problems using Boost.Threads
by anthony kelly 25 Jan '08

25 Jan '08

Hello Boost Users, I am trying to use Boost.Thread in my application. I am compiling on Windows for Windows using Cygwin GCC with the "no-cygwin" option. Boost.Build is far too complicated for my application, so I am using a simple Makefile. I don't mind whether I use Windows threads or POSIX threads, but I have tried both and I can't get either to work. Re: Windows threads, I have static linked to "boost_thread-gcc34-mt-1_34_1" (no plan to use unsupported tss), and I defined BOOST_THREAD_USE_LIB and BOOST_HAS_WINTHREADS, however I get a lot of error messages about pthread functions not found. Have you any idea why I get error messages about using pthread functions when BOOST_HAS_WINTHREADS is defined? Is there some other macro(s) which should be defined? Re: POSIX threads, I have static linked to "boost_thread-gcc34-mt-1_34_1" and "pthreadGC2" (pthreads-win32), and I defined BOOST_THREAD_USE_LIB and BOOST_HAS_PTHREADS, however I get an error message that nanosleep is not found. I read that nanosleep is part of libc, so I added "c" to my list of static libraries, which finds nanosleep, but then I receieve other errors which looks as though I am linked to multiple C runtimes. What is the correct approach to link nanosleep in this configuration? If you have any suggestion, please let me know! Thanks in advance for your time. Kind regards, Anthony P.S. Is there a resource which can tell me which are the prerequisite libraries for boost libraries (e.g. pthreads or winsock)? So far, I have had to guess appropriate libraries based on internet searches for the function names appearing in "undefined reference" errors, but this seems very clumsy!

1 0

[ublast] sparse matrix - resize with preserve
by Olivier Jouny 24 Jan '08

24 Jan '08

I would like to use the resize method of mapped_matrix with preserve = true but it does not work. By looking at the implementation, a comment says not implemented :-( Can you confirm ? If it's the case, does somebody has a work around ? Regards, Olivier. _________________________________________________________________ Comme 9 millions de français, créez un compte Hotmail ! http://www.windowslive.fr/hotmail/default.asp

1 0

Storing types in a non-template class?
by Robert Dailey 24 Jan '08

24 Jan '08

Hi, I have a singleton class called InputRegistry. Before the class is used, the user must call: InputRegistry::Initialize(); I would like this function to take a parameter that defines the type of a critical section object, which is unknown at compile time (template). I need to store this type in my InputRegistry class so that the unknown critical section type can be constructed later. How can I do this without making InputRegistry a template class? If I make InputRegistry a template, that basically means I need to redesign the class, as the singleton pattern would not apply if it were a template class. I'm hoping boost can help here. Thanks in advance.

4 3

[Singleton] pre-review question
by Ovanes Markarian 24 Jan '08

24 Jan '08

Hello, I am currently evaluating boost.singleton for our application. I hope it is not too late to write a review. As I understood from the docs, Singleton instance is immediately (when all global and static variables are initilized) created. In our context we need singleton to be created on first request (as it is described in GoF). The problem is following: Our application must parse config file which is accessible after entering main and store the options/parameters parsed in a singleton class. If I initialize singleton immediately (before entering main) I can not parse command line options and have to trick around with some additional functions which do it for me (but a constructor would be a much better choice, when I try for the first time to access the singleton). Is that possible with the current implementation? I know I could overload operator-> and check that options are parsed, but I think this is a not so obvious scenario for everyone. With Kind Regards, Ovanes

2 5

cygwin buit boost libraries do not work
by Steven Woody 24 Jan '08

24 Jan '08

Hi folks, I have buit boost in cygwin followed instruction in the boost getting start documentation. it seemed successful and many libraries ( including dynamic and static ones ) were put into /usr/local/lib. I then try to compile a very simple example code listed as blow, which are a copy on the boost documentaion. #include <boost/regex.hpp> #include <iostream> #include <string> int main() { std::string line; boost::regex pat( "^Subject: (Re: |Aw: )*(.*)" ); while (std::cin) { std::getline(std::cin, line); boost::smatch matches; if (boost::regex_match(line, matches, pat)) std::cout << matches[2] << std::endl; } } What I did is, g++ -I/usr/local/include RegexSample.cpp -o RegexSample -L/usr/local/lib -lboost_regex but the resulting RegexSample.exe can not run properly -- it simply exist with an error code ( 53 ) and do nothing for me. My question is: is there anything different for Boost on Cygwin? The procedure I used to buid Boost on Cygwin is provided as below. Please check it can tell me if there is anything wrong. Thanks! --------- 1, build the bjam using './build.sh gcc' under the tools/jam/src directory and then copy bjam.exe to a place searchable; 2. 'bjam --builddir=~/tmp --toolset=gcc --layout=system --without-python install' ----------- -- woody then sun rose thinly from the sea and the old man could see the other boats, low on the water and well in toward the shore, spread out across the current.

3 6

Re: [Boost-users] Multiple templates singletons
by Hicham Mouline 24 Jan '08

24 Jan '08

I omited to show the definition of those templates TreeImplBase.They would have the map as a private member,as protected members, functions to insert instances in the map, called by the ctors of derived templates. I will look at boost::fusion thanks, ----- Original Message ----- From: "Ovanes Markarian" To: boost-users(a)lists.boost.org Subject: Re: [Boost-users] Multiple templates singletons Date: Thu, 24 Jan 2008 00:02:01 +0100 Hello! I have a question to this construct: template <typename contractType, typename treeType> class TreeImplBase; Each specialization of these holds a static std::map< boost::tuple<int,enum1,enum2,enum3>, const TreeImplBase* >; which contains the list of the derived template classes. How can you instantiate a pointer to the class temlate? Class template is an incomplete type. You will not be able to specify a pointer to it. I think you should use boost::fusion inplace. This lib, offers a compile time map (if you map the same time to an instance of it, use a fusion set), which can map a type to an instance. This is what you need. http://spirit.sourceforge.net/dl_more/fusion_v2/libs/fusion/doc/html/fusion… http://spirit.sourceforge.net/dl_more/fusion_v2/libs/fusion/doc/html/fusion… Any comments are appreciated, Perhaps this is a known design pattern, perhaps there is an easier way of doing this, perhaps there is a boost library which helps with this, Good Luck, Ovanes _______________________________________________ Boost-users mailing list Boost-users(a)lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users

1 0

iostreams and serial ports help
by Rob Desbois 24 Jan '08

24 Jan '08

Hi, I'm a newcomer to boost and having trouble determining if what I want to achieve is possible. I want some sort of boost::iostreams object wrapping a serial port so that it conforms to the usual interfaces - this would be extremely useful not only for its use, but also to enable injection of a stringstream instance as a mock object when testing. It seems that the likely candidate for this is file_descriptor but I'm not entirely sure. First off, file_descriptor refines SeekableDevice, which a serial port isn't - does this mean that I can't do it? After trying to compile I realised that a SeekableDevice provides read, write and seek functions, nothing else. Am I seriously misunderstanding the point of boost::iostreams? If I am then can anybody point me in a better direction? My ultimate requirement is a serial port access class that looks just like a C++/boost iostream. I cannot use C++ iostreams directly as I require access to file-descriptor which is not available. Anyone able to help a lost person please? TIA, --rob

4 8

[function] using with CRTP
by Russell L. Carter 23 Jan '08

23 Jan '08

Hi, This is probably a novice template question. I am curious what I am doing wrong here, where I'm trying to get a function object holding a base member of a CRTP class. gcc 4.2.3 complains about: type/value mismatch at argument 1 in template parameter list for 'template<class Derived> template<template<template<class> class MHBase> template<class Derived> template<class> class MHBase> struct MHBase< <template-parameter-1-1> >::fooFO Any help appreciated! Thanks, Russell #include <boost/function.hpp> #include <iostream> class MHBAccessor { template<class> friend class MHBase; template<class Derived> static int foo(Derived& derived, int i) { return derived.foo_(i); } }; typedef boost::function<int (int)> FooFunc; template <class Derived> struct MHBase { Derived& derived() { return static_cast<Derived&>(*this); } int foo(int i) { return MHBAccessor::foo(this->derived(), i); } template <template <class> class MHBase > struct fooFO { fooFO(MHBase<Derived>& mhb) : mhb_(mhb) {} void operator()(int i) { mhb_.foo(i); } private: MHBase<Derived>& mhb_; }; FooFunc getFooFunc() { FooFunc ff = fooFO<MHBase<Derived> >(*this); return ff; } }; class Put : public MHBase<Put> { friend class MHBAccessor; int foo_(int i) { return i; } }; int main(int argc, char **argv) { Put put; std::cout << put.foo(123) << std::endl; FooFunc ff = put.getFooFunc(); std::cout << ff(456) << std::endl; return 0; }

4 5

iostreams::stream: flush doesn't work but endl does
by Rob Desbois 23 Jan '08

23 Jan '08

Hi all, I've created my own iostreams device (for a serial port) which for debugging purposes prints out calls to read()/write(). I'm constructing a stream on the device like this: serial_device serial("/dev/ttyS3", serial_device::Baud38400); stream<serial_device> s(serial); If I do this: s << "hello world" << endl; serial.write() is called and does its debug output, but with this: s << "hello world\r" << flush; I get nothing despite using 'flush'. I've also tried calling s.flush() manually but get the same lack of effect. How come flush isn't working as it should here? TIA, --rob -- Rob Desbois

2 2

Solve linear system (tridiagonal matrix)
by nisha kannookadan 23 Jan '08

23 Jan '08

Hi everybody I need to solve a linear system Ax = b, where A is a tridiagonal matrix, is there an easy way to do it? I programmed an algo, which computes the inverse of the tridiagonal matrix with the usmani-algo and uses prod then, but this is quite time intense. I want to use the lapack bindings, but I dont get along with it. Im bit confused, if I need to install something additional to it, the way I tried out gesv, its not working and I get error msgs. I searched through out the www, but its hard to find good info. If somebody has already the code to solve linear equations with lapack, I would thankful, if he could post it here. Thank you Nisha K _________________________________________________________________ Express yourself instantly with MSN Messenger! Download today it's FREE! http://messenger.msn.click-url.com/go/onm00200471ave/direct/01/

2 1

Property Tree
by Glasgow, Craig C 23 Jan '08

23 Jan '08

Hello I'm using boost property_tree with the variant data type: typedef boost::variant<std::string, std::vector<std::string> > registry_entry; i.e I want to store either strings or vectors of strings at each key node. I'm using the inserters and extractors at the bottom of this email (similar to those included in the custom_data.cpp in the property_tree/examples folder). Does anyone know how I can write the tree out in xml format? The provided write_xml method can only handle property trees of string data type. Is there anyway to either tweak the code (e.g. modify the inserter and extractors) or wrap the variant class such that property_tree::write_xml will correctly render the tree as xml? Regards Craig // Custom extractor - converts data from boost::any to Type template<class Type> struct pt_extractor { inline bool operator()(const boost::any &data, Type &extracted, const std::locale &loc) const { extracted = boost::any_cast<Type>(data); return true; // Success } }; // Custom inserter - converts data from Type to boost::any template<class Type> struct pt_inserter { inline bool operator()(boost::any &data, const Type &to_insert, const std::locale &loc) const { data = to_insert; return true; // Success } }; // Custom property_tree traits - using custom extractor and inserter template<class Ch> struct pt_traits { // Character type to be used by ptree keys typedef Ch char_type; // Key type to be used by ptree typedef std::basic_string<Ch> key_type; // Data type to be used by ptree typedef boost::any data_type; // Extractor to be used by ptree template<class Type> struct extractor: public pt_extractor<Type> { }; // Inserter to be used by ptree template<class Type> struct inserter: public pt_inserter<Type> { }; // Key comparison function inline bool operator()(const key_type &key1, const key_type &key2) const { return _stricmp(key1.c_str(), key2.c_str()) < 0; } }; typedef boost::property_tree::basic_ptree<pt_traits<char> > pt_rep; typedef boost::variant<std::string, std::vector<std::string> > registry_entry; ***************************************************************************** This communication is sent by the Standard Bank Plc or one of its affiliates The registered details of Standard Bank Plc are: Registered in England No. 2130447, Registered Office 25 Dowgate Hill London EC4R 2SB Authorised and Regulated by the Financial Services Authority. More information on Standard Bank is available at www.standardbank.com Everything in this email and any attachments relating to the official business of Standard Bank Group Limited and any or all subsidiaries, the Company, is proprietary to the Company. It is confidential, legally privileged and protected by relevant laws. The Company does not own and endorse any other content. Views and opinions are those of the sender unless clearly stated as being that of the Company. The person or persons addressed in this email are the sole authorised recipient. Please notify the sender immediately if it has unintentionally, or inadvertently reached you and do not read, disclose or use the content in any way and delete this e-mail from your system. The Company cannot ensure that the integrity of this email has beenmaintained nor that it is free of errors, virus, interception or interference. The sender therefore does not accept liability for any errors or omissions in the contents of this message which arise as a result of e-mail transmission. If verification is required please request a hard-copy version. This message is provided for informational purposes and should not be construed as a solicitation or offer to buy or sell any securities or related financial instruments. *****************************************************************************

1 0

asio socket options
by Eugene M. Kim 23 Jan '08

23 Jan '08

Greetings, On UNIX-like platforms, boost::asio::detail::socket_option::{boolean,integer} templates are very useful when defining socket option classes for non-standard options that are supported by the platform, e.g.: namespace so = boost::asio::detail::socket_option; typedef so::boolean<SOL_SOCKET, SO_REUSEADDR> reuse_addr; typedef so::boolean<SOL_SOCKET, SO_REUSEPORT> reuse_port; typedef so::boolean<SOL_SOCKET, SO_ERROR> pending_error; typedef so::integer<SOL_SOCKET, SO_SNDLOWAT> send_low_watermark; typedef so::integer<SOL_SOCKET, SO_RCVLOWAT> recv_low_watermark; typedef so::boolean<IPPROTO_IP, IP_HDRINCL> include_header; typedef so::boolean<IPPROTO_IP, IP_DONTFRAG> do_not_fragment; typedef so::integer<IPPROTO_IP, IP_TOS> type_of_service; typedef so::integer<IPPROTO_IP, IP_TTL> time_to_live; ... and so on. It is not suitable to include these non-standard options in the asio library itself, but it seems justifiable to move the boolean/integer templates out of boost::asio::detail and into boost::asio. Also, many socket options take a C struct as an argument, so a template for such options might be useful as well: template <int Level, int Name, class Value> class typed : public Value { typedef Value value; template <typename Protocol> int level(const Protocol &) const { return Level; } template <typename Protocol> int name(const Protocol &) const { return Name; } template <typename Protocol> value *data(const Protocol &) { return (value *)this; } template <typename Protocol> std::size_t size(const Protocol &) const { return sizeof(Value); } template <typename Protocol> void resize(const Protocol &, std::size_t s) const { if (s != sizeof(Value)) throw std::length_error("resizing fixed-length socket option"); } }; Then it becomes trivial to define and use such socket options: typedef typed<SOL_SOCKET, SO_SNDTIMEO, struct ::timeval> send_timeout; typedef typed<SOL_SOCKET, SO_RCVTIMEO, struct ::timeval> recv_timeout; send_timeout sndtimeo; sndtimeo.tv_sec = 1; sndtimeo.tv_usec = 0; Would this template be appropriate for inclusion in a future version of asio? Just my $.02. Cheers, Eugene

1 0

meta-programming library
by noman javed 23 Jan '08

23 Jan '08

Hi I need to work with boost-mpl library. Should I install the whole boost library or is there any possibility to just install boost-mpl ? Thanks in anticipation Noman Javed --------------------------------- Be a better friend, newshound, and know-it-all with Yahoo! Mobile. Try it now.

2 1

The Scanner Business
by jono 23 Jan '08

23 Jan '08

hello possums! i have the following rules declared in a spirit grammar definition (using boost 1.8.4 )..... //----------------------- typedef typename lexeme_scanner<ScannerT>::type lexeme_scanner_t; rule<lexeme_scanner_t> legal_initial_p ,legal_p ; //----------------------- and initialized in definition::definition() .... //----------------------- legal_initial_p = alpha_p | ch_p('_') |'<'|'>'|'='|'?'|'/'|'*'|'&'|'-'|'+'|'%' |'!'|'$'|':'|'@' ; legal_p = legal_initial_p | digit_p | ch_p('.') ; //----------------------- the above code compiles and functions as expected under MSVC 7.1 but causes g++ 4.1.1 to hang. if i use type rule<ScannerT> for legal_p and legal_initial_p, the code compiles on both, but doesn't work properly. i'm using |dynamic_distinct_parser ||so i can handle the '.' character as a syntactic token in certain contexts. |||//---------------------- |some declarations ... || struct keywords : symbols<Object> { ... } keywords_p; || ||dynamic_distinct_parser<ScannerT> keyword_p; ||||| //----------------------|some definitions ...| | ||chset<> distinct(chset<>(legal_ident) | chset<>("0-9")); distinct_directive<> keyword_d(distinct); distinct_parser<> keyword_p(distinct); |||| || ||rule<ScannerT, object_closure::context_t> ||Symbol_p = lexeme_d [ ( legal_initial_p >> *legal_p ) - keywords_p ] [ Symbol_p.value_ = intern_(arg1, arg2) ] ; || ||| dot_p = keyword_p("."); //---------------------- |// .... ||it's quite possible there's a better way to do the parsing, but it was fiddly, and i don't have any more time to mess with it. (the rest of the grammar is a recursive scheme reader, hence the closures) i would be very grateful for any suggestions as to what might be causing g++ to go walkabout! cheers, jono jp(a)day-one.com | | |

1 0

Regex newbie problem
by Barberi 23 Jan '08

23 Jan '08

Newbie to boost. I used the MSVC 8.0 installer, and got the first couple of example programs to work, but failed on the regex example. I'm using MSVC 2008 (aka MSVC-9.0) (Express Edition) IDE, in Debug configuration. The linker can't resolve several symbols with libboost_regex_vc80-mt-gd-1_34_1.lib. Stripped of the qualifiers these are: _Orphan_all(void) const, ~_Container_base(void), _Swap_all(...), and insert(...). I switched to release configuration, now the linker looks in libboost_regex_vc80-mt-1_34_1.lib and can't find insert(...). Any suggestions? Should I wait for a MSVC-9.0 installer, and go back to MSVC 2005 in the meantime? Stephen B.

3 8

Boost, HP-UX and gcc
by Martin_Thomas＠McAfee.com 22 Jan '08

22 Jan '08

Hello, I am trying to use use Boost::Regex on HP 11i with gcc 4.2.2 and I notice that in config/platform/hpux.hpp there is a define to turn off wide-char support unless using aCC and have a1 source eg: #if !(defined(__HP_aCC) || !defined(_INCLUDE__STDC_A1_SOURCE)) # define BOOST_NO_SWPRINTF # define BOOST_NO_CWCTYPE #endif This will never allow gcc to use wregex even if wchar support is available. Is this because there are so few people using Zgcc oin HP-ux or am I missing something? On the assumption that it is related to usage, I have added the following to allow gcc 4 to use wregex: #if defined(__GNUC__) # if (__GNUC__ >3) # undef BOOST_NO_CWCTYPE # endif #endif Is there anything obviously wrong with this? Thanks // Martin

4 8

[program_options] crash on application exit
by ope 22 Jan '08

22 Jan '08

Hi, to get the main() more clean, I moved the command line options parser to an own class (some stuff missing): ---8--- class CommandLineOptions { public: CommandLineOptions( int argc, char** argv ); private: po::options_description generic_options; po::options_description qt4_common; #if defined( Q_OS_UNIX ) || defined( Q_OS_MAC ) po::options_description qt4_x11; #endif #if defined( Q_OS_WIN ) po::options_description qt4_win; #endif #if defined DEBUG po::options_description qt4_dbg; #endif po::options_description hidden_options; po::options_description usage_options; po::options_description cmdline_options; po::positional_options_description pos_options; po::variables_map vm; }; --->8--- with ---8<--- CommandLineOptions::CommandLineOptions( int argc, char** argv ) : generic_options("Generic options"), qt4_common("Qt options"), #if defined( Q_OS_UNIX ) || defined( Q_OS_MAC ) qt4_x11( "Qt X11 specific options" ), #endif #if defined( Q_OS_WIN ) qt4_win( "Qt Windows specific options" ), #endif #if defined( DEBUG ) qt4_dbg( "Qt debug options" ), #endif hidden_options("Hidden options") { generic_options.add_options() ( "version,v", "print version string") ( "help,h", "produce help message") ; hidden_options.add_options()(...); qt4_common.add_options()(... #if defined( Q_OS_UNIX ) || defined( Q_OS_MAC ) qt4_x11.add_options()(... #if defined( Q_OS_WIN ) qt4_win.add_options()(... #if defined( DEBUG )(... usage_options .add( generic_options ) .add( qt4_common ) #if defined( Q_OS_UNIX ) || defined( Q_OS_MAC ) .add( qt4_x11 ) #elif defined( Q_OS_WIN ) .add( qt4_win ) #endif #if defined( DEBUG ) .add( qt4_dbg ) #endif ; cmdline_options.add( usage_options ).add( hidden_options ); pos_options.add( "input-file", -1 ); po::store(po::command_line_parser(argc, argv). options(cmdline_options).positional(pos_options).run(), vm); po::notify(vm); if (vm.count( "version" )) { std::cout << Version() << "\n"; } ... } --->8--- and the main(): ---8<--- int main( int argc, char **argv ) { try { logging::init(); CommandLineOptions cmdline( argc, argv ); QT_REQUIRE_VERSION( argc, argv, QT_VERSION_STR ); Q_INIT_RESOURCE( workbench ); QApplication app( argc, argv ); QMainWindow wb; wb.show(); // using boost.logging v2 L_ << "execute application."; return app.exec(); ... --->8--- It works fine until the app exits. My gdb session wasn't very useful to me: (gdb) target exec ./edas (gdb) run Starting program: /.../edas2/edas Using host libthread_db library "/lib/i686/cmov/libthread_db.so.1". (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) [Thread debugging using libthread_db enabled] [New Thread 0xb70506c0 (LWP 13445)] [New Thread 0xb6aa5b90 (LWP 13450)] 28:31 [1] execute application. Program received signal SIGSEGV, Segmentation fault. [Switching to Thread 0xb70506c0 (LWP 13445)] 0x08051980 in ?? () (gdb) backtrace #0 0x08051980 in ?? () If I comment out my CommandLineOptions class instance it works. Thanks, Olaf

1 0

boost::filesystem::rename - behavior/code vs. docs
by Andriy Gapon 22 Jan '08

22 Jan '08

This is what documentation on boost::filesystem::rename states: http://www.boost.org/libs/filesystem/doc/tr2_proposal.html template <class Path1, class Path2> void rename(const Path1& from_p, const Path2& to_p); Requires: Path1::external_string_type and Path2::external_string_type are the same type. Effects: Renames from_p to to_p, as if by POSIX rename(). Postconditions: !exists(from_p) && exists(to_p), and the contents and attributes of the file originally named from_p are otherwise unchanged. [Note: If from_p and to_p resolve to the same file, no action is taken. Otherwise, if to_p resolves to an existing file, it is removed. A symbolic link is itself renamed, rather than the file it resolves to being renamed. -- end note] Actual effect is that an exception is thrown when renaming a file to an existing file on the same fs. rename(2) performs an 'atomic' rename in this case without problems, but in the boost code there is a specila check against "too permissive" posix rename(): http://boost.cvs.sourceforge.net/boost/boost/libs/filesystem/src/operations… 1126 rename_api( const std::string & from, const std::string & to) 1127 { 1128 // POSIX is too permissive so must check 1129 fs::system_error_type dummy; 1130 if ( fs::exists( status_api( to, dummy ) ) ) 1131 return EEXIST; 1132 return std::rename( from.c_str(), to.c_str() ) != 0 1133 ? errno : 0; 1134 } And rename testing code also checks for the above behavior. I think that this is a huge difference between the docs and the code, and this is very counter-intuitive too. -- Andriy Gapon

1 0