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

[bind] Template deduction
by Robert Dailey 05 Mar '08

05 Mar '08

Hi, Suppose I have the following code: template< typename T > void DoFoo( int number, T object ) { } void main() { std::list<std::string> myStrings; // Assume myStrings is full of something... std::for_each( myStrings.begin(), myStrings.end(), boost::bind( &DoFoo, 5, _1 ) ); } On MSVC 9, this will not compile because DoFoo is a template, and at some point, boost::bind can't take a template as the function to call, or it can't deduce the function's template arguments. Is there a way I can make this work? Thanks.

2 1

boost posix time
by Aljaz 04 Mar '08

04 Mar '08

Hello I'm developing server and client applications and I need ping-pong mechanism to get ping time betwen them (time needed for data to arrive).. I've been thinking the best way would be to send Code: boost::posix_time::microsec_clock::universal_time()from server to client, which would then just reply with its own universal_time() .. The first problem is I dont know how to store universal_time() as an int/long.. I know I can store it as a string but this would take extra cpu to parse the data each time (with boost posix library).. All the data I send is serialized - maybe I could just serialize the Code: boost::posix_time::ptimeobject and send over? The next thing is what if server's time zone is gmt+1 and client's gmt+3? Will that affect the the time if I do the subtraction to get the ping time? If so, what can I do about it? Maybe anyone has better solution in mind? Thanks a lot for help

2 1

Ann: Floating Point Utilities Review starts today
by John Maddock 04 Mar '08

04 Mar '08

The review of Johan Rade's floating point utilities starts today. Code and docs can be downloaded from : http://www.boost-consulting.com/vault/index.php?action=downloadfile&filenam… The library consists of three parts: 1) Floating point classification routines: these are optimised implementations of the C99 and C++ TR1 functions fpclassify, isinf, isnan, isnormal and isfinite. From Boost-1.35 onwards these are already a part of Boost.Math (see http://svn.boost.org/svn/boost/trunk/libs/math/doc/sf_and_dist/html/math_to…) so if accepted the two implementations will get merged. The review here should focus on the implementation used, and testing on whatever platforms you have available - in particular are there any circumstances (compiler optimisation settings etc) where this implementation breaks? 2) Sign manipulation functions: implementations of the C99 and C++ TR1 functions copysign and signbit, plus the changesign function. Two of these (signbit and copysign) are currently undocumented members of Boost.Math, and again the two implementations will get merged if this library is accepted. Again the main focus of the review here is the implementation, and testing thereof especially in the presence of compiler optimisations. 3) C++ locale facets: these will read and write non-finite numbers in a portable and round-trippable way: that is not otherwise possible with current C++ std library implementations. These are particularly useful for number-serialisation for example. Since the design is already specified by the C++ standard for these facets, your review here should focus on implementation, testing, documentation, and perhaps where in Boost these should best be placed if accepted. These look to be a useful collection of utilities, so I'll look forward to your reviews, Regards, John Maddock, Floating Point Utilities Review Manager.

9 34

Boost::random: Using linear congruential etc.
by Hans Mull 04 Mar '08

04 Mar '08

Hi! A few days ago I started to try out the boost::random library. Now I have a problem: If you don't want to use the "special cases" like minstd_rand[0] for linear_congruential hellekalek1995 for inverse congruential algorithm, how do you use them? I have non-constant parameters (which are not know at compile time). It would be nice if you give me some examples about how to use these "general case" algorithms. Best regards, Hans

1 0

boost::iostreams zlib compression
by j.c. 04 Mar '08

04 Mar '08

Ok I think this sucks. :-) Here is my code to implement zlib compression/decompression via boost. However since is uses streams I cannot simply pass a buffer. This code is intended for compress socket buffers, but is pathetically slow. HAs anyone done this? What can I improve? The copy seems to be the bottle neck. thx jc Here is my code: #include <boost/iostreams/copy.hpp> #include <boost/iostreams/filter/zlib.hpp> #include <boost/iostreams/device/file.hpp> #include <boost/iostreams/stream.hpp> #include <boost/iostreams/filtering_stream.hpp> #include <boost/iostreams/detail/config/zlib.hpp> using namespace boost::iostreams; std::string compress(std::string data) { std::stringstream compressed; std::stringstream decompressed; decompressed << data; boost::iostreams::filtering_streambuf<boost::iostreams::input> out; out.push(boost::iostreams::zlib_compressor()); out.push(decompressed); boost::iostreams::copy(out, compressed); return compressed.str(); } std::string decompress(std::string data) { std::stringstream compressed; std::stringstream decompressed; compressed << data; boost::iostreams::filtering_streambuf<boost::iostreams::input> out; out.push(boost::iostreams::zlib_decompressor()); out.push(compressed); boost::iostreams::copy(out, decompressed); return decompressed.str(); } int main() { std::string data = compress("boo!"); std::cout << data << std::endl; std::cout << decompress(data) << std::endl; return 0; }

3 6

compiling 64 bit boost on Mac OS 10.5
by Tom Smith 04 Mar '08

04 Mar '08

Hi, I've been trying and failing to get a 64 bit build of the boost libraries on 10.5. I always seem to end up with 32 bit binaries. I've tried various permutations involving bjam address-model=64 (I last tried to build a week ago so I'm hazy on the actual command line). I only need x64, not ppc64. Cheers, Tom

3 5

iterator_adapter/iterator_tag/std::distance bug
by Martin Wartens 04 Mar '08

04 Mar '08

I used boost::iterator_adapter with a std::vector iterator, all parameters set to boost::use_default. (So this is the same as a plain std::vector iterator) With this configuration, the iterator tag is recognized as "input_iterator_tag" not "random_iterator_tag" as it must be. This hurts because the iterator tag is used to select the right std::distance function. I could solve the problem by explicitly setting the traversal type template parameter to "std::random_access_iterator_tag", while "boost::random_access_traversal_tag" also didn't work. The compiler is VC8.

2 2

serialization - xml
by Ambardekar, Mukund 03 Mar '08

03 Mar '08

Hi, I used boost serialization to write my data in binary format as in below way - std::ofstream ofs(catalogFileName.c_str(),std::ios::binary); boost::archive::binary_oarchive oa(ofs); //write oa & m_myclass; // object of CMyClass type containing some data (say name) This worked correctly. Now I wanted to write into xml format. Did the below changes std::ofstream ofs(catalogFileName.c_str(),std::ios::binary); boost::archive::xml_oarchive oa(ofs); //write oa & m_myclass; But it is giving me below error - d:\tools\boost-1feb08\boost_1_34_1\boost\archive\basic_xml_oarchive.hpp( 83) : error C2027: use of undefined type 'boost::STATIC_ASSERTION_FAILURE<x>' Can you help me to identify the problem? FYI - I have made no changed CMyClass looks like as below. And I have made no changes in it (same one used for binary and xml) class CMyClass { public: Part(const std::wstring& Name); private: std::wstring m_name; friend class boost::serialization::access; template<class Archive> void serialize(Archive& ar, const unsigned int) { ar & m_name; } }; Thanks in advance. Regards, Mukund

2 1

Minimal human-readable boost::serialization ?
by Eric Newhuis 03 Mar '08

03 Mar '08

I looks like the last time this was mentioned in this mailing list was in 2005. Has anyone considered anything like boost::archive::json_oarchive etc.? JSON looks easy to read in log files. I want name-value pairs in my serialization. But I don't want the verbosity of XML. I think JSON would be ideal.

2 1

Better debugger support for boost::optional with VS2005 compiler
by Daniel Krügler 03 Mar '08

03 Mar '08

Hello, just recently I introduced boost::optional as default type to represent optional/nullable types in a sub-project of our company, which generates C++ code out of UML models. While this worked seamless, programmers not yet used to boost::optional complained that they could not really see, what such type "contains" if they try to observe it in the debugger. I searched for a solution, only finding the discussion around the article "how to see the actual value of a 'boost::optional<__int64>' in VS2005 debugger" from 2006-07-26, which did not gave a satisfactory answer. Please ignore this message, if the here described solution (or a better one), which I worked out this weekend, is already well-known (but I would appreciate if someone could give me a link to the already known solution). My approach was to modify the unofficially supported debugger configuration file autoexp.dat, which typically is found at ${SomePath}\Microsoft Visual Studio 8\Common7\Packages\Debugger\autoexp.dat If you never have heard of this file, you should make a copy of it, before doing the modifications I discuss here. A more detailled description what can be done with this file, can be read here (but you don't need to do that, if you just want to get the results I propose here): http://www.virtualdub.org/blog/pivot/entry.php?id=120 Just use your preferred text editor to open the autoexp.dat, jump at the end of it and go five rows back. You should be +/- 1 row about the line with the content: ; This section lets you define your own errors for the HRESULT display. Go one row back (row 1957 in my file), which should be the end of the so-called [Visualizer] section. Now insert here the text provided at the end of this article and just start you debugger session (you don't need to restart the studio) - you should have a nice view on each boost::optional, even those which are optionals of references. I tested the visualizer for the following boost versions: 1.30.0 1.31.0 1.32.0 1.33.0 1.33.1 1.34.1 and all of them worked. I would like to thank the boost developers for their great contributions (where boost::optional is only one of them) and I hope you can even more enjoy boost::optional with this trick. Unfortunately I have not found similar solutions for other compilers yet, e.g for gcc, which probably is due to my lack of experience with them but I would appreciate if someone else publishes them, too. Greetings from Bremen, Daniel Krügler ;------------------------------------------------------------------------------ ; Visualizer for boost::optional, boost::optional_detail::optional_base ;------------------------------------------------------------------------------ boost::optional<*>|boost::optional_detail::optional_base<*> { preview ( #if($e.m_initialized) ( *(($T1 *) (&$e.m_storage.dummy_.data[0])) ) #else ( "?" ) ) children ( #if($e.m_initialized) ( #(value: *(($T1 *) (&$e.m_storage.dummy_.data[0])) ) ) #else ( "?" ) ) } ;------------------------------------------------------------------------------

3 3

[filesystem] bug???
by Ovanes Markarian 03 Mar '08

03 Mar '08

Hello *! A short question: removing the leaf from the path using the remove_leaf() results in my case (MS VC7.1, boost 1.34.1) in the path with the trailing slash at the end. Next call to remove_leaf() results in removing this trailing slash, which is not correct... Is it a bug, if the doc states to remove_leaf: basic_path& remove_leaf(); Effects: If has_branch_path() then remove the last filename from the stored path. If that leaves the stored path with one or more trailing slash elements not representing root-directory, remove them. Many thanks, Ovanes

1 0

A possible bug in Boost.property_tree with UNICODE XML file?
by Tan, Tom (Shanghai) 03 Mar '08

03 Mar '08

I am trying read_xml() on an xml-format config file, which starts like this: <?xml version="1.0" encoding="UTF-16"> ..... This file is well-formed and verified by "XML Copy Writer", it can also be open by IE or Filefox with no problem. I am using this code snippet to parse this file: try{ read_xml<wptree>(xmlfile,pt); } catch(xml_parser-error e) { e.what(); } and it throws an exception saying "expected <". By looking into the code I found that, this caused by the fact that the xmlfile starts with a Unicode Byte Order Mark (BOM)-0xFFFE . "fffe 3c00 3f00 7800 ......" However, read_xml() cannot handle this BOM byte and throws. It's easy to fix it by simply discard the BOM byte, but I wonder whether there is already a flag to enable/disable this that I am unaware of. Can anyone help to clarify? Thanks, tom

1 0

Re: [Boost-users] templates: parts of functions
by Lang Stefan 03 Mar '08

03 Mar '08

Hi, As I am not overly familiar with boost or it's mpl library yet, I cannot provide any specific advice along these lines. According to your description, my advice would be to create helper template classes or functions that rely only on some of the template parameters of your actual template class, but not N. For these helper functions and classes you do not need partial specialization, and thus no duplicate code. Depending on the complexity of your functions you might also extract those parts that do not depend on N into more template functions. If these helpers need access to private members of your makeTree class, you can declare them friend. (you might need to replicate those friend declarations in your partial specializations) For those parts of the code that do depend on N, but do not need a different implementation, you may still provide templated helper classes. The compiler will still generate multiple class implementations for different values of N, but you won't need to replicate the actual code yourself. Hope this helps. Cheers, Stefan > ------------------------------ > > Message: 6 > Date: Sat, 1 Mar 2008 14:38:34 +0100 > From: "Hicham Mouline" <hicham(a)mouline.org> > Subject: [Boost-users] templates: parts of functions > To: <boost-users(a)lists.boost.org> > Message-ID: <000001c87ba1$8cfe0d30$a6fa2790$@org> > Content-Type: text/plain; charset="us-ascii" > > Hi, > > > > I have a function makeTree whose behavior depends on 3 > different policies > > typename T1 > > typename T2 > > int N > > > > For performance reasons, I need to write the 2- and 3- specializations > alongside the general N case. That is, the 2- and 3- default > > instantiations generated from the N case by the compiler are > not optimal. > > > > As it is impossible to write function templates partial > specializations, I > encapsulate the > > > > template< typename T1, typename T2, int N> void > makeTree(Head<N>&, const > T2&, const C&) ; > > > > function in a class template > > > > template< typename T1, typename T2, int N> class makeTree{ > > void makeTree(Head<N>&, const T2&, const C&); > > }; > > and write the partial specializations <T1,T2,2> and <T1,T2,3>. > > > > > > Now, this makeTree function is quite large. It includes code that : > > . doesn't depend on any of the template arguments > > . code that depends only T1 > > . code that depends only on T2 > > . code that depends only on N > > . code that depends only on <T1,N>. > > > > I am worried about code bloat, and the effect on overall > performance of the > actual code generated even if it is not run > > (I noticed substantial effects depending on the size of > object files and > final shared library generated). > > > > Besides this, I suppose it is a better design to have these > segments of code > put in template functions > > with the right template arguments. > > > > Is it MPL the library that I should look into for such work? > > Perhaps nothing in Boost particularly facilitates this work, > and std c++ > templates are enough, > > > > Thoughts, comments are appreciated, > > > > Rds, >

1 0

Why does runtime-debugging set _STLP_USE_DYNAMIC_LIB?
by Eric Woodruff 03 Mar '08

03 Mar '08

I've been trying to understand why the debug version of boost was attempting linkage with the DLL version of stlport (which I didn't build and I'm not using) and now I see that the runtime-debugging flag is specifying that the dynamic library must be used. This causes the static stlport to use the stlpdx cross namespace rather than the stlpd namespace. It is not obvious to me why debugging should require a dynamic library. Under code inspection this looks like a copy-paste oversight. if [ $(rproperties).get <runtime-debugging> ] = "on" { usage-requirements += <define>_STLP_DEBUG=1 <define>_STLP_DEBUG_UNINITIALIZED=1 ; } if [ $(rproperties).get <*runtime-debugging*> ] = "on" { usage-requirements += <define>_STLP_USE_DYNAMIC_LIB=1 ; } Why would it be checking for runtime-debugging twice? I think the second if should check for <runtime-link> which has the possible values "static" or "dynamic". This problem looks related to this thread: http://lists.boost.org/boost-users/2007/09/30905.php, but I couldn't find the Boost.Build message that is a follow-up to that one. I haven't found a posted solution either. Looking at the CVS HEAD revision of stlport.jam, it hasn't changed since 1_34_1. So if this is a bug, it has not been fixed yet. Thanks, Eric

2 1

Re: [Boost-users] fatal error C1128: object file format limit exceeded : more than 65, 279 sections -------PROBLEM SOLVED
by jono 03 Mar '08

03 Mar '08

i thought i had broken the translation unit in two, but it was a clever illusion ... doh.

1 0

fatal error C1128: object file format limit exceeded : more than 65, 279 sections
by jono 02 Mar '08

02 Mar '08

hi, i've received this error compiling a project under visual studio 7.1. "fatal error C1128: object file format limit exceeded : more than 65,279 sections" the problem appeared when i added a couple more specializations to a 'binary operator' visitor for manipulating boost variant objects. before this, there were 20 or 30 similar specializations compiling without errors or warnings. the visitor and helpers are as follows: =========================== template<BinaryOp op, typename ObjectT, typename ObjectU> Object binary_op(const ObjectT &arg1, const ObjectU &arg2) { return undefined; } template<BinaryOp op, typename ObjectT> Object binary_op(const ObjectT &arg1) { return undefined; } template<BinaryOp op> class BinaryOpVisitor : public boost::static_visitor<Object> { public: template<typename ObjectT> Object operator()( const ObjectT &arg0) const; { return binary_op<op, ObjectT>(arg0); } template<typename ObjectT, typename ObjectU> Object operator()( const ObjectT &arg0, const ObjectU &arg1) const { return binary_op<op, ObjectT, ObjectU>(arg0, arg1); } }; ======================== and specialized: ======================== template<> INLINE Object binary_op<ADD>(const Integer& arg0, const Integer& arg1){return make<Integer>(arg0 + arg1);} .... .... .... .... etc etc ======================== the variant type looks like this.... ======================== typedef boost::make_recursive_variant< Nil ,Integer ,Real ,String ,Error ,Class ,Character ,Stream ,Boolean ,boost::recursive_wrapper< Symbol > ,boost::recursive_wrapper< Lambda > ,boost::recursive_wrapper< Function > ,boost::recursive_wrapper< Pair > >::type Atom; ======================== i imagine there are some combinatorial proliferations of template instantiation going on in the compiler. the operators were declared inline, but removing that didn't fix the error. i've restructured the source code in other ways, but without success so far. visual studio 8 has a compiler switch /bigobj to resolve this problem: "By default, an object file can hold up to 65,536 (2^16) addressable sections. /bigobj increases that address capacity to 4,294,967,296 (2^32)." i haven't tried compiling this project with g++ since making these changes (haven't checked them in) so i don't know what its limits are. does anyone has experience with solving this problem. is it a problem with my visitor class? is there a way i can break my recursive variant down for the compiler? cheers, Jono Poff

1 0

a difference of boost.function between g++4.13 and vc8
by 钱晓明 02 Mar '08

02 Mar '08

I wrote some code using boost.function. Compiling it both in g++4.13 and vc8, I find that I can not declare a struct in function with operator() and use boost.function to call it when I use g++. Just like this: int main() { struct int_div { float operator()(int x, int y) const { return ((float)x)/y; }; }; boost::function<float (int x, int y)> f = int_div(); std::cout << f(5, 3) << std::endl; return 0; } When I compile the code above in g++, it give me the error : 请求从 'main()::int_div' 转换到非标量类型 'boost::function<float ()(int, int), std::allocator<void> >'(I use ubuntu chinese.) but when I move the int_div definition outside of main(), It is ok! And the code above VC8 can compile it! why?

2 1

Re: [Boost-users] [boost] Serialization of Range library's sub_range
by brass goowy 02 Mar '08

02 Mar '08

Brian Wood writes: >Recently, I suggested there should be serialization support wrt >the Range library - >http://lists.boost.org/Archives/boost/2008/01/132294.php. >I'm writing to inform you of my plans in this regard. This input: > >Msgs > (sub_range<list<short> >) @OUT >} > >to the C++ Middleware Writer, http://home.seventy7.com/cgi-bin/samb.cgi, >results in this output: > >#include <boost/range/sub_range.hpp> >#include <list> >#include <Buffer.hh> >using boost::sub_range; > > >struct Msgs >{ >inline >Msgs() {} >inline >~Msgs() {} > >inline >int >Send(Buffer* buf, const sub_range<list<short> >& about1) >{ > unsigned int headCount = 0; > headCount = about1.size(); > if (!buf->Receive(&headCount, sizeof(int))) { > buf->ews_.SetErrorWords(1, __FILE__, __LINE__); > return 0; > } > list<short>::const_iterator rit1 = boost::begin(about1); > list<short>::const_iterator rit2 = boost::end(about1); > for (; rit1 != rit2; ++rit1) { > if (!buf->Receive(&(*rit1), sizeof(short))) { > buf->ews_.SetErrorWords(1, __FILE__, __LINE__); > return 0; > } > } > > if (!buf->SendStoredData()) { > buf->ews_.SetErrorWords(2, __FILE__, __LINE__); > return 0; > } > return 1; >} > >}; > >This functionality isn't available online yet. G-d willing, it >will be available by the middle of next month. Update: the functionality is now available online. Brian Wood Ebenezer Enterprises www.webebenezer.net

1 0

Boost file system idiocy
by Patrick Loney 02 Mar '08

02 Mar '08

Can't find how to seek to eof, currently using: fs.seekp( (std::streamoff)boost::filesystem::file_size( SKDO::ArchiveFilePath( keyname ) ) ); please help ****************************************************************************** "This message and any attachments are solely for the intended recipient and may contain confidential and privileged information. If you are not the intended recipient, any disclosure, copying, use, or distribution of the information included in this message and any attachments is prohibited. If you have received this communication in error, please notify us by reply e-mail and immediately and permanently delete this message and any attachments. Thank you." Interactive Transaction Solutions Ltd (2473364 England) Registered Office: Systems House, Station Approach Emsworth PO10 7PW ********************************************************************** Ce message �lectronique contient des informations confidentielles � l'usage unique des destinataires indiqu�s, personnes physiques ou morales. Si vous n'�tes pas le destinataire voulu, toute divulgation, copie, ou diffusion ou toute autre utilisation de ces informations, est interdite. Si vous avez re�u ce message �lectronique par erreur, nous vous remercions d'en avertir son exp�diteur imm�diatement par email et de d�truire ce message ainsi que les �l�ments attach�s. Interactive transaction Solutions SAS- France (RCS Pontoise : 489 397 877) Si�ge social : Parc Saint Christophe, 10, Avenue de l�Entreprise 95865 Cergy-Pontoise Cedex ______________________________________________________________________ This email has been scanned by the MessageLabs Email Security System. For more information please visit http://www.messagelabs.com/email ______________________________________________________________________

2 1

Couting occourance of specified characters
by Olaf Peter 01 Mar '08

01 Mar '08

Hi, the snipped attached counts the accourance of a specified character. Is there a simpler way using the boost libs? This solution looks quite overkill. It should be fast ;-) Thanks, Olaf

2 1

Bind + Function + Arguments
by QbProg 01 Mar '08

01 Mar '08

Hello there, I've started using boost function and bind features , and I'd like to find a way to do the following. You can take it as suggestion, if it is not implemented :) I would like to use the bind function on object members in the same way indipendently of the number of arguments. For example: function <void (int , int ) > FunctionOne; function < bool (int ) > FunctionTwo; class Implementor { public: void ImplOne ( int , int ); bool ImplTwo ( int ); }; void test () { Implementor i; FunctionOne = bind ( &Implementor::ImplOne , &i , _1 , _ 2); FunctionTwo = bind (&Implemenetor::ImplTwo, &i , _1); /* I Would be nice to do something like this, with parameter "guessing" */ // FunctionOne = GenericBind ( &Implementor::ImplOne ,&i); // FunctionTwo = GenericBind ( &Implementor::ImplTwo, &i); /* This will greatly remove verbosity! */ /* Actually I do */ #define Bind0(A,B) bind(A,B) #define Bind1(A,B) bind(A,B,_1) #define Bind2(A,B) bind(A,B,_1,_2) ...... FunctionOne = Bind2 ( &Implementor::ImplOne , &i ); FunctionTwo = Bind1 ( &Implementor::ImplTwo , &i); } It would be usefoul to make library callbacks more usable, insteat of writing a long list of arguments each time. It is possible with current implementation? I ask you for any suggestion, Thank you QbProg

2 3

[Review] Proto review starts today, March 1st
by Hartmut Kaiser 01 Mar '08

01 Mar '08

Hi all, The review of Eric Nieblers Proto library starts today, March 1st 2008, and will end on March 14th. I really hope to see your vote and your participation in the discussions on the Boost mailing lists! --------------------------------------------------- About the library: Proto is a framework for building Domain Specific Embedded Languages in C++. It provides tools for constructing, type-checking, transforming and executing expression templates. More specifically, Proto provides: * An expression tree data structure. * Operator overloads for building the tree from an expression. * Utilities for defining the grammar to which an expression must conform. * An extensible mechanism for immediately executing an expression template. * An extensible set of tree transformations to apply to expression trees. * A mechanism for giving expressions additional behaviors and members. Documentation is here: http://boost-sandbox.sourceforge.net/libs/proto Download proto.zip from here: http://www.boost-consulting.com/vault/index.php?directory=Template%20Metapro gramming Proto is a very important infrastructure library, IHMO. It has been used as the backbone for several other library writing efforts already, such as Xpressive and the rewrite of Spirit (there it has been used for all three parts: the parser, the lexer and the generator modules), and it is a key enabling technology in the long-planned unification of Phoenix and the Lambda Library. --------------------------------------------------- Please always state in your review, whether you think the library should be accepted as a Boost library! Additionally please consider giving feedback on the following general topics: - 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? Regards Hartmut Review Manager

1 0

templates: parts of functions
by Hicham Mouline 01 Mar '08

01 Mar '08

Hi, I have a function makeTree whose behavior depends on 3 different policies typename T1 typename T2 int N For performance reasons, I need to write the 2- and 3- specializations alongside the general N case. That is, the 2- and 3- default instantiations generated from the N case by the compiler are not optimal. As it is impossible to write function templates partial specializations, I encapsulate the template< typename T1, typename T2, int N> void makeTree(Head<N>&, const T2&, const C&) ; function in a class template template< typename T1, typename T2, int N> class makeTree{ void makeTree(Head<N>&, const T2&, const C&); }; and write the partial specializations <T1,T2,2> and <T1,T2,3>. Now, this makeTree function is quite large. It includes code that : . doesn't depend on any of the template arguments . code that depends only T1 . code that depends only on T2 . code that depends only on N . code that depends only on <T1,N>. I am worried about code bloat, and the effect on overall performance of the actual code generated even if it is not run (I noticed substantial effects depending on the size of object files and final shared library generated). Besides this, I suppose it is a better design to have these segments of code put in template functions with the right template arguments. Is it MPL the library that I should look into for such work? Perhaps nothing in Boost particularly facilitates this work, and std c++ templates are enough, Thoughts, comments are appreciated, Rds,

1 0

static assert
by Hicham Mouline 01 Mar '08

01 Mar '08

hi, I would like to avoid template instantiations with ContractType whose static functions st1 and st2 return an enum... template <typename ContractType> class C { BOOST_STATIC_ASSERT( !(ContractType::st1()==A && ContractType::st2()==P) ); ... }; Is there a way to make this work?

4 4

Re: [Boost-users] A request for help using the Boost Unit Test Library
by Tim Allison 01 Mar '08

01 Mar '08

> > Tim Allison <tallison <at> meicompany.com> writes: > >> I ran the -E option on the Gnu compiler. Here is what I hope is the >> relevant output (at the end of the 38K resulting file!): > > [...] > >> # 20 "/usr/include/boost/test/unit_test.hpp" 2 3 4 > > [...] > >> BOOST_AUTO_TEST_CASE( test1 ) > > Looks like this is Boost 1.33. You need 1.34 to build examples in new docs. > > Gennadiy > > Where do I find examples for Boost 1.33? Based on your previous remarks, I used: http://www.boost.org/libs/test/example/unit_test_example_01.cpp It is copyrighted 2005. Tim Allison

1 0

BGL: Can bundled properties not be a struct or class name?
by Jean-Olivier Racine 29 Feb '08

29 Feb '08

Good day, I have a graph with bundled properties attached to both the vertices and edges. typedef long int ASElementID; enum ASRelationships { NoRelationship, IsContaining, IsBeside, IsAbove, IsCommentedBy, IsMadeOf }; typedef boost::adjacency_list<boost::vecS, boost::listS, boost::bidirectionalS, ASElementID, ASRelationships> Graph; When using the write_graphviz and writing my property writer, I get the error_property_not_found error. Reading post 18294 (http://thread.gmane.org/gmane.comp.lib.boost.user/18294) I see what I should probably do. My problem is that I cannot give the address of ASElementID since it is an int. Is it a limitation? Do I *have* to use a struct for it to work? Something like: struct IDType {int ID}; Thanks and have a nice weekend! Jean-Olivier

1 0

[boost test] question on exception testing
by Ovanes Markarian 29 Feb '08

29 Feb '08

Hello *, I have a question regarding the tools BOOST_<level>_THROW: Why is it required to specify the exception being thrown? What if I don't know which exception is thrown, but I know that it should be thrown? My use case is: I have a worker class which executes some functionality. I know that the functionality will throw, but I would like to test that my class will remain in the consistent state after the exception was thrown? With BOOST_<level>_THROW(class_inst.execute()) I could spare the try{}catch(...){} block and make other test assertions, but unfortunately it is not possible, due to lacking of this utility. I would propose a throw guard into the lib, which does not produce any assertion, but swallows an exception. My intention is: I have to develop inside of a huge framework. The functionality I am trying to execute will only work on a target hardware. Since I am inside a debugger and not going to test the the executed functionality itself, but check the state of my class I think such a tool might help. With Kind Regards, Ovanes

2 4

Re: [Boost-users] A request for help using the Boost Unit Test Library
by Tim Allison 29 Feb '08

29 Feb '08

> > Message: 2 > Date: Thu, 28 Feb 2008 18:44:34 +0000 (UTC) > From: Gennadiy Rozental <rogeeff(a)gmail.com> > Subject: Re: [Boost-users] A request for help using the Boost Unit > Test Library > To: boost-users(a)lists.boost.org > Message-ID: <loom.20080228T184232-270(a)post.gmane.org> > Content-Type: text/plain; charset=us-ascii > > Tim Allison <tallison <at> meicompany.com> writes: >>> What compiler are u using? Can you post compilation command? These errors > may >>> only mean that you doing something funny during compilation and compiler > do >>> not see BOOST_AUTO_TEST_CASE macro. >> I am using Gnu compiler 4.1.3-29. I have reduced the number of >> switches, so that the actual line used to build is: >> g++ -g -I../lib -I/usr/X11R6/include -I../../../source/ecs/shared >> -DOS_IS_LINUX -c FuserTests.cc > > Where is boost include? Can you generate preprocessed code? Do you see boost > test code? what do u see in place of BOOST_AUTO_TEST_CASE( test1 )? > > Gennadiy > I ran the -E option on the Gnu compiler. Here is what I hope is the relevant output (at the end of the 38K resulting file!): inline test_case* make_test_case( callback0<> const& test_func, const_string tc_name ) { return new test_case( ut_detail::normalize_test_case_name( tc_name ), test_func ); } template<typename UserTestCase> inline test_case* make_test_case( void (UserTestCase::*test_method )(), const_string tc_name, boost::shared_ptr<UserTestCase> const& user_test_case ) { return new test_case( ut_detail::normalize_test_case_name( tc_name ), ut_detail::user_tc_method_invoker<UserTestCase>( user_test_case, test_method ) ); } } } # 1 "/usr/include/boost/test/detail/enable_warnings.hpp" 1 3 4 # 246 "/usr/include/boost/test/unit_test_suite.hpp" 2 3 4 # 20 "/usr/include/boost/test/unit_test.hpp" 2 3 4 # 5 "FuserTests.cc" 2 BOOST_AUTO_TEST_CASE( test1 ) { boost::test_tools::tt_detail::check_impl( (2 == 1), boost::wrap_stringstream().ref() << boost::unit_test::literal_string( "2 == 1", sizeof( "2 == 1" ) - 1 ), boost::unit_test::literal_string( "FuserTests.cc", sizeof( "FuserTests.cc" ) - 1 ), (std::size_t)12, boost::test_tools::tt_detail::CHECK, boost::test_tools::tt_detail::CHECK_PRED, 0 ); } BOOST_AUTO_TEST_CASE( test2 ) { int i = 0; boost::test_tools::tt_detail::check_frwd( boost::test_tools::tt_detail::equal_impl_frwd(), boost::wrap_stringstream().ref() << "", boost::unit_test::literal_string( "FuserTests.cc", sizeof( "FuserTests.cc" ) - 1 ), (std::size_t)23, boost::test_tools::tt_detail::CHECK, boost::test_tools::tt_detail::CHECK_EQUAL , i, "i" , 2, "2" ); boost::test_tools::tt_detail::check_frwd( boost::test_tools::tt_detail::equal_impl_frwd(), boost::wrap_stringstream().ref() << "", boost::unit_test::literal_string( "FuserTests.cc", sizeof( "FuserTests.cc" ) - 1 ), (std::size_t)25, boost::test_tools::tt_detail::CHECK, boost::test_tools::tt_detail::CHECK_EQUAL , i, "i" , 0, "0" ); } Tim

2 1

Filter packages from a device e.g. network
by Hansi 29 Feb '08

29 Feb '08

Hello, I have a question regarding to filter packages from a device. What is the best way (high performance and not to much copies of the data) to filter data. e.g. I need a filter to make packages from a tcp stream. In this case I have first a field with the length of the package. Best regards Hansjörg

1 0

Using bind with function objects
by Olaf Peter 29 Feb '08

29 Feb '08

Hi, this test file doesn't compile, but why? $ LANG=en g++ bind.cpp -o bind && ./bind /usr/include/boost/bind.hpp: In member function 'R boost::_bi::list2<A1, A2>::operator()(boost::_bi::type<R>, F&, A&, long int) [with R = int, F = Foo, A = boost::_bi::list2<const char*&, const int&>, A1 = boost::arg<1> (*)(), A2 = boost::arg<2> (*)()]': /usr/include/boost/bind/bind_template.hpp:89: instantiated from 'typename boost::_bi::result_traits<R, F>::type boost::_bi::bind_t<R, F, L>::operator()(A1&, const A2&) [with A1 = const char*, A2 = int, R = int, F = Foo, L = boost::_bi::list2<boost::arg<1> (*)(), boost::arg<2> (*)()>]' bind.cpp:22: instantiated from here /usr/include/boost/bind.hpp:278: error: void value not ignored as it ought to be Thanks, Olaf

4 5

[Graph] A* with an implicit graph?
by Ian Downes 28 Feb '08

28 Feb '08

I'd like to do shortest path on an implicit graph - i.e. building it as I go, on examining a vertex I'd like to add the adjacent edges before continuing. Is this supported? If so, can someone point me in the right direction? If not, how could I do this in boost? The boost astar algorithm claims to support implicit graphs using a modified version of the boost bfs algorithm (nonconst_bfs.hpp). This modified version didn't exist in my distribution so I followed the comments in a document from the authors (link below) and removed the const requirements for the graph in breadth_first_search.hpp. The example (implicit example from pdf below, pg 12) compiles ok but seg faults when trying to add a vertex (vertex_t v = add_vertex(vert_prop(white_color), g); page 16 of pdf) in the visitor (some adds succeed first). Any ideas? http://www.cs.rpi.edu/~beevek/research/astar_bgl04.pdf thanks Ian

1 0

problem while using lapack binding
by Baris Gokce 28 Feb '08

28 Feb '08

Hello, I have downloaded binding of boost from http://news.tiker.net/dl/software/boost-bindings and copied them under the boost/numeric. In my code, I am using 'gesvd' function of lapack. There is no compilation error, but while linking following undefined reference to '_dgesvd' error occurs: ../../lib/libTemp.a(MatrixOperationsDefault.o): In function `_ZN16MatrixOperations6MO_svdERN5boost7numeric5ublas6matrixIdNS2_15basic_row_majorIjiEENS2_15unbounded_arrayIdSaIdEEEEESA_SA_SA_': /cygdrive/e/workspace3_2/Locomotion/BOUN/Locomotion/src/MatrixOperations.cc:159: undefined reference to `_dgesvd_' ../../lib/libTemp.a(MatrixOperationsDefault.o): In function `_ZN16MatrixOperations6MO_svdERN5boost7numeric5ublas6matrixIdNS2_15basic_row_majorIjiEENS2_15unbounded_arrayIdSaIdEEEEESA_SA_SA_': /cygdrive/c/boost/boost/numeric/ublas/exception.hpp:127: undefined reference to `_dgesvd_' collect2: ld returned 1 exit status make: *** [test.exe] Error 1 Do I have to compile boost with added binding files again? If yes, how? Best Regards...

1 0

Re: [Boost-users] A request for help using the Boost Unit Test Library
by Tim Allison 28 Feb '08

28 Feb '08

> ------------------------------ > > Message: 4 > Date: Thu, 28 Feb 2008 16:56:15 +0000 (UTC) > From: Gennadiy Rozental <rogeeff(a)gmail.com> > Subject: Re: [Boost-users] A request for help using the Boost Unit > Test Library > To: boost-users(a)lists.boost.org > Message-ID: <loom.20080228T165213-642(a)post.gmane.org> > Content-Type: text/plain; charset=us-ascii > > Tim Allison <tallison <at> meicompany.com> writes: > >> // each test module could contain no more then one ..... function defined >> // alternatively you could define init function yourself >> #define BOOST_TEST_MAIN >> #include <boost/test/unit_test.hpp> >> >> //___________________________________________________________________// >> >> // most frequently you implement test cases as a .... utomatic registration >> BOOST_AUTO_TEST_CASE( test1 ) > > You mean to say error is on above line, right? > >> The error messages are: >> FuserTests.cc:9: error: expected constructor, destructor, or type >> conversion before '(' token >> FuserTests.cc:18: error: expected constructor, destructor, or type >> conversion before '(' token > > What compiler are u using? Can you post compilation command? These errors may > only mean that you doing something funny during compilation and compiler do > not see BOOST_AUTO_TEST_CASE macro. I am using Gnu compiler 4.1.3-29. I have reduced the number of switches, so that the actual line used to build is: g++ -g -I../lib -I/usr/X11R6/include -I../../../source/ecs/shared -DOS_IS_LINUX -c FuserTests.cc Tim Allison > > Gennadiy > > > >

2 1

lexical_cast not throwing exception when converting negative value to UINT
by Kris Wragg 28 Feb '08

28 Feb '08

Dear subscribers, I am using lexical_cast to cast a string to an UINT, unfortunately, it is not throwing an exception if a value is negative. Example: std::wstring failString = L"-1"; UINT nTest = boost::lexical_cast<UINT>(failString); This does not throw a bad_lexical_cast exception, which it should, as -1 is not a valid unsigned integer. Perhaps I am using the wrong function, or using this one incorrectly, I've only just started using boost after moving to a new project that's not based on ancient legacy code! Regards, Kris Wragg ________________________________ This email and all attachments transmitted with it contain confidential information belonging to the sender. It is intended only for the addressee(s) stated above. Should you not be the intended recipient then any disclosure, copying, distribution or the taking of any action in reliance on the contents of this email is strictly prohibited. If you have received this transmission in error, please notify the sender immediately and permanently delete this message and destroy any copies. Internet communications cannot be guaranteed to be timely, secure, error-free or virus-free and Sender expressly does not warrant or guarantee that this e-mail is virus free. Recipient is advised to protect itself by taking the necessary steps to ensure that appropriate security measures are in place and to utilize a virus scanner before opening any attachment or down loading any information or documentation using links or other contents in this e-mail. You may opt out of receiving future communications from us by replying to this e-mail and replacing the subject line with UNSUBSCRIBE.

2 1

[mpl] Hierarchy Generation
by Scott Meyers 28 Feb '08

28 Feb '08

I have a set of types representing features code might have, e.g., typedef mpl::vector<ExceptionSafe, ThreadSafe, ReallyFast, GPLed> Features; These can occur in any combination as sets of features, e.g., tyepdef FeatureSet<ExceptionSafe> F1; typedef FeatureSet<ThreadSafe, ExceptionSafe> F2; typedef FeatureSet<ReallyFast, ExceptionSafe, GPLed> F3; There should be an implicit conversion from one FeatureSet to another if and only if we don't lose any features in the conversion. So converting from F1 to F2 or F3 is okay (since both F2 and F3 contain ExceptionSafe), but converting from F2 to F3 is not okay (because we'd lose ThreadSafe). Furthermore, a conversion that adds fewer features is preferable to one that adds more, so the conversion from F1 to F2 is better than the conversion from F1 to F3 (because the former adds only one feature while the latter adds two). I'm looking for a way to automatically generate an inheritance hierarchy that will give me the behavior I've described. For the above example, the hierarchy has four levels. The root is all the features, because any subset of features is permitted to be implicitly convertible to the set of all features. Root: FeatureSet<ExceptionSafe, ThreadSafe, ReallyFast, GPLed> The leaves are the individual features: Leaves: FeatureSet<ExceptionSafe> FeatureSet<ThreadSafe> FeatureSet<ReallyFast> FeatureSet<GPLed> The next level up is pairs of features: One level FeatureSet<ExceptionSafe, ThreadSafe> up from FeatureSet<ExceptionSafe, ReallyFast> leaves: FeatureSet<ExceptionSafe, GPLed> ... The next level up (which is one below the root) is triplets of features: One level FeatureSet<ExceptionSafe, ThreadSafe, ReallyFast> down from FeatureSet<ExceptionSafe, ThreadSafe, GPLed> root: FeatureSet<ExceptionSafe, ReallyFast, GPLed> ... In general, given a list of n features, I need a hierarchy n levels deep, with one class at the root, n classes at the leaves, and m-wise combinations of the n features at the intervening levels (where m varies from 2 to n-1). Fortunately, I don't need to worry about permutations of the ordering of the features, because I already have a way to canonically order them. (Not a great way, but it works, and that's a different problem.) Is there some way for me to use the MPL to automatically generate the hierarchy I need? Thanks, Scott

3 9

Re: [Boost-users] A request for help using the Boost Unit Test Library
by Tim Allison 28 Feb '08

28 Feb '08

I'm sorry, I completely missed your reply. Based on your comment that the new docs are for 1.34, The contents of FuserTest.cc are cut and pasted from example 1: // each test module could contain no more then one 'main' file with init function defined // alternatively you could define init function yourself #define BOOST_TEST_MAIN #include <boost/test/unit_test.hpp> //____________________________________________________________________________// // most frequently you implement test cases as a free functions with automatic registration BOOST_AUTO_TEST_CASE( test1 ) { // reports 'error in "test1": test 2 == 1 failed' BOOST_CHECK( 2 == 1 ); } //____________________________________________________________________________// // each test file may contain any number of test cases; each test case has to have unique name BOOST_AUTO_TEST_CASE( test2 ) { int i = 0; // reports 'error in "test2": check i == 2 failed [0 != 2]' BOOST_CHECK_EQUAL( i, 2 ); BOOST_CHECK_EQUAL( i, 0 ); } //____________________________________________________________________________// // EOF The error messages are: FuserTests.cc:9: error: expected constructor, destructor, or type conversion before '(' token FuserTests.cc:18: error: expected constructor, destructor, or type conversion before '(' token Tim Allison > > Today's Topics: > > 1. Re: A request for help using the Boost Unit Test Library > (Gennadiy Rozental) > > > ---------------------------------------------------------------------- > > Message: 1 > Date: Thu, 21 Feb 2008 20:05:42 +0000 (UTC) > From: Gennadiy Rozental <rogeeff(a)gmail.com> > Subject: Re: [Boost-users] A request for help using the Boost Unit > Test Library > To: boost-users(a)lists.boost.org > Message-ID: <loom.20080221T200258-326(a)post.gmane.org> > Content-Type: text/plain; charset=utf-8 > > Tim Allison <tallison <at> meicompany.com> writes: >> I am using boost 1.33.1-42 <at> 586 on Suse 10.2 (If it matters, my computer >> is a Latitude D510) and Gnu compiler 4.1.3-29 <at> 586 (as listed by Smart >> Package Manager). > > The new docs cover Boost.Test starting 1.34 > >> When I run the make file, it echos out the following line: >> g++ -Wno-deprecated -frepo -m32 -g -I../lib -I/usr/X11R6/include >> -I../../../source/ecs/shared -DOS_IS_LINUX -c FuserTests.cc >> >> FuserTests.cc is the file into which I have attempted to copy the >> examples. > > Would you care to share with us the file content? > >> The case I am most interested in running uses BOOST_AUTO_TEST_CASE( >> functionName ). This line produces the error: >> FuserTests.cc:16: error: expected constructor, destructor, or type >> conversion before ?(? token > > Obviosly Boost.Test headers are not included. > > Gennadiy > > > >

2 1

Boost::Function: Getting a pointer to a template function
by Hans Mull 28 Feb '08

28 Feb '08

Hi! I want to use a boost::function pointer to a template function. The template type should be determinated at the time the boost::function object is initialized. Example: template<class T> inline string addNumber(T inputNumber) { return lexical_cast<string>(inputNumber + 2); } template<class numType> inline boost::function<string(numType)> getFunction(numType number) { boost::function<string(numType)> f; f = //Which code can I use here? If I call f() I want to call addNumber<numType>(number) } ---------------------- I want to use it like that: void callFunction() { boost::function<string(double)> f; f = getFunction<unsigned long>(); for(int i = 0; i < 10000;i++) { std::cout << f(rand()) << endl; } } Thanks in advance, Hans

2 2

Looking for a boost graph algorithm
by Christian Sturz 28 Feb '08

28 Feb '08

Hi, I'm looking for an appropriate Boost Graph algorithm for the following problem: my graph (based on adjacency_list) represents the control-flow graph of a program where each vertex is a statement and the control-flow relations between the statements are modeled by edges. Now I'm looking for an algorithm that finds all vertices that can be reached from a defined statement. So, what I need is a back-traversal of the graph, from the given vertex to all predecessors and so forth. I didn't find any algorithm in the Boost Graph Library. The "strong_components" algorithm might be OK since it assumes a directed graph (as is mine) but the algorithm seems to operate in the direction of the directed edges. Obviously I need the the opposite direction. Do you know any approach/algorithm that might solve my problem? Regards, Chris -- Psssst! Schon vom neuen GMX MultiMessenger gehört? Der kann`s mit allen: http://www.gmx.net/de/go/multimessenger

2 3

Problems connecting to boost repository
by Lang Stefan 28 Feb '08

28 Feb '08

Hello List, I am trying to download (parts of) the current version of boost as described in http://svn.boost.org/trac/boost/wiki/BoostSubversion. I do have a subversion client installed (version 1.4.2) as well as TortoiseSVN for windows. However, so far I have been unable to download any file(s) from the repository at http://svn.boost.org. The IT staff advised me to use a proxy server and a specific port to establish that sort of internet connection, and I found and set the appropriate options for TortoiseSVN. Afterwards, when I tried to use TortoiseSVN for downloading from the boost repository, some administrative data was transferred, but still no actual files. The error messages I got were: REPORT request failed on '/svn/boost/!svn/vcc/default' REPORT of '/svn/boost/!svn/vcc/default': 400 Bad Request (http://svn.boost.org) Can anyone tell me what's wrong, or how to access the boost svn repository properly? I would try out the command line client directly, but am not sure how to set the proxy server and port. Mit freundlichen Grüssen / Kind regards Stefan Lang

2 1

Problems compiling programs with graphviz_read()
by Vivek Menon 28 Feb '08

28 Feb '08

Hello all,I am a newbie to Boost Graph Libraries. I am working on a Mac OSX (Intel) platform. I have compiled the libraries using the bjam executable from boost_1_34_1/tools/jam/src/bin.macosxx86/bjam This resulted in formation of libraries in bin.v2/libs/... Now I would like to compile the edge-connectivity.cpp file in boost_1_34_1/libs/graph/examples directory with the following command: g++ -I$BOOST edge-connectivity.cpp -o test1 $BOOST/bin.v2/libs/graph/build/darwin/debug/link-static/libboost_graph-d-1_34_1.a This gives me the following errors: Undefined symbols: "boost::read_graphviz(std::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, boost::subgraph<boost::adjacency_list<boost::vecS, boost::vecS, boost::undirectedS, boost::property<boost::vertex_attribute_t, std::map<std::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::less<std::basic_string<char, std::char_traits<char>, std::allocator<char> > >, std::allocator<std::pair<std::basic_string<char, std::char_traits<char>, std::allocator<char> > const, std::basic_string<char, std::char_traits<char>, std::allocator<char> > > > >, boost::no_property>, boost::property<boost::edge_attribute_t, std::map<std::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::less<std::basic_string<char, std::char_traits<char>, std::allocator<char> > >, std::allocator<std::pair<std::basic_string<char, std::char_traits<char>, std::allocator<char> > const, std::basic_string<char, std::char_traits<char>, std::allocator<char> > > > >, boost::property<boost::edge_index_t, int, boost::no_property> >, boost::property<boost::graph_graph_attribute_t, std::map<std::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::less<std::basic_string<char, std::char_traits<char>, std::allocator<char> > >, std::allocator<std::pair<std::basic_string<char, std::char_traits<char>, std::allocator<char> > const, std::basic_string<char, std::char_traits<char>, std::allocator<char> > > > >, boost::property<boost::graph_vertex_attribute_t, std::map<std::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::less<std::basic_string<char, std::char_traits<char>, std::allocator<char> > >, std::allocator<std::pair<std::basic_string<char, std::char_traits<char>, std::allocator<char> > const, std::basic_string<char, std::char_traits<char>, std::allocator<char> > > > >, boost::property<boost::graph_edge_attribute_t, std::map<std::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::less<std::basic_string<char, std::char_traits<char>, std::allocator<char> > >, std::allocator<std::pair<std::basic_string<char, std::char_traits<char>, std::allocator<char> > const, std::basic_string<char, std::char_traits<char>, std::allocator<char> > > > >, boost::property<boost::graph_name_t, std::basic_string<char, std::char_traits<char>, std::allocator<char> >, boost::no_property> > > >, boost::listS> >&)", referenced from: _main in cctmckT6.o ld: symbol(s) not found collect2: ld returned 1 exit status Can some one explain what's going on? I have tried using -L option as well with no success: g++ -I$BOOST edge-connectivity.cpp -o test1 -L$BOOST/boost_1_34_1/bin.v2/libs/graph/build/darwin/debug/link-static/ -lboost_graph-d-1_34_1.a The examples that do not use graphviz_read() are easily compiled. I would also like to know if someone uses a jam file that lists all the dependencies for the example files. This way I can compile all the example files in the /libs/graph/examples in one shot with no linking problems. Thanks, Vivek

1 0

[enable_if] Compiler Error with MSVC 7.1
by Christian Henning 28 Feb '08

28 Feb '08

Hi there, I'm trying to compile the following code using MSVC 7.1 but it fails. Using MSVC 8 works which leaves me with the impression that it might be a compiler bug. #include <boost/utility/enable_if.hpp> #include <boost/mpl/bool.hpp> using namespace boost; template < typename A > struct is_A : public mpl::true_ {}; template < typename B > struct is_B : public mpl::true_ {}; template< typename A > struct enable_if_A : public enable_if< typename is_A< A >::type > {}; template< typename B > struct enable_if_B : public enable_if< typename is_B< B >::type > {}; template < typename One > inline void foo( typename enable_if_A< One >::type* ptr = 0 ) {} template < typename One > inline void foo( typename enable_if_B< One >::type* ptr = 0 ) {} int _tmain(int argc, _TCHAR* argv[]) { return 0; } After a while of tinkering I got it to compile when inheriting the enable_if helper classes from enable_if<...>::type instead of from enable_if<...>. The following works: #include <boost/utility/enable_if.hpp> #include <boost/mpl/bool.hpp> using namespace boost; template < typename A > struct is_A : public mpl::true_ {}; template < typename B > struct is_B : public mpl::true_ {}; template< typename A > struct enable_if_A : public enable_if< typename is_A< A >::type >::type {}; template< typename B > struct enable_if_B : public enable_if< typename is_B< B >::type >::type {}; template < typename One > inline void foo( typename enable_if_A< One >* ptr = 0 ) {} template < typename One > inline void foo( typename enable_if_B< One >* ptr = 0 ) {} int _tmain(int argc, _TCHAR* argv[]) { return 0; } Is that a known issue with MSVC 7.1? Or is it always better to inherit from enable_if<...>::type? Regards, Christian

1 0

The Graph Library Applied to Trees
by Charles Brockman 28 Feb '08

28 Feb '08

I'm considering using the Boost Graph Library to represent binary trees in my application. This will be a planar acyclic directed graph with an adjacency-list representation. The vertices will have properties bundled in an exterior class. A. While building the tree I need to track the depth of each branch and the greatest depth of any branch, that is, the distance from the source vertex to the farthest target vertex. For that I plan to use the distance_recorder. Is that reasonable? Are there pitfalls? Is there a better way? B. I'll need to calculate the total length of the tree by which I mean the sum of all edges. I plan to do a depth-first search and increment a counter with a visitor event point such as discover_vertex. Is that the preferred method? C. On occasion I'll have to swap vertices between two independent trees. My plan is to just swap the bundled properties of each vertex. At first I considered the remove_vertex function but the edges connected to a vertex must be removed first with clear_vertex so before I remove the vertex I'll have to record the source and the target vertices. Then I'll have to add the replacement vertex into the correct position. I can imagine that blowing up in my face. I would appreciate any guidance on this problem. -- Charles Brockman

1 0

gzip_compressor cannot be flushed and closed properly
by Mengda Wu 27 Feb '08

27 Feb '08

Hi, I have a code block trying to output a series of gzip files. I wish to have all the files flushed after the block. But the files can only be flushed after exiting the whole program. The files still have zero size after I try to close them. I am using std::vector to store all the pointers to boost filtering_ostream. Can you help? //Open iostreams char filename[20]; std::vector<boost::iostreams::filtering_ostream *> os_vector; for(i=0; i<4; i++) { sprintf(filename, "file_%d.gz", i ); std::ofstream *of = new std::ofstream(filename, std::ios_base::binary); boost::iostreams::filtering_ostream* os = new boost::iostreams::filtering_ostream; os->push(boost::iostreams::gzip_compressor()); os->push(*of); os_vector->push_back(os); } //Output something for(i=0; i<4; i++) { boost::iostreams::filtering_ostream* os = os_vector[i]; os<<"Output something here" << std::endl; } //Close streams for(i=0; i<4; i++) { boost::iostreams::filtering_ostream* os = os_vector[i]; os->strict_sync(); os->pop(); os->reset(); } Thanks, Mengda

3 6

[interprocess] Many-to-many communication
by Boris 27 Feb '08

27 Feb '08

I have an application which needs to communicate with several other applications on the same machine. I thought to use message queues but quickly ran into a problem: Either I create as many threads as there are message queues which all block and wait for incoming data or I constantly poll all message queues. The later should be easily possible with try_receive(). However it doesn't sound like a good idea to poll continuously? And I assume integrating the message queue somehow into Boost.Asio is not something easily possible? Thus for many-to-many communication shared memory is probably the way to go? Boris

3 2

BGL: Bug in depth_first_search
by Christian Sturz 27 Feb '08

27 Feb '08

Hi, as it seems I've found a bug in the depth_first_search algorithm. I want to use it with my own ColorMap and a start vertex. Here is the code taken from http://www.boost.org/libs/graph/example/dfs-example.cpp with some slight modifications: ------------------------------------ include <boost/graph/adjacency_list.hpp> #include <boost/graph/depth_first_search.hpp> #include <boost/pending/integer_range.hpp> #include <boost/pending/indirect_cmp.hpp> #include <iostream> using namespace boost; template < typename TimeMap > class dfs_time_visitor:public default_dfs_visitor { typedef typename property_traits < TimeMap >::value_type T; public: dfs_time_visitor(TimeMap dmap, TimeMap fmap, T & t) : m_dtimemap(dmap), m_ftimemap(fmap), m_time(t) { } template < typename Vertex, typename Graph > void discover_vertex(Vertex u, const Graph & g) const { put(m_dtimemap, u, m_time++); } template < typename Vertex, typename Graph > void finish_vertex(Vertex u, const Graph & g) const { put(m_ftimemap, u, m_time++); } TimeMap m_dtimemap; TimeMap m_ftimemap; T & m_time; }; int main() { // Select the graph type we wish to use typedef adjacency_list < vecS, vecS, directedS > graph_t; typedef graph_traits < graph_t >::vertices_size_type size_type; // Set up the vertex names enum { u, v, w, x, y, z, N }; char name[] = { 'u', 'v', 'w', 'x', 'y', 'z' }; // Specify the edges in the graph typedef std::pair < int, int >E; E edge_array[] = { E(u, v), E(u, x), E(x, v), E(y, x), E(v, y), E(w, y), E(w, z), E(z, z) }; #if defined(BOOST_MSVC) && BOOST_MSVC <= 1300 graph_t g(N); for (std::size_t j = 0; j < sizeof(edge_array) / sizeof(E); ++j) add_edge(edge_array[j].first, edge_array[j].second, g); #else graph_t g(edge_array, edge_array + sizeof(edge_array) / sizeof(E), N); #endif // Typedefs typedef boost::graph_traits < graph_t >::vertex_descriptor Vertex; Vertex root = source( edge_array[2],g ); typedef size_type* Iiter; // discover time and finish time properties std::vector < size_type > dtime(num_vertices(g)); std::vector < size_type > ftime(num_vertices(g)); size_type t = 0; dfs_time_visitor < size_type * >vis(&dtime[0], &ftime[0], t); std::vector<default_color_type> color(num_vertices( g ) ); depth_first_search(g, visitor(vis), &color[0], root ); return 0; } ------------------------------------ So, I've added a ColorMap and a start vertex as 3. and 4. argument to "depth_first_search". This code can't be compiled with the gcc on a Debian Etch machine. These are the compiler errors: dfs.cc: In function `int main()': dfs.cc:47: warning: unused variable 'name' /usr/include/boost/graph/depth_first_search.hpp: In function `void boost::depth_first_search(const VertexListGraph&, DFSVisitor, ColorMap, typename boost::graph_traits<G>::vertex_descriptor) [with VertexListGraph = main()::graph_t, DFSVisitor = boost::bgl_named_params<dfs_time_visitor<main::size_type*>, boost::graph_visitor_t, boost::no_property>, ColorMap = boost::default_color_type*]': dfs.cc:76: instantiated from here /usr/include/boost/graph/depth_first_search.hpp:197: error: 'struct boost::bgl_named_params<dfs_time_visitor<main::size_type*>, boost::graph_visitor_t, boost::no_property>' has no member named 'initialize_vertex' dfs.cc:76: instantiated from here /usr/include/boost/graph/depth_first_search.hpp:200: error: 'struct boost::bgl_named_params<dfs_time_visitor<main::size_type*>, boost::graph_visitor_t, boost::no_property>' has no member named 'start_vertex' dfs.cc:76: instantiated from here /usr/include/boost/graph/depth_first_search.hpp:207: error: 'struct boost::bgl_named_params<dfs_time_visitor<main::size_type*>, boost::graph_visitor_t, boost::no_property>' has no member named 'start_vertex' /usr/include/boost/graph/depth_first_search.hpp: In function `void boost::detail::depth_first_visit_impl(const IncidenceGraph&, typename boost::graph_traits<G>::vertex_descriptor, DFSVisitor&, ColorMap, TerminatorFunc) [with IncidenceGraph = boost::adjacency_list<boost::vecS, boost::vecS, boost::directedS, boost::no_property, boost::no_property, boost::no_property, boost::listS>, DFSVisitor = boost::bgl_named_params<dfs_time_visitor<main::size_type*>, boost::graph_visitor_t, boost::no_property>, ColorMap = boost::default_color_type*, TerminatorFunc = boost::detail::nontruth2]': /usr/include/boost/graph/depth_first_search.hpp:201: instantiated from `void boost::depth_first_search(const VertexListGraph&, DFSVisitor, ColorMap, typename boost::graph_traits<G>::vertex_descriptor) [with VertexListGraph = main()::graph_t, DFSVisitor = boost::bgl_named_params<dfs_time_visitor<main::size_type*>, boost::graph_visitor_t, boost::no_property>, ColorMap = boost::default_color_type*]' dfs.cc:76: instantiated from here /usr/include/boost/graph/depth_first_search.hpp:103: error: 'struct boost::bgl_named_params<dfs_time_visitor<main::size_type*>, boost::graph_visitor_t, boost::no_property>' has no member named 'discover_vertex' /usr/include/boost/graph/depth_first_search.hpp:201: instantiated from `void boost::depth_first_search(const VertexListGraph&, DFSVisitor, ColorMap, typename boost::graph_traits<G>::vertex_descriptor) [with VertexListGraph = main()::graph_t, DFSVisitor = boost::bgl_named_params<dfs_time_visitor<main::size_type*>, boost::graph_visitor_t, boost::no_property>, ColorMap = boost::default_color_type*]' dfs.cc:76: instantiated from here /usr/include/boost/graph/depth_first_search.hpp:120: error: 'struct boost::bgl_named_params<dfs_time_visitor<main::size_type*>, boost::graph_visitor_t, boost::no_property>' has no member named 'examine_edge' /usr/include/boost/graph/depth_first_search.hpp:201: instantiated from `void boost::depth_first_search(const VertexListGraph&, DFSVisitor, ColorMap, typename boost::graph_traits<G>::vertex_descriptor) [with VertexListGraph = main()::graph_t, DFSVisitor = boost::bgl_named_params<dfs_time_visitor<main::size_type*>, boost::graph_visitor_t, boost::no_property>, ColorMap = boost::default_color_type*]' dfs.cc:76: instantiated from here /usr/include/boost/graph/depth_first_search.hpp:123: error: 'struct boost::bgl_named_params<dfs_time_visitor<main::size_type*>, boost::graph_visitor_t, boost::no_property>' has no member named 'tree_edge' /usr/include/boost/graph/depth_first_search.hpp:201: instantiated from `void boost::depth_first_search(const VertexListGraph&, DFSVisitor, ColorMap, typename boost::graph_traits<G>::vertex_descriptor) [with VertexListGraph = main()::graph_t, DFSVisitor = boost::bgl_named_params<dfs_time_visitor<main::size_type*>, boost::graph_visitor_t, boost::no_property>, ColorMap = boost::default_color_type*]' dfs.cc:76: instantiated from here /usr/include/boost/graph/depth_first_search.hpp:127: error: 'struct boost::bgl_named_params<dfs_time_visitor<main::size_type*>, boost::graph_visitor_t, boost::no_property>' has no member named 'discover_vertex' /usr/include/boost/graph/depth_first_search.hpp:201: instantiated from `void boost::depth_first_search(const VertexListGraph&, DFSVisitor, ColorMap, typename boost::graph_traits<G>::vertex_descriptor) [with VertexListGraph = main()::graph_t, DFSVisitor = boost::bgl_named_params<dfs_time_visitor<main::size_type*>, boost::graph_visitor_t, boost::no_property>, ColorMap = boost::default_color_type*]' dfs.cc:76: instantiated from here /usr/include/boost/graph/depth_first_search.hpp:133: error: 'struct boost::bgl_named_params<dfs_time_visitor<main::size_type*>, boost::graph_visitor_t, boost::no_property>' has no member named 'back_edge' /usr/include/boost/graph/depth_first_search.hpp:201: instantiated from `void boost::depth_first_search(const VertexListGraph&, DFSVisitor, ColorMap, typename boost::graph_traits<G>::vertex_descriptor) [with VertexListGraph = main()::graph_t, DFSVisitor = boost::bgl_named_params<dfs_time_visitor<main::size_type*>, boost::graph_visitor_t, boost::no_property>, ColorMap = boost::default_color_type*]' dfs.cc:76: instantiated from here /usr/include/boost/graph/depth_first_search.hpp:136: error: 'struct boost::bgl_named_params<dfs_time_visitor<main::size_type*>, boost::graph_visitor_t, boost::no_property>' has no member named 'forward_or_cross_edge' /usr/include/boost/graph/depth_first_search.hpp:201: instantiated from `void boost::depth_first_search(const VertexListGraph&, DFSVisitor, ColorMap, typename boost::graph_traits<G>::vertex_descriptor) [with VertexListGraph = main()::graph_t, DFSVisitor = boost::bgl_named_params<dfs_time_visitor<main::size_type*>, boost::graph_visitor_t, boost::no_property>, ColorMap = boost::default_color_type*]' dfs.cc:76: instantiated from here /usr/include/boost/graph/depth_first_search.hpp:141: error: 'struct boost::bgl_named_params<dfs_time_visitor<main::size_type*>, boost::graph_visitor_t, boost::no_property>' has no member named 'finish_vertex' /usr/include/boost/graph/depth_first_search.hpp: In member function `void boost::DFSVisitorConcept<Visitor, Graph>::constraints() [with Visitor = boost::bgl_named_params<dfs_time_visitor<main::size_type*>, boost::graph_visitor_t, boost::no_property>, Graph = main()::graph_t]': /usr/include/boost/concept_check.hpp:48: instantiated from `void boost::function_requires(boost::mpl::identity<T>*) [with Concept = boost::DFSVisitorConcept<boost::bgl_named_params<dfs_time_visitor<main::size_type*>, boost::graph_visitor_t, boost::no_property>, main()::graph_t>]' /usr/include/boost/graph/depth_first_search.hpp:191: instantiated from `void boost::depth_first_search(const VertexListGraph&, DFSVisitor, ColorMap, typename boost::graph_traits<G>::vertex_descriptor) [with VertexListGraph = main()::graph_t, DFSVisitor = boost::bgl_named_params<dfs_time_visitor<main::size_type*>, boost::graph_visitor_t, boost::no_property>, ColorMap = boost::default_color_type*]' dfs.cc:76: instantiated from here /usr/include/boost/graph/depth_first_search.hpp:36: error: 'struct boost::bgl_named_params<dfs_time_visitor<main::size_type*>, boost::graph_visitor_t, boost::no_property>' has no member named 'initialize_vertex' /usr/include/boost/graph/depth_first_search.hpp:37: error: 'struct boost::bgl_named_params<dfs_time_visitor<main::size_type*>, boost::graph_visitor_t, boost::no_property>' has no member named 'start_vertex' /usr/include/boost/graph/depth_first_search.hpp:38: error: 'struct boost::bgl_named_params<dfs_time_visitor<main::size_type*>, boost::graph_visitor_t, boost::no_property>' has no member named 'discover_vertex' /usr/include/boost/graph/depth_first_search.hpp:39: error: 'struct boost::bgl_named_params<dfs_time_visitor<main::size_type*>, boost::graph_visitor_t, boost::no_property>' has no member named 'examine_edge' /usr/include/boost/graph/depth_first_search.hpp:40: error: 'struct boost::bgl_named_params<dfs_time_visitor<main::size_type*>, boost::graph_visitor_t, boost::no_property>' has no member named 'tree_edge' /usr/include/boost/graph/depth_first_search.hpp:41: error: 'struct boost::bgl_named_params<dfs_time_visitor<main::size_type*>, boost::graph_visitor_t, boost::no_property>' has no member named 'back_edge' /usr/include/boost/graph/depth_first_search.hpp:42: error: 'struct boost::bgl_named_params<dfs_time_visitor<main::size_type*>, boost::graph_visitor_t, boost::no_property>' has no member named 'forward_or_cross_edge' /usr/include/boost/graph/depth_first_search.hpp:43: error: 'struct boost::bgl_named_params<dfs_time_visitor<main::size_type*>, boost::graph_visitor_t, boost::no_property>' has no member named 'finish_vertex' So, there seem to be a bug in the algorithm. When calling it as defined in the exampl file from the Boost side without a ColorMap and a vertex, it works fine. Could you verify this and tell me if you can reproduce it? Thank you. Regards, Christian -- GMX startet ShortView.de. Hier findest Du Leute mit Deinen Interessen! Jetzt dabei sein: http://www.shortview.de/?mc=sv_ext_mf@gmx

1 0

[serialization] class template with virtual base
by Richard Crossley 27 Feb '08

27 Feb '08

Hi, I've been searching the archives but am not finding an answer... Is there any way to serialise D<> via a pointer to its base B when the types used to parameterise D<> are not known in advance? i.e. How do I register D<> with the serialisation library? For Example... struct B{ virtual void fn() =0; }; template<typename T> struct D : B{ virtual void fn(){}; }; B* b = new D<char>(); .... anArchive & b; Thanks, Richard.

2 3

Re: [Boost-users] boost::crc32 - visual studio crt - bug?!
by gfsdgadas＠Safe-mail.net 27 Feb '08

27 Feb '08

I do not like the idea of changing my compile flags (I really do like that option for my debug builds!) just because of boost crc32. What about changing the header the request way, marking all casts using &0xFF? I do not see any reason for those safe marks not being introduces into boost crc.

4 4

Boost PP lib
by Hicham Mouline 27 Feb '08

27 Feb '08

hi, in an ealier post, i was helped to write explicit template instantiations as: #define FINH1 (EC)(EP)(AC)(AP) #define FINH2 (C)(D)(E)(F)(G) #define FINH3 (2)(3) #define M(r, product) template class Tree<BOOST_PP_SEQ_ENUM(product)>; BOOST_PP_SEQ_FOR_EACH_PRODUCT(M, (FINH1)(FINH2)(FINH3)) That nicely generated: template class Tree<EC,C,2>; template class Tree<EC,C,3>; template class Tree<EC,D,2>; template class Tree<EC,D,3>; ... what i need now is coupling FINH2 and FINH3 (i don't want their crossproduct) as in: #define FINH1 (EC)(EP)(AC)(AP) #define FINH2 (C,2)(D,2)(E,3)(F,3)(G,3) #define M(r, product) template class Tree<BOOST_PP_SEQ_ENUM(product)>; BOOST_PP_SEQ_FOR_EACH_PRODUCT(M, (FINH1)(FINH2)) But that does't work because the commas in FINH2 are taken as macro arguments which fails... ideas are appreciated... regards,

2 1

Boost Interprocess docs
by Wolfram Stering 27 Feb '08

27 Feb '08

Hello List, if I wanted to read Boost.Interprocess documentation, where would I look for it? I want to evaluate Boost.Interprocess for use in a setuid application with requires a GUI part using a backend (setuid) and a frontend (gui) process. Boost.Interprocess is scheduled to be included in release 1.35, if I remember correctly. thanks for your help, best regards, -wolfi -- Wolfram Stering (Entwicklung) HALE electronic GmbH Eugen-Müller-Straße 18, 5020 Salzburg, Austria Tel: +43 (662) 439011 550 Fax: +43 (662) 439011 9 http://www.hale.at/ Firmenbuchnummer: FN 66801m HG Salzburg -- Scanned by MailScanner.

1 0

circular_buffer iterator problem
by abir basak 27 Feb '08

27 Feb '08

Hi, I have a boost svn trunk checkout (latest one). It seems boost circular_buffer has a iterator problem for converting from non const to const iterator, on visual studio 7.1 (2003) release build only. Can someone check for it ? (It used to work on boost 1.34) for eg, using namespace boost::assign; typedef boost::circular_buffer<int> CB; std::vector<int> v; v+=1,2,3,4,5,6,7,8,9,10,11,12; CB buf(10,v.begin(),v.end()); CB::iterator it = buf.begin(); CB::const_iterator it1 = buf.begin();/// This line doesn't compile on release build only. a second (and little different) question, why boost::assign doesn't work with circular_buffer directly which has 2 iterator construct like most other STL containers ? thanks abir -- Abir Basak, Member IEEE Software Engineer, Read Ink Technologies B. Tech, IIT Kharagpur email: abir(a)abirbasak.com homepage: www.abirbasak.com

1 0

boost::lambda::bind internal object storage
by Zian Smith 27 Feb '08

27 Feb '08

Hello, Can someone explain why I see 5 constructor calls here, but only 4 destructor calls? Is the bind function neglecting to destroy one of the internal copies made or am I missing something? I understand that if I use the "var" function, the problem is eliminated, but I'd still like to figure this out. I'm using boost 1.33.1 -Thanks! //================================= #include <iostream> #include <vector> #include <algorithm> #include "boost/lambda/bind.hpp" using namespace std; using namespace boost::lambda; struct woot { woot() { cout << "woot constr\n"; } woot(const woot& o) { cout << "woot copy-constr\n"; } ~woot() { cout << "woot destr\n"; } }; void do_nothing(const woot& w) { } int main() { woot w; vector<char> v; for_each(v.begin(), v.end(), bind(do_nothing, w)); } //================================= Output: ------- woot constr woot copy-constr woot copy-constr woot copy-constr woot destr woot destr woot copy-constr woot destr woot destr ____________________________________________________________________________________ Never miss a thing. Make Yahoo your home page. http://www.yahoo.com/r/hs

1 0

ASIO ssl::stream read/write_some question.
by Derrick Hathaway 27 Feb '08

27 Feb '08

Could someone please shed some light on ASIO SSL reads and writes for me? The names of the read and write functions for the ssl::stream (read_some and write_some) imply to me that the buffer(s) may not be completely filled/sent with a read/write. Do I need to have some intermediate handler that will begin another async_read/write_some until my buffer is filled/flushed? The documentation doesn't seem to indicate that this is necessary. It makes sense from the example that the reads may not fill the provided buffer before invoking the handler. So it may be necessary to call read_some multiple times until the expected number of bytes is received. But is this also the case with writes, and if so, how do I manage multiple writes if I provide a vector of buffers to be written with and async_write_some? Thanks in advance. Derrick

2 2

boost.serialization question
by Aljaz 27 Feb '08

27 Feb '08

Hello I use boost.serialization for data transfer betwen my server and client applications. The example of serialized data looks like: " 3b22 serialization::archive 4 0 1 2 7 request 1 0 0 1 0 1 2 1 " I've been wondering if there is any way to hide serialization:.archive in the data, maybe replace it with something like 'sa' .. Thanks a lot for help Best regards Aljaz

2 1

Re: [Boost-users] Boost.Test unit testing examples: stale links
by Tim Allison 27 Feb '08

27 Feb '08

To see example 1, use: http://www.boost.org/libs/test/example/unit_test_example_01.cpp The obvious changes are needed for the other examples. (Someone else found these links and posted them at the almost unused forums on SourceForge.) From what I have been able to figure out, the most up to date examples and documentation are found at: http://www.patmedia.net/~rogeeff/html/utf/user-guide.html. I would like to know if you can compile the examples. I can not. I sent an email a week or so ago requesting help compiling the examples with g++. So far, I have not had a response. Tim Allison > ------------------------------ > > Message: 6 > Date: Tue, 26 Feb 2008 22:33:16 +0100 > From: Bram Kuijper <a.l.w.kuijper(a)rug.nl> > Subject: [Boost-users] Boost.Test unit testing examples: stale links > To: boost-users(a)lists.boost.org > Message-ID: <47C4859C.5000302(a)rug.nl> > Content-Type: text/plain; charset=ISO-8859-1; format=flowed > > Hi all, > > I don't know if this is the right place for it, but I wanted to report a > number of stale links on the Boost.Test website. It regards the examples > of the Unit Testing Framework, example 1 to 5 on page > http://www.boost.org/libs/test/doc/examples/index.html > : > > http://www.boost.org/libs/test/doc/examples/unit_test_example1.html > http://www.boost.org/libs/test/doc/examples/unit_test_example2.html > http://www.boost.org/libs/test/doc/examples/unit_test_example3.html > http://www.boost.org/libs/test/doc/examples/unit_test_example4.html > http://www.boost.org/libs/test/doc/examples/unit_test_example5.html > > since I really would like to see some examples of this testing > framework, I wonder if it is possible to have these examples back online? > > thanks, > > Bram Kuijper > >

2 1

[IOStreams] Inconsistency b/w docs and code of stream()
by Nitay Joffe 27 Feb '08

27 Feb '08

Hey guys, I'm working with the boost::iostreams library to create a stream object from a Device class I've implemented. I'm trying to call the constructor that allows setting of the buffer size and it is failing to compile on 64 bit. I used the documenation at http://www.boost.org/libs/iostreams/doc/guide/generic_streams.html#stream The docs say the constructor is of the form: stream(const T& t, std::streamsize buffer_size = default_value, std::streamsize pback_size = default_value); My code to create the stream looks like: MyDevice myDevice(...) MyStream(myDevice, (std::streamsize)500); On a 32 bit machine this works fine, but on a 64 bit one it does not compile. The parameters I'm passing in are not matching up with a constructor in the stream class so it tries to call MyDevice with the wrong set of parameters. I believe the culprit is the file iostreams/detail/push_params.hpp, which defines the buffer and pback size using: #define BOOST_IOSTREAMS_PUSH_PARAMS() \ , int buffer_size = -1 , int pback_size = -1 \ /**/ This macro is then used by stream and friends in their constructor declarations. On 64 bit machines std::streamsize is a long int, so the constructor parameters do not match up and compilation fails. Thanks, -n

2 1

serialization: classes with pimpls
by Noah Roberts 26 Feb '08

26 Feb '08

Any ideas? The only options I can come up with are: 1) break the pimpl idiom by removing the opaqueness or implementing private getter/setter functions for the serialize template. Still get the exception safety of using the pimpl but lose the compile-time decoupling. 2) only implement the serialize function for certain types of archives. This one seems the most interesting actually. 3) write string writers/parsers for serialize to use, which pretty much destroys the whole point of using boost::serialization in the first place. Maybe I'm missing a good option?

1 0

iterator_traits and shared_ptr
by Patrik Jonsson 26 Feb '08

26 Feb '08

Hi, I'm trying to transparently get the type of a shared_ptr or a pod pointer using iterator_traits::value_type. This fails on shared_ptr<T> with the following message: /usr/lib/gcc/x86_64-redhat-linux/4.1.2/../../../../include/c++/4.1.2/bits/stl_iterator_base_types.h: In instantiation of 'std::iterator_traits<boost::shared_ptr<int> >': crap.cc:5: instantiated from here /usr/lib/gcc/x86_64-redhat-linux/4.1.2/../../../../include/c++/4.1.2/bits/stl_iterator_base_types.h:129: error: no type named 'iterator_category' in 'class boost::shared_ptr<int>' /usr/lib/gcc/x86_64-redhat-linux/4.1.2/../../../../include/c++/4.1.2/bits/stl_iterator_base_types.h:131: error: no type named 'difference_type' in 'class boost::shared_ptr<int>' It seems shared_ptr needs difference_type and iterator_category for this to work. Is there another way to accomplish what I'm trying to do? Regards, /Patrik

2 1

[variant] make_recursive_variant containing a variant ?
by Alban Linard 26 Feb '08

26 Feb '08

Hi, i am trying to build a recursive variant containing another variant, as described in the following source code: #include <cstdlib> #include <vector> #include <string> #include <boost/variant.hpp> typedef ::boost::variant < int , bool > first_variant; typedef ::boost::make_recursive_variant < first_variant , ::std::string , ::std::vector< ::boost::recursive_variant_ > >::type second_variant; int main() { first_variant f(true); second_variant s(f); return EXIT_SUCCESS; } It does not compile (g++-4.0 and g++-4.3). It seems that the "first_variant" part of second_variant is not used. Why this behavior ? Is there a workaround ? Thanks, Alban Linard

2 5

Boost.Test unit testing examples: stale links
by Bram Kuijper 26 Feb '08

26 Feb '08

Hi all, I don't know if this is the right place for it, but I wanted to report a number of stale links on the Boost.Test website. It regards the examples of the Unit Testing Framework, example 1 to 5 on page http://www.boost.org/libs/test/doc/examples/index.html : http://www.boost.org/libs/test/doc/examples/unit_test_example1.html http://www.boost.org/libs/test/doc/examples/unit_test_example2.html http://www.boost.org/libs/test/doc/examples/unit_test_example3.html http://www.boost.org/libs/test/doc/examples/unit_test_example4.html http://www.boost.org/libs/test/doc/examples/unit_test_example5.html since I really would like to see some examples of this testing framework, I wonder if it is possible to have these examples back online? thanks, Bram Kuijper

1 0

Re: [Boost-users] boost exceptions on MacOS 10.5
by David Thulson 26 Feb '08

26 Feb '08

Hello, I have had similar problems in OSX 10.5 where exceptions are not caught, but with a different project. We are using CMake. I'm not sure how familiar you are with CMake, but it is a build system that allows you to generate projects for "native" development tools. In OSX, you can create Xcode projects or makefiles. If I use Xcode to build the project, I cannot catch exceptions consistently. If I use makefiles, it works fine. You can see the discussion on the CMake list here: http://www.cmake.org/pipermail/cmake/2008-January/019091.html The compiler flags that are used are significantly different, so that seems like a likely problem. Unfortunately, I have not had time to finish tracking the issue down. I'm not really familiar with building boost so maybe this is obvious, but how are you building it? What compiler flags are you using? David Thulson

3 2

BGL: ColorMap for depth_first_search
by Christian Sturz 26 Feb '08

26 Feb '08

Hi, I've an urgent question on a BGL algorithm. I want to use the depth_first_search algorithm from the Boost Graph Library. However, I've problems with the definition of my own ColorMap. To illustrate this, here is an example (taken from http://www.boost.org/libs/graph/example/dfs-example.cpp) int main() { // Select the graph type we wish to use typedef adjacency_list < vecS, vecS, directedS > graph_t; typedef graph_traits < graph_t >::vertices_size_type size_type; // Set up the vertex names enum { u, v, w, x, y, z, N }; char name[] = { 'u', 'v', 'w', 'x', 'y', 'z' }; // Specify the edges in the graph typedef std::pair < int, int >E; E edge_array[] = { E(u, v), E(u, x), E(x, v), E(y, x), E(v, y), E(w, y), E(w, z), E(z, z) }; graph_t g(edge_array, edge_array + sizeof(edge_array) / sizeof(E), N); // Typedefs typedef boost::graph_traits < graph_t >::vertex_descriptor Vertex; typedef size_type* Iiter; // discover time and finish time properties std::vector < size_type > dtime(num_vertices(g)); std::vector < size_type > ftime(num_vertices(g)); size_type t = 0; dfs_time_visitor < size_type * >vis(&dtime[0], &ftime[0], t); // Some arbitrary vertex as root (start point of DFS) Vertex root = source( edge_array[2],g ); depth_first_search(g, visitor(vis), myColorMap, root ); return 0; } My problem is the definition of "myColorMap" which must be a model of Read/Write Property Map. However, I have no idea how to define a working ColorMap. It would be fully sufficient to take the default color map which is used when DFS is invoked with "depth_first_search(g, visitor(vis));" Do you have an idea how I define a ColorMap that I can use for "myColorMap"? Any help is appreciated. Thank you. Christian -- Psst! Geheimtipp: Online Games kostenlos spielen bei den GMX Free Games! http://games.entertainment.web.de/de/entertainment/games/free

1 0

[noob] I have pointer, ptr_list::release wants iterator
by Szczepan Hołyszewski 26 Feb '08

26 Feb '08

I have a pointer to an object that is in a ptr_list, and I want to "detach" the object from the list without deleting it. Problem: ptr_list::release() wants an iterator, not a pointer. A naive attempt at std::finding the iterator fails because ptr_list's indirection kicks in and std::find ends up wanting to compare values instead of pointers. Luckily my objects aren't equality comparable so I get a compile error instead of an elusive bug. As a workaround I just walk the list with BOOST_FOREACH() but it feels kludgy. Is there a "designed" way to feed a ptr_list an element pointer and milk the corresponding iterator? Is there a direct way do detach an item from ptr_list without deleting it? Did I overlook something obvious? Thanks in advance, Szczepan Holyszewski

2 1

BGL: ColorMap for depth_first_search
by Christian Sturz 25 Feb '08

25 Feb '08

Hi, I want to use the depth_first_search algorithm. Since I need the DFS version where a start vertex is specified, I also have to define a ColorMap. That's the function I want to use: template <class Graph, class DFSVisitor, class ColorMap> void depth_first_search(const Graph& g, DFSVisitor vis, ColorMap color, typename graph_traits<Graph>::vertex_descriptor start) However, I'm not able to define a working ColorMap. Since the map is irrelevant for me, the default ColorMap would be fully sufficient. Could you give me a hint how to create it? And another short question. I want to iterate the graph backwards via DFS. Could I use the "reverse_graph adaptor" with: depth_first_search( boost::make_reverse_graph(G), visitor( vis ) ...)? Regards, Christian -- GMX startet ShortView.de. Hier findest Du Leute mit Deinen Interessen! Jetzt dabei sein: http://www.shortview.de/?mc=sv_ext_mf@gmx

1 0

Linking problem with read_graphviz
by Michaela Kedikova 25 Feb '08

25 Feb '08

Hello, I'm new to boost(I'm working with boost_1_34_1) and I have a problem with read_graphviz. I'm working with MSVC 8. I've already built all external libraries using bjam and link them to my project. I also built libbgl-viz.lib, which contains all the source files in libs/graph/src.(As it was suggested in a thread) But I still have a link-error when I try to compile the graphviz.cpp example. I've searched in the mailing-list for solutions of this problem, but I didn't find anything, which helped me. Now I hope that somebody could give me a hint, what the problem may be. (My suggestion is, whether there is a missing object in a library or there are some difficulties between boost1_34_1 and MSVC 8.0). I post the error message below: 1>test_graphviz.obj : error LNK2019: unresolved external symbol "void __cdecl boost::read_graphviz(class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> > const &,class boost::subgraph<class boost::adjacency_list<struct boost::vecS,struct boost::vecS,struct boost::undirectedS,struct boost::property<enum boost::vertex_attribute_t,class std::map<class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> >,class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> >,struct std::less<class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> > >,class std::allocator<struct std::pair<class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> > const ,class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> > > > >,struct boost::no_property>,struct boost::property<enum boost::edge_attribute_t,class std::map<class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> >,class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> >,struct std::less<class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> > >,class std::allocator<struct std::pair<class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> > const ,class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> > > > >,struct boost::property<enum boost::edge_index_t,int,struct boost::no_property> >,struct boost::property<enum boost::graph_graph_attribute_t,class std::map<class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> >,class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> >,struct std::less<class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> > >,class std::allocator<struct std::pair<class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> > const ,class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> > > > >,struct boost::property<enum boost::graph_vertex_attribute_t,class std::map<class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> >,class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> >,struct std::less<class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> > >,class std::allocator<struct std::pair<class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> > const ,class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> > > > >,struct boost::property<enum boost::graph_edge_attribute_t,class std::map<class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> >,class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> >,struct std::less<class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> > >,class std::allocator<struct std::pair<class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> > const ,class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> > > > >,struct boost::property<enum boost::graph_name_t,class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> >,struct boost::no_property> > > >,struct boost::listS> > &)" (?read_graphviz@boost@@YAXABV?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@AAV?$subgraph@V?$adjacency_list@UvecS@boost@@U12@UundirectedS@2@U?$property@W4vertex_attribute_t@boost@@V?$map@V?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@V12@U?$less@V?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@@2@V?$allocator@U?$pair@$$CBV?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@V12@@std@@@2@@std@@Uno_property@2@@2@U?$property@W4edge_attribute_t@boost@@V?$map@V?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@V12@U?$less@V?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@@2@V?$allocator@U?$pair@$$CBV?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@V12@@std@@@2@@std@@U?$property@W4edge_index_t@boost@@HUno_property@2@@2@@2@U?$property@W4graph_graph_attribute_t@boost@@V?$map@V?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@V12@U?$less@V?$basic_string@DU?$char_traits@D@std@@V?$allocator@ D@2@@std@@@2@V?$allocator@U?$pair@$$CBV?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@V12@@std@@@2@@std@@U?$property@W4graph_vertex_attribute_t@boost@@V?$map@V?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@V12@U?$less@V?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@@2@V?$allocator@U?$pair@$$CBV?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@V12@@std@@@2@@std@@U?$property@W4graph_edge_attribute_t@boost@@V?$map@V?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@V12@U?$less@V?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@@2@V?$allocator@U?$pair@$$CBV?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@V12@@std@@@2@@std@@U?$property@W4graph_name_t@boost@@V?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@Uno_property@2@@2@@2@@2@@2@UlistS@2@@boost@@@1@@Z) referenced in function _main 1>D:\Mihaela\Projects_Cpp\Test_Boost\Debug\Test_graphviz.exe : fatal error LNK1120: 1 unresolved externals Sorry for my bad english. Thanks in advance, Mishe __________________________________________________________ Sent from Yahoo! Mail. A Smarter Inbox. http://uk.docs.yahoo.com/nowyoucan.html

1 0

betweenness centrality
by ivanballatore＠libero.it 25 Feb '08

25 Feb '08

Hi, i'm an italian developper and i have a question about the function brandes_betweenness_centrality. I would know if it's possbile to run brandes_betweenness_centrality only on a subset of nodes of a graph, because the execution of the function on the entire graph is too long. I want to calculate the betweenness approximation starting from the value computed for this subset of vetexes. Sorry for my english(i'm italian...).

1 0

subgraph with vertex lists
by Boris Lau 25 Feb '08

25 Feb '08

Hello, I'm new to boost and a bit confused about how to choose the vertex list type. My graph has around 50.000, and will be pruned down to roughly 500, so I use clear_vertex and remove_vertex a lot. According to http://www.boost.org/libs/graph/doc/adjacency_list.html, this works well only with VL=listS, otherwise I cannot safely remove vertices. Is that true? Ok, let's assume for now I need to use VL=listS in order to be able to safely remove vertices. How can I use the subgraph adaptor? It seems to require the vertex_index property, so I tried this: typedef subgraph< adjacency_list<listS, listS, undirectedS, property<vertex_index_t, int>, property<edge_index_t, int> > > Graph; which gives me nasty errors. It does make sense, if adding vertex_index to listS does not work. I was surprised by this, because adding edge_index does do the job if I use EL=listS: typedef subgraph< adjacency_list<listS, vecS, undirectedS, property<vertex_index_t, int>, property<edge_index_t, int> > > Graph; So, back to the topic. I need to be able to remove vertices and edges, and I need to use subgraphs. What can I do? Thanks for any help, Boris

1 1

fusion library not in boost 1.34.1?
by Robert Dailey 25 Feb '08

25 Feb '08

Hi, I've been working out of the development branch for boost for a while and I always took for granted that fusion was part of the library. However, after downgrading back to 1.34.1, I notice it isn't in there. Is there a reason for this? How do I go about downloading a version of fusion that will work with 1.34.1? Thanks.

3 3

Building boost 1.34.1 with MSVC 8 (64-bit)
by Eric Woodruff 25 Feb '08

25 Feb '08

Hello, I hope John Maddock is reading this as he had some knowledge of the Intel compiler problem. Similar to the thread quoted below, I am getting the "Windows cannot open this file" dialog when trying to build boost using VC8 as 64-bit. (Doesn't have /this/ problem when I don't set address-model=64 or use the VC7.1 x86 compiler for 32-bit builds.) This is with boost-jam-3.1.16-1-ntx86. I've tried adding the full path to cl.exe in user-config.jam without success. For historical reasons, I am using an nmake Makefile (under a Cygwin environment) to fit the boost build into an existing build system. It essentially operates as follows: tar jxf boost_1_34_1.tar.bz2 echo using msvc : 8.0 : "C:/WINDDK/3790.1830/bin/win64/x86/amd64/cl.exe" ; > boost_1_34_1\tools\build\v2\user-config.jam echo using stlport : 5.1.5 : c:\cygwin\home\ewoodruff\build\trunk\winnt\amd64\x64-WINNT\TOOLS\usr\include\stlport c:\cygwin\home\ewoodruff\build\trunk\winnt\amd64\x64-WINNT\TOOLS\usr\lib ; >> boost_1_34_1\tools\build\v2\user-config.jam sed -ie 's,\\\\,/,g' boost_1_34_1\tools\build\v2\user-config.jam cd boost_1_34_1 && "c:\cygwin\home\ewoodruff\build\trunk\winnt\amd64\boost\boost-jam-3.1.16-1-ntx86\bjam.exe" "-q" "--toolset=msvc-8.0" "--prefix=c:\cygwin\home\ewoodruff\build\trunk\winnt\amd64\x64-WINNT\TOOLS" "runtime-link=shared" "link=static" "threading=multi" "stdlib=stlport-5.1.5" "address-model=64" "--without-python" "--without-wave" stage install Once the first directory leaf is created is when the .rcp file tries to get run. Here's the content of that file: $ cat bin.v2/libs/serialization/build/msvc-8.0/debug/address-model-64/link-static/stdlib-stlport-5.1.5/stdlib-stlport-5.1.5/threading-multi/basic_archive.obj.rsp "libs\serialization\src\basic_archive.cpp" -Fo"bin.v2\libs\serialization\build\msvc-8.0\debug\address-model-64\link-static\stdlib-stlport-5.1.5\stdlib-stlport-5.1.5\threading-multi\basic_archive.obj" -TP /Z7 /Od /Ob0 /W3 /GR /MDd /Zc:forScope /Zc:wchar_t /favor:blend /wd4675 /EHs /Gy -c -DBOOST_ALL_NO_LIB=1 -D_STLP_DEBUG=1 -D_STLP_DEBUG_UNINITIALIZED=1 -D_STLP_USE_DYNAMIC_LIB=1 "-I." "-Ic:\cygwin\home\ewoodruff\build\trunk\winnt\amd64\x64-WINNT\TOOLS\usr\include\stlport" I'm also curious, is stdlib-stlport-5.1.5 supposed to be in the path twice? Thanks in advance, Eric > > > *Subject:* [Boost-users] Building boost 1.34.1 > > > *From:* James Emerton > > > *Date:* 2007-08-31 18:08:46 > > ------------------------------------------------------------------------ > I'm having difficulties building boost 1.34.1 using the intel > toolchain. Once bjam tries to invoke the compiler, I get the "Windows > cannot open this file" dialog, where the file is an .obj.rsp file. > > It would appear that perhaps the compiler itself is not making it into > the command line, thus causing the system to try executing the .rsp > file directly. I am not having this problem if I use the msvc > toolchain. > > > Command to invoke the build: > bjam --toolset=intel-9.1 > Compiler: > Intel(R) C++ Compiler for 32-bit applications, Version 9.1 > Build 20060519Z > bjam: > Boost.Jam Version 3.1.14. OS=NT. > > > Some environment variables: > > ICPP_COMPILER91=C:\Program Files\Intel\Compiler\C++\9.1 > INTEL_BASE_MSVC_TOOLSET=vc-7_1 > INTEL_LICENSE_FILE=C:\Program Files\Common Files\Intel\Licenses > INTEL_PATH=C:\Program Files\Intel\Compiler\C++\9.1\Ia32 > INTEL_SHARED=C:\Program Files\Common Files\Intel\Shared Files > > I've run the iclvars batch script, icl.exe is on the path, and I've > used this environment to build boost 1.33 successfully. I've tried > various debug levels of bjam, and I haven't seen anything that makes a > problem obvious to me. > > Thanks > James > _______________________________________________ > Boost-users mailing list > Boost-<email removed> > http://lists.boost.org/mailman/listinfo.cgi/boost-users > > ------------------------------------------------------------------------ > > > *From:* John Maddock > > > *Date:* 2007-08-31 19:17:38 > > ------------------------------------------------------------------------ > James Emerton wrote: > /> I'm having difficulties building boost 1.34.1 using the intel/ > /> toolchain. Once bjam tries to invoke the compiler, I get the "Windows/ > /> cannot open this file" dialog, where the file is an .obj.rsp file./ > />/ > /> It would appear that perhaps the compiler itself is not making it into/ > /> the command line, thus causing the system to try executing the .rsp/ > /> file directly. I am not having this problem if I use the msvc/ > /> toolchain./ > > Yep, it's bug report already: http://zigzag.cs.msu.su/boost.build/ticket/137 > > It does work if you add an entry for intel to a user-config.jam in your HOME > directory, mine looks like: > > using intel-win : 9.1 : "c:/program > files/intel/compiler/c++/9.1/ia32/bin/icl.exe" ; > using intel-win : 10.0 : "c:/program > files/intel/compiler/c++/10.0.025/ia32/bin/icl.exe" ; > using intel-win : 8.1 : "c:/program > files/intel/cpp/compiler80/ia32/bin/icl.exe" ; > #using intel-win : 7 : "c:/program > files/intel/cpp/compiler70/ia32/bin/icl.exe" ; > #using intel-win : 6 : "c:/program > files/intel/cpp/compiler60/ia32/bin/icl.exe" ; > > See the Boost.Build docs for more info on setting this up. And of course in > most cases you can always build the source from your IDE should you prefer > :-) > > HTH, John. > > _______________________________________________ > Boost-users mailing list > Boost-<email removed> > http://lists.boost.org/mailman/listinfo.cgi/boost-users >

3 4

boost::crc32 - visual studio crt - bug?!
by gfsdgadas＠Safe-mail.net 24 Feb '08

24 Feb '08

I want to compute the Crc32 checksum of a bunch of bytes. To do so, I simply use crc_32_type and process_bytes. Unfortunately this does not work in debug mode. Visual Studio (2008) breaks at line 574 of crc.hpp. // Compare a byte to the remainder's highest byte static unsigned char index( value_type rem, unsigned char x ) { return x ^ rem; } Error message: Run-Time Check Failure #1 - A cast to a smaller data type has caused a loss of data. If this was intentional, you should mask the source of the cast with the appropriate bitmask. For example: char c = (i & 0xFF); Changing the code in this way will not affect the quality of the resulting optimized code. I do not like changing the boost source code, so what can I do to workaround this problem?

2 1

GPGPU
by Carlo Alberto Ferraris 24 Feb '08

24 Feb '08

Just a casual question: has anyone considered the possibility to write and include a GPGPU library in boost? CAF

3 3

[call_traits] call_traits<enum>::param_type is const reference
by Sergei Politov 23 Feb '08

23 Feb '08

As I can see, call_traits for enum types defines param_type as const reference. I cannot find a reason for it. It is possible to use is_enum to determine whether type is enum or not, and define param_type for enums similar to value_type. Could anybody describe the situation?

1 0

Re: [Boost-users] All orderings of boost::graph::topological_sort()
by Matan Nassau 23 Feb '08

23 Feb '08

Excuse me for the double-post: my previous post had a bad subject line... Thank you very much for your prompt reply, Stephen. On Fri, Feb 22, 2008 at 12:51, Stephen Nuchia wrote: > For any significant-sized input the number of orderings will be too > large to do anything useful with. At the moment my input consists of about 32 vertices, although I'd like not to count on it to remain this way. I can probably assume the number of nodes will remain in the order of magnitude of hundreds or less. I'm talking about a schedule generating program for an undergraduate student. It generates a schedule for a semester for a student, given some constraints (time of day, courses done already and so on) and the course dependency (this course must not be taken before that one and so on). I would very much like to support other complexities such as some courses may be taken while a certain dependent is being taken, while others must be taken before any of their dependents may be taken. But I leave such complexities as worries for future versions. For the time being, all I want is to be able to generate all the different schedules (i.e., all various topo sorts). Indeed, something like std::next_permutation() would make a lot of sense. It would be fast and flexible. My next task would then be to find an algorithm to quickly estimate the number of different orderings. :) > For an input small enough that the > set of feasible orderings will fit in memory I would simply generate all > permutations and screen for feasibility rather than attempting to create > an enumeration algorithm. I don't follow. Can I change the ordering of vertices without ruining the graph's integrity? How? Would this affect the toposort output? > Create a comparability matrix and use it to > prune the permutation tree, perhaps? N^2 bits doesn't sound > prohibitive. I don't follow... > This has probably been studied but I'm at work so I'll speculate. For > larger inputs with many constraints it may be worthwhile to create an > enumeration algorithm. > > For purposes of empirical testing of build-ordering systems it might be > sufficient to go into the algorithm and randomize the selection > processes whenever it makes arbitrary choices. This may or may not be > easily achieved by randomly pre-sorting the input, depending on the > internals of the algorithm. If the sort is stable that will work. That is the reason for my query on my OP in the first place. I don't quite understand how manipulate the input vertices of the depth_first_search(). For what I understand, depth_first_search() works on a set of start vertices (which have no incoming edges), and for each of these the algorith dives in to the depth of the vertex's tree. Permuting the order of these vertices would give me advance me a lot, but I don't know how to do that. > If there is enough interest to justify adding an efficient deterministic > enumerator for poset orders I'll write one. Graph-theoretical > algorithms are a fun way to spend the weekend :-) I personally think the topological_sort() algorithm begs for an algorithm that would allow iterating over all unique orderings. As I stated above, I think something of the sort of std::next_permutation() would make a lot of sense, especially because graph algorithms, and finding all different toposorts in particular, tend to be expensive operations. Simply asking for one ordering at a time gives a lot of flexibility to the user with the added benefit of efficiency and integration with the STL's way of doing things. -- Matan

2 1

Using boost::barrier dynamically
by Gonzalo Garramuño 23 Feb '08

23 Feb '08

I'm trying to have several threads synced with a boost::barrier. But I'm finding its design very limiting (or, more likely, my understanding of its use is too poor). I want to have a barrier created dynamically (based on some parameters that determine how many threads run). And I also want to be able to delete the barrier at any point in time (from a thread that does not wait on the barrier). I expect this to force all threads that are waiting on the barrier to continue (ie. send a notify_all). However, boost::barrier's destructor just does nothing, so I get a hang. What's the right way to achieve this? PS. On a separate note... is there any reason why all of boost primitives have their variables private? I really would find it pretty useful to sometimes be able to query, for example, the lock count of some objects. -- Gonzalo Garramuño ggarra(a)advancedsl.com.ar AMD4400 - ASUS48N-E GeForce7300GT Xubuntu Gutsy

1 0

Re: [Boost-users] All orderings of boost::graph::topological_sort()"
by Matan Nassau 23 Feb '08

23 Feb '08

Thank you very much for your prompt reply, Stephen. On Fri, Feb 22, 2008 at 12:51, Stephen Nuchia wrote: > For any significant-sized input the number of orderings will be too > large to do anything useful with. At the moment my input consists of about 32 vertices, although I'd like not to count on it to remain this way. I can probably assume the number of nodes will remain in the order of magnitude of hundreds or less. I'm talking about a schedule generating program for an undergraduate student. It generates a schedule for a semester for a student, given some constraints (time of day, courses done already and so on) and the course dependency (this course must not be taken before that one and so on). I would very much like to support other complexities such as some courses may be taken while a certain dependent is being taken, while others must be taken before any of their dependents may be taken. But I leave such complexities as worries for future versions. For the time being, all I want is to be able to generate all the different schedules (i.e., all various topo sorts). Indeed, something like std::next_permutation() would make a lot of sense. It would be fast and flexible. My next task would then be to find an algorithm to quickly estimate the number of different orderings. :) > For an input small enough that the > set of feasible orderings will fit in memory I would simply generate all > permutations and screen for feasibility rather than attempting to create > an enumeration algorithm. I don't follow. Can I change the ordering of vertices without ruining the graph's integrity? How? Would this affect the toposort output? > Create a comparability matrix and use it to > prune the permutation tree, perhaps? N^2 bits doesn't sound > prohibitive. I don't follow... > This has probably been studied but I'm at work so I'll speculate. For > larger inputs with many constraints it may be worthwhile to create an > enumeration algorithm. > > For purposes of empirical testing of build-ordering systems it might be > sufficient to go into the algorithm and randomize the selection > processes whenever it makes arbitrary choices. This may or may not be > easily achieved by randomly pre-sorting the input, depending on the > internals of the algorithm. If the sort is stable that will work. That is the reason for my query on my OP in the first place. I don't quite understand how manipulate the input vertices of the depth_first_search(). For what I understand, depth_first_search() works on a set of start vertices (which have no incoming edges), and for each of these the algorith dives in to the depth of the vertex's tree. Permuting the order of these vertices would give me advance me a lot, but I don't know how to do that. > If there is enough interest to justify adding an efficient deterministic > enumerator for poset orders I'll write one. Graph-theoretical > algorithms are a fun way to spend the weekend :-) I personally think the topological_sort() algorithm begs for an algorithm that would allow iterating over all unique orderings. As I stated above, I think something of the sort of std::next_permutation() would make a lot of sense, especially because graph algorithms, and finding all different toposorts in particular, tend to be expensive operations. Simply asking for one ordering at a time gives a lot of flexibility to the user with the added benefit of efficiency and integration with the STL's way of doing things. -- Matan

1 0

All orderings of boost::graph::topological_sort()
by Matan Nassau 22 Feb '08

22 Feb '08

Hello, In implementing a dependency graph subsystem, one of the tasks is figuring out all the possible ways of traversing the graph with respect to the dependencies. boost::graph::topological_sort() looks like a reasonable place to start; however, topological_sort() only gives /one/ linear ordering, not all possible orderings. I digged into the code and found that the algorithm is in fact a call to depth_first_search with a custom visitor. The 0-in-degree vertices are not mentioned in the algorithm, and I suspect manipulating the order of these vertices would be a good start in generating different orderings. I also digged into the depth_first_search() code, but there too it doesn't seem to be straightforward to change the order in which those start-nodes are visited. I am also aware that even if I would find a way to change the order in which the 0-in-degree vertices are visited it will still not give me all the various orderings; I think that repeating the procedure with a bredth_first_search() will give me yet more different orderings, and with I'll have all the different orderings. I may have missed something obvious but I really don't know how to tackle the problem of generating various orderings. Any advice, please? Thank you! -- Matan

2 1

void_cast patch
by VennGrayDog 22 Feb '08

22 Feb '08

I found #pragma warning(push) without #pragma warning(pop) difs attached Index: void_cast.hpp =================================================================== --- void_cast.hpp (revision 43370) +++ void_cast.hpp (working copy) @@ -221,6 +221,10 @@ } // namespace serialization } // namespace boost +#ifdef BOOST_MSVC +# pragma warning(pop) +#endif + #include <boost/config/abi_suffix.hpp> // pops abi_suffix.hpp pragmas #endif // BOOST_SERIALIZATION_VOID_CAST_HPP

1 0

Re: [Boost-users] Problem using boost::bind with VC2005 (compiler crashes)
by Daniel Lidström 22 Feb '08

22 Feb '08

> > The compiler seems to crash when trying to compile the > > RailCaterpillarControllerGUI constructor. > > Any ideas why? > > I'm using 8.0.50727.867 (vsvista.050727-8600), boost-1.34.1 > > Try adding the argument placeholders to your bind call: > > // cbToggleDeadZone should be called by widget's callback > CBF_Callback callback( > boost::bind( > &RailCaterpillarControllerGUI::cbToggleDeadZone, > this, > _1, > _2 > )); > > > HTH / Johan Of course, forgot the placeholders... Thanks! -- Daniel

1 0

boost::filesystem::detail::not_found_error with Boost installer on VC2005
by Rainer Thaden 22 Feb '08

22 Feb '08

Hi all, yes, I know, there has been a thread some days ago, but I just registered and don't know how to reply to a mail that I don't have. Anyway, there seems to be a bug in boost 1.34.1 with VC 2005 when using directory_iterator. I used the filesystem before and everything worked fine. Now (since I use the iterator), I get 1>HD2.obj : error LNK2019: Verweis auf nicht aufgelöstes externes Symbol ""unsigned int boost::filesystem::detail::not_found_error" (?not_found_error@detail@filesystem@boost@@3IA)" in Funktion ""private: unsigned int __thiscall boost::filesystem::basic_directory_iterator<class boost::filesystem::basic_path<class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> >,struct boost::filesystem::path_traits> >::m_init(class boost::filesystem::basic_path<class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> >,struct boost::filesystem::path_traits> const &)" (?m_init@?$basic_directory_iterator@V?$basic_path@V?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@Upath_traits@filesystem@boost@@@filesystem@boost@@@filesystem@boost@@AAEIABV?$basic_path@V?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@Upath_traits@filesystem@boost@@@23@@Z)". I looked into the filesystem dll with depends and found the function there with ordinal 48 (0x30) (in both the debug and the release Version for multithreading). Also in the lib, I found it. Rather strange. The signatures of the linker and the ones found in the DLL also match. I used the installer from Boost-Consulting. In the thread some time ago, a responder claimed, he could compile successfully. Maybe he used a self-compiled boost version? I am going to test that, but maybe somebody has an answer right now? Best regards, Rainer Thaden

1 1

Problem using boost::bind with VC2005 (compiler crashes)
by Daniel Lidström 22 Feb '08

22 Feb '08

Hello! I'm trying to use boost::bind and boost::function to perform a callback mechanism. However, the VC2005 compiler crashes for some reason. Let me show the code right away: #include <boost/bind.hpp> #include <boost/function.hpp> #include <boost/any.hpp> #include <boost/noncopyable.hpp> #include <iostream> #include <tchar.h> struct Fl_Widget; typedef boost::function<void (Fl_Widget&, const boost::any&)> CBF_Callback; struct Fl_Widget { CBF_Callback cbf_callback_; boost::any user_data_; //Three calls to set or probe that functor callback variable. CBF_Callback cbf_callback() const {return cbf_callback_;} void cbf_callback(CBF_Callback c, const boost::any& p) {cbf_callback_=c; user_data_=p;} void cbf_callback(CBF_Callback c) {cbf_callback_=c;} //Calls to do the functor callback void do_cbf_callback() {cbf_callback_(boost::ref(*this),user_data_);} // override user_data_ void do_cbf_callback(const boost::any& user_data) {cbf_callback_(boost::ref(*this),user_data);} }; struct GUI : private boost::noncopyable { static GUI& Instance() { static GUI theGUI; return theGUI; } void print() { std::cout << "Inside GUI" << std::endl; } private: GUI(){} }; struct RailCaterpillarControllerGUI { Fl_Widget widget; RailCaterpillarControllerGUI() { // cbToggleDeadZone should be called by widget's callback CBF_Callback callback(boost::bind(&RailCaterpillarControllerGUI::cbToggleDeadZone, this)); widget.cbf_callback(callback); } void cbToggleDeadZone(Fl_Widget& widget, const boost::any& data) { // assume data is reference to the GUI instance boost::any_cast<boost::reference_wrapper<GUI> >(data).get().print(); } }; int _tmain(int argc, TCHAR** argv) { RailCaterpillarControllerGUI controller; controller.widget.do_cbf_callback(boost::ref(GUI::Instance())); } Compiling gives this: 1>------ Build started: Project: test, Configuration: Debug Win32 ------ 1>Compiling... 1>test.cpp 1>c:\users\daniel.sbg\programming\boost_1_34_1\boost\bind.hpp(1575) : warning C4180: qualifier applied to function type has no meaning; ignored 1> c:\users\daniel.sbg\programming\boost_1_34_1\boost\bind.hpp(1609) : see reference to class template instantiation 'boost::_bi::add_cref<Pm,I>' being compiled 1> with 1> [ 1> Pm=void (__thiscall RailCaterpillarControllerGUI::* )(Fl_Widget &,const boost::any &), 1> I=1 1> ] 1> c:\users\daniel.sbg\programming\test\test.cpp(48) : see reference to class template instantiation 'boost::_bi::dm_result<Pm,A1>' being compiled 1> with 1> [ 1> Pm=void (__thiscall RailCaterpillarControllerGUI::* )(Fl_Widget &,const boost::any &), 1> A1=RailCaterpillarControllerGUI * 1> ] 1>c:\users\daniel.sbg\programming\boost_1_34_1\boost\mem_fn.hpp(318) : warning C4180: qualifier applied to function type has no meaning; ignored 1> c:\users\daniel.sbg\programming\boost_1_34_1\boost\bind\bind_template.hpp(22 5) : see reference to class template instantiation 'boost::_mfi::dm<R,T>' being compiled 1> with 1> [ 1> R=void (Fl_Widget &,const boost::any &), 1> T=RailCaterpillarControllerGUI 1> ] 1>Project : error PRJ0002 : Error result 255 returned from 'C:\Program Files\Microsoft Visual Studio 8\VC\bin\cl.exe'. The compiler seems to crash when trying to compile the RailCaterpillarControllerGUI constructor. Any ideas why? I'm using 8.0.50727.867 (vsvista.050727-8600), boost-1.34.1 Thanks in advance! -- Daniel

2 1

regexp - how to do an anti-match on a word
by David Ward 22 Feb '08

22 Feb '08

Boosters: I'm having trouble determining a regular expression that will do the following: 1. Match the following strings: TermGeMstreamIPTriplePlay_11.0.IpTxMstream_0.DefaultGateway TermGeMstreamIPTriplePlay_11.0.VplsTxMstream_0.CurMbps TermGeMstreamIPTriplePlay_11.0.MplsTxMstream_0.AvgMbps TermGeMstreamIPTriplePlay_11.0.MacTxMstream_0.Address 2. But not match the string: TermGeMstreamIPTriplePlay_11.0.TrafficTxMstream_0.RampStart So, essentially match the strings with IpTx, VplsTx, MplsTx and MacTx, but not match strings with TrafficTx. Right now, I have: .*\.(IpTx|MacTx|MplsTx|VplsTx)Mstream_0\..* This works, however, it is not very robust because there may be more entries added to the matching words over time. Is there someway to do an anti-match on the word "TrafficTx"? Something like: .*\.(!TrafficTx)Mstream_0\..* Thanks, David

2 1

"Re: Boost, HP-UX and gcc"
by stacie＠dcindigos.com 22 Feb '08

22 Feb '08

I'm on 11.23 itanium using gcc 4.1.1 and the latest build of 1.35 beta (feb 21) With my setup, c++config,h indicates that the header wchar.h does exist, however gcc does *not* set _GLIBCPP_USE_WCHAR_T nor_GLIBCXX_USE_WCHAR_T. My build keeps throwing No WCHAR_MIN and WCHAR_MAX errors, and suggests that I adjust integer_traits<>. I was going to just let it slide but unfortunately I received this error while trying to build using the Unit Test framework. Since unit testing is my friend, I would really like to get this working. I tried a work-around from the boost mailing list, but the code that I found in my files was not quite like what he modified: ***** Added code at top of bad header file: #ifdef hpux #include <wchar.h> #undef WCHAR_MIN /* (wchar_t)0 */ #define WCHAR_MIN 0 #endif * error simply changed to: /home/mlinck/develop/api/tp-sources-trunk/boost_1_33_1/boost/regex/v4/regex_traits_defaults.hpp:94:28: error: missing binary operator before token "(" * tried a different workaround at the code where the error is thrown. Noticed that broken code gets skipped for HP_aCC compiler. Figure we can maybe get away with same for gcc compiler: => #if !defined HP_aCC && !defined hpux * seems to compile. Had to apply to boost/regex/v4/perl_matcher.hpp as well ****** How would I modify integer_traits<> ? Since that is what is suggested...... Thanks!

3 2

Re: [Boost-users] "Re: Boost, HP-UX and gcc"
by stacie＠dcindigos.com 22 Feb '08

22 Feb '08

Hi Boris, No help there. I'm still getting the same errors. I'm willing to look at it some more if you have any more suggestions. How can I turn off the building of the python and serialization libraries? There are other libraries that may be broken with HP, too and I'd like to exclude those from the build. Stacie > -------- Original Message -------- > Subject: Re: [Boost-users] "Re: Boost, HP-UX and gcc" > From: "Boris Gubenko" <Boris.Gubenko(a)hp.com> > Date: Thu, February 21, 2008 4:19 pm > To: <boost-users(a)lists.boost.org> > Cc: Boris Gubenko <Boris.Gubenko(a)hp.com> > > > I don't think John's fix in http://svn.boost.org/trac/boost/changeset/42890 made > it to 1.35. You migh want to apply it locally and see if it fixes your problem. > > Thanks, > Boris > > ----- Original Message ----- > From: <stacie(a)dcindigos.com> > To: <boost-users(a)lists.boost.org> > Sent: Thursday, February 21, 2008 3:36 PM > Subject: [Boost-users] "Re: Boost, HP-UX and gcc" > > > > I'm on 11.23 itanium using gcc 4.1.1 and the latest build of 1.35 beta > > (feb 21) > > > > With my setup, c++config,h indicates that the header wchar.h does exist, > > however gcc does *not* set _GLIBCPP_USE_WCHAR_T nor_GLIBCXX_USE_WCHAR_T. > > > > My build keeps throwing No WCHAR_MIN and WCHAR_MAX errors, and suggests > > that I adjust integer_traits<>. > > I was going to just let it slide but unfortunately I received this error > > while trying to build using the Unit Test framework. Since unit testing > > is my friend, I would really like to get this working. > > > > I tried a work-around from the boost mailing list, but the code that I > > found in my files was not quite like what he modified: > > > > ***** > > Added code at top of > > bad header file: #ifdef hpux #include <wchar.h> #undef > > WCHAR_MIN /* (wchar_t)0 */ #define WCHAR_MIN 0 #endif > > * error simply changed > > to: > > /home/mlinck/develop/api/tp-sources-trunk/boost_1_33_1/boost/regex/v4/regex_traits_defaults.hpp:94:28: > > error: missing binary operator before token "(" > > * tried a different workaround at the code where the error > > is thrown. Noticed that broken code gets skipped for > > HP_aCC compiler. Figure we can maybe get away with same > > for gcc compiler: > > => #if !defined HP_aCC && !defined hpux > > > > * seems to compile. Had to apply to > > boost/regex/v4/perl_matcher.hpp as well > > ****** > > > > > > How would I modify integer_traits<> ? Since that is what is > > suggested...... > > > > > > Thanks! > > > > > > > > _______________________________________________ > > Boost-users mailing list > > Boost-users(a)lists.boost.org > > http://lists.boost.org/mailman/listinfo.cgi/boost-users > > > > _______________________________________________ > Boost-users mailing list > Boost-users(a)lists.boost.org > http://lists.boost.org/mailman/listinfo.cgi/boost-users

2 1

BGL: Finding a vertex by the vertex_index_t property
by Torsten Sadowski 21 Feb '08

21 Feb '08

Hi, is there a simple possibility to find a vertex in a adjacency_list graph by the vertex_index_t property? My uninspired way would be to build a map<int, vertex_descriptor> but there might be something more elegant. Cheers, Torsten

2 1

[variant] ideas needed
by ope 21 Feb '08

21 Feb '08

Hi, I've the following code: ---8<--- #include <boost/variant.hpp> #include <iostream> #include <string> struct available : public boost::static_visitor<bool> { template <typename T, typename U> bool operator()( const T&, const U& ) const { return false; // cannot compare different types } template <typename T> bool operator()( const T& lhs, const T& rhs ) const { static const std::string NA( "NA" ); return lhs != NA; //??? } }; using namespace boost; using namespace std; int main() { variant< double, std::string > R; variant< double, std::string > L; R = 3.14; L = "NA"; cout << R << endl; cout << L << endl; bool r = boost::apply_visitor(available(), R ); cout << ( r ? "true" : "false" ) << endl; bool l = boost::apply_visitor(available(), L ); cout << ( l ? "true" : "false" ) << endl; } --->8--- which does not compile. Anyway, from a data parser I get a double or a "NA" (not available) string. This test shall handle both cases. To process using the double I have to check if it available (than use defaults). Unfortunately, the static_visitor<bool> needs two arguments. How can I get the syntax (or simpler) as used in the example? Thanks Olaf

4 5

A request for help using the Boost Unit Test Library
by Tim Allison 21 Feb '08

21 Feb '08

I am working on a project that uses the Boost Multi-Index library and Boost shared_ptr, so I know I can use the Boost libraries on our project. However, I am unable to build so much as an example from the Boost Unit Test Library, and need help. (I assume the problem is a conflict with compiler switches, or something along those lines.) I set up a test directory, into which I will link the source code I wish to test. The Make file I am using is a modified form of one we use to build my project, so I know it will build both the module using some Boost libraries, and our old code, some of which is more than 10 years old. I am using boost 1.33.1-42@586 on Suse 10.2 (If it matters, my computer is a Latitude D510) and Gnu compiler 4.1.3-29@586 (as listed by Smart Package Manager). When I run the make file, it echos out the following line: g++ -Wno-deprecated -frepo -m32 -g -I../lib -I/usr/X11R6/include -I../../../source/ecs/shared -DOS_IS_LINUX -c FuserTests.cc FuserTests.cc is the file into which I have attempted to copy the examples. (The examples are both from the patmedia.net/rogeeff site, and the official boost site, as corrected within the discussion forum.) The case I am most interested in running uses BOOST_AUTO_TEST_CASE( functionName ). This line produces the error: FuserTests.cc:16: error: expected constructor, destructor, or type conversion before ‘(’ token Is there something special I need to do to get this to compile in C++? -- Tim Allison Chief Engineer MIDCON Millennium Engineering and Integration (719) 264-4302 Work (719) 264-4318 Fax

2 1

[has_xxx] How can it work at all???
by Ovanes Markarian 21 Feb '08

21 Feb '08

Hello *! As already asked many times in the list about the broken HAS_XXX ( BOOST_MPL_HAS_XXX_TRAIT_DEF<http://www.boost.org/libs/mpl/doc/refmanual/has-xxx-trait-def.html>) mpl I would like to ask following: If I use this code (on MSVC 7.1 compiler) and I have a type like: struct test { int xxx; }; and use the macro namespace x { BOOST_MPL_HAS_XXX_TRAIT_DEF(xxx); } This will result in the following code fragment: -------------------------------------------------------------------------------------------- /// included code from has_xxx.hpp namespace boost { namespace mpl { namespace aux { template< typename T > struct msvc71_sfinae_helper { typedef void type; }; } } } namespace sn { template< typename T, typename U = void > struct has_xxx_impl_ { static const bool value = false; typedef boost::mpl::bool_<value> type; }; template< typename T > struct has_xxx_impl_< T , typename boost::mpl::aux::msvc71_sfinae_helper< typename T::xxx >::type > { static const bool value = true; typedef boost::mpl::bool_<value> type; }; template< typename T, typename fallback_ = boost::mpl::bool_<false> > struct has_xxx : has_xxx_impl_<T> { }; } -------------------------------------------------------------------------------------------------- For me this will clearly fail, since T::xxx is not a type but an integer member of the struct. If xxx would be a member function it would not be a type as well, so has_xxx can only determine if there is some nested class type or a typedef within the inspected type. Is this correct? Many thanks, Ovanes

4 7

[bind, function] bind a cast
by Alexander Gutenev 21 Feb '08

21 Feb '08

Hi Is there any easy way to bind an explicit type conversion ? For example, I want to supress compiler warning here with static_cast: #include <boost/bind.hpp> #include <boost/function.hpp> int main(void) { boost::function<long (void)> foo; boost::function<short (void)> bar; bar = foo; // compiler warning. bar = boost::bind<short>(foo); // the same warning //bar = boost::bind(static_cast<short>, foo); // does not compile return 0; } If not, I think it might be added. Casts are useful sometimes not only to avoid warning, but also to get code compiled.

2 3

[Thread] Error compiling from svn trunk (tss.hpp)
by Luiz Marques 21 Feb '08

21 Feb '08

Hi, I'm trying to compile my code with boost.thread from SVN trunk (gcc 4.1.2, Fedora 7). I get the following error : /include/boost/thread/pthread/tss.hpp:55: error: invalid conversion from void* to MyThread* The code being compiled creates a thread_specific_ptr with a custom cleanup function, and is OK with boost 1.34 . Looking at the specified header, in line 55 there is a call to the custom cleanup_function: void operator()(void* data) { cleanup_function(data); } It seems like a static_cast<T*> is missing in the parameter. I changed it locally and the problem was solved. Can anyone confirm this problem ? Thanks, Luiz Marques

1 0

Re: [Boost-users] [regex] Why pattern anchored at buffer end ($)alsomatch for (\x0c)FF, (\x0d)CR
by Chandan Nilange 21 Feb '08

21 Feb '08

I am restarting this old thread. My Question: How to modify Boost regex library, to match perl behavior, such that, I only get pattern matches after/before \n and ignore (\x0c)FF, (\x0d)CR for $ & ^ anchored patterns ? Regards, Chandan John Maddock wrote: >Chandan Nilange wrote: > > >>Hello John, >> >>I not very clear. >> >>So do you mean that - >> >>$ & ^ matching pattern after/before any (\x0c)FF, >>(\x0d)CR is right PCRE /m modifier behaviour. >> >> > >I believe so, but deciding what the right behaviour should be is not >necessarily easy: Perl has the luxury that it controls File IO as well as >regexes: in contrast Boost.Regex is designed to accept text from all kinds >of sources and platforms. Most applications will in any case regard an >isolated \r or \n as a line break, as well as \r\n. Likewise in what sense >does a form-feed *not* start a new line? :-) > >HTH, John. > >_______________________________________________ >Boost-users mailing list >Boost-users(a)lists.boost.org >http://lists.boost.org/mailman/listinfo.cgi/boost-users > > ____________________________________________________________________________________ Be a better friend, newshound, and know-it-all with Yahoo! Mobile. Try it now. http://mobile.yahoo.com/;_ylt=Ahu06i62sR8HDtDypao8Wcj9tAcJ

2 4

which toolset for VC8 on XP64
by Andreas Fabri 21 Feb '08

21 Feb '08

Hello, With the precompiled boost libs from boost-consulting I get linker errors error LNK2019: unresolved external symbol "private: void __cdecl boost::detail::tss::init(class boost::function1<void,void *,class std::allocator<class boost::function_base> > *)" (?init@tss@detail@boost@@AEAAXPEAV?$function1@XPEAXV?$allocator@Vfunction_base@boost@@@std@@@3@@Z) referenced in function "public: __cdecl boost::detail::tss::tss(class boost::function1<void,void *,class std::allocator<class boost::function_base> > *)" (??0tss@detail@boost@@QEAA@PEAV?$function1@XPEAXV?$allocator@Vfunction_base@boost@@@std@@@2@@Z) I started a shell such that Microsoft Visual Studio 8/VC/BIN/amd64/CL.EXE is the compiler on the $PATH I then run bjam --toolset=msvc" but the problem persists. Is there another toolset I have to choose? thanks in advance, andreas

4 4

Link errors using VC++ 9.0, with Boost libs compiled with VC++ 8.0
by David Hartman 21 Feb '08

21 Feb '08

I'm compiling and linking using VC++ 9.0, and using Boost libraries built with VC++ 8.0. Unresolved external link errors are returned. Has anyone else had this problem? I'm going to build the Boost libraries using VC++ 9.0, for use with VC++ 9.0, but I was wondering if anyone else has tried to migrate to VC++ 9.0, and how they handled it. An example of the link error is below. 28>libboost_unit_test_framework-vc80-mt-gd.lib(xml_report_formatter.obj) : error LNK2019: unresolved external symbol "__declspec(dllimport) public: __thiscall std::_Container_base::~_Container_base(void)" (__imp_??1_Container_base@std@@QAE@XZ) referenced in function "protected: void __thiscall std::vector<struct std::pair<char,char const *>,class std::allocator<struct std::pair<char,char const *> > >::_Tidy(void)" (?_Tidy@?$vector@U?$pair@DPBD@std@@V?$allocator@U?$pair@DPBD@std@@@2@@std@@IAEXXZ) David Hartman

3 2

[program_options] Unicode problem
by Sergei Politov 21 Feb '08

21 Feb '08

The following example could not be compiled: ================================================================= #include <string> #include <boost/program_options.hpp> namespace po = boost::program_options; int main() { std::wstring test; po::options_description desc("Allowed options"); desc.add_options() ("test", po::wvalue<std::wstring>(&test)->default_value(L"value"), "description"); } ================================================================= It is due to wvalue::default_value tries to do lexical_cast<std::string>(std::wstring). Also I cannot change default_value to default_value("value") as soon as it accepts only something convertible to std::wstring. In other words it is impossible to specify default_value for wstring option.

3 8

HP-UX with GNU
by stacie＠dcindigos.com 21 Feb '08

21 Feb '08

A couple of questions for the HP gurus about Boost on HP-UX with GNU GCC. I see that the development builds target this configuration, but the 1.34 release does not, so I'm not sure that I can even do this with the 1.34 build. I am new to using bjam so be gnetle. Platform: HP-UX, on itanium ia64, 11.23, building for 64 bit linker: hp-ux GCC:4.1.1 Boost:1.34 User config file: using gcc: : :<cxxflags>-mlp64 <linkflags>-mlp64 Is this suffcient, or do I need to indicate the archtiecture (ia64 on itanium) in some way? Threading: gcc.jam indicates that GCC on HP-UX does not support threading. We have built many libraries with threading using GCC so is this an old comment or is it specific to the gnu linker and not HP-UXs. Can I modify a config file to support threading correctly? Failing libs Zlibs and bzlib builds fail. How do I disable building them? Can the build fail and it's not a problem for the rest of the libraries. Do I need to modify any of the source to work correctly? Would I be better off using a development build? Will boost work correctly on HP 11.31? Thanks!

4 3

[variant] Compilation problem : solved
by Olivier Tournaire 20 Feb '08

20 Feb '08

That was my mistake. Forget to #include <string> ... Regards -- Le temps des cerises reviendra. Dans l'immédiat, c'est le temps des noyaux. Courage.

1 0

[variant] Compilation problem
by Olivier Tournaire 20 Feb '08

20 Feb '08

Hi all, I am trying to use boost::variant. First of all, I tried to compile and run the example from the documentation (http://www.boost.org/doc/html/variant.html) However, I am having a compilation error. Here it is : C:\Boost\include\boost-1_35\boost\variant\detail\variant_io.hpp(64) : error C2679: '<<' binaire : aucun opérateur trouvé qui accepte un opérande de partie droite de type 'const T1' (ou il n'existe pas de conversion acceptable) --> error C2679: binary '<<' : no operator found which takes a right-hand operand of type 'type' (or there is no acceptable conversion) No '<<' operator for std::string ? Where is my mistake ? Best regards, Olivier -- Le temps des cerises reviendra. Dans l'immédiat, c'est le temps des noyaux. Courage.

1 0

Boost 1.35 build problem
by Johan Råde 20 Feb '08

20 Feb '08

I downloaded Boost 1.35 Head from SVN. I built it with the command: bjam --toolset=msvc-7.1 --NO_ZLIB=0 -sZLIB_SOURCE=C:/Qt/3.3.8/src/3rdparty/zlib install Then I tried to rebuild our app. I got the linking error: LINK : fatal error LNK1104: cannot open file 'libboost_serialization-vc71-mt-1_35.lib' The folder C:/Boost/lib does not contain the file libboost_serialization-vc71-mt-1_35.lib. It seems it was not built. The folder contains boost_serialization-vc71-mt-1_35.dll boost_serialization-vc71-mt-1_35.lib boost_serialization-vc71-mt.lib but no other xxx_serialization_yyy libs. The same happens with the other Boost libs, not just with the serialization lib. Does anyone have any idea what the problem might be, or what additional information I should provide? Thank you, Johan Råde

4 12

Re: [Boost-users] HP-UX with GNU
by stacie＠dcindigos.com 20 Feb '08

20 Feb '08

I submitted this question to the boost-build list, which is where I should have sent it originally. I have some questions about how you faked Boost builder out, but I don't want to continue it here. > -------- Original Message -------- > Subject: Re: [Boost-users] HP-UX with GNU > From: raulh39(a)ya.com > Date: Wed, February 20, 2008 3:13 pm > To: boost-users(a)lists.boost.org > > > On Wed, 20 Feb 2008 11:58:19 -0700 > stacie(a)dcindigos.com wrote: > > > > Threading: > > gcc.jam indicates that GCC on HP-UX does not support threading. We have > > built many libraries with threading using GCC so is this an old comment > > or is it specific to the gnu linker and not HP-UXs. Can I modify a > > config file to support threading correctly? > > I can't help for the rest of your questions, but we use GCC+HP-UX+Boost in our work without any problem. > > The only real difference with your environment is that we work in HP-UX 11.23 but on PA-RISC... > > We had to "force" Boost to believe that this environment is OK for building multi-thread versions of the libraries. > As you have said, gcc.jam indicates that GCC on HP-UX does not support threading. > _______________________________________________ > Boost-users mailing list > Boost-users(a)lists.boost.org > http://lists.boost.org/mailman/listinfo.cgi/boost-users

1 0

Why not overloading && and || in bind?
by Jean-Pierre Bergamin 20 Feb '08

20 Feb '08

Hello everyone Is there a reason why boost::bind does not overload the logical operators || and &&? They definitively would come in more handy to build up complicated logical expressions than using the std::logical_and and std::logical_or function objects. Adding... BOOST_BIND_OPERATOR( ||, logical_or ) BOOST_BIND_OPERATOR( &&, logical_and ) ...to the operator definitions in bind.hpp works as expected with the code below (tested with MSVC 2008). So - why not? :-) Regards James ////////////////////////////////////////////////// // Code example: ////////////////////////////////////////////////// struct employee { employee(int id, const string &name) : id(id), name(name) {} int id; string name; }; ostream &operator<<(ostream &o, const employee &e) { o << e.id << " - " << e.name << endl; return o; } typedef vector<employee> emp_container; emp_container staff; staff.reserve(4); staff.push_back(employee(1, "John")); staff.push_back(employee(2, "Paula")); staff.push_back(employee(3, "Sara")); staff.push_back(employee(4, "Geroge")); emp_container::const_iterator e = find_if(staff.begin(), staff.end(), bind(&employee::name, _1) == "Paula" || bind(&employee::name, _1) == "John" || bind(&employee::id, _1) >= 4 ); if (e != staff.end()) { cout << "Found: " << *e << endl; }

7 13

compile problem using multi_index_container
by Dwight Kelly 20 Feb '08

20 Feb '08

Hello, I've encountered a problem compiling code that uses boost multi_index_container. I'm using both Microsoft Visual Studio 2005 and Apple's gcc 4.0.1. Visual Studio was very picky about how I used namespaces. I finally got it working with the code below. The key to success was the 'using namespace boost::multi_index' line. If I attempted to prefix the members directly, eg. typedef typename boost::multi_index::nth_index_const_iterator<0>::type left_const_iterator; or namespace mi=boost::multi_index; typedef typename mi::nth_index_const_iterator<0>::type left_const_iterator; it would fail to compile. However, when I attempt to compile the same code on Mac OS X using gcc 4.0.1 I get the following errors. Does anyone know why? src.cpp:98:error: wrong number of template arguments (1, should be 2) /boost/boost/multi_index_container_fwd.hpp:46: error: provided for 'template<class MultiIndexContainer, int N> struct boost::multi_index::nth_index_const_iterator' src.cpp:98: error: expected ';' before 'left_const_iterator' src.cpp:99: error: wrong number of template arguments (1, should be 2) /boost/boost/multi_index_container_fwd.hpp:46: error: provided for 'template<class MultiIndexContainer, int N> struct boost::multi_index::nth_index_const_iterator' src.cpp:99: error: expected ';' before 'right_const_iterator' -----code---- using namespace boost::multi_index; 88 template<class A, class B=String> 89 class SimpleBiMap : public multi_index_container< 90 pair<A,B>, 91 indexed_by< 92 ordered_unique<member<pair<A,B>, A, &pair<A,B>::first> >, 93 ordered_unique<member<pair<A,B>, B, &pair<A,B>::second> > 94 > > 95 { 96 public: 97 typedef pair<A,B> mytype; 98 typedef typename nth_index_const_iterator<0>::type left_const_iterator; 99 typedef typename nth_index_const_iterator<1>::type right_const_iterator; 100 101 SimpleBiMap(const mytype init[], int N) { 102 for (int i=0; i<N; i++) { 103 insert(init[i]); 104 } 105 } 106 107 left_const_iterator find(const A &key) const { return get<0>().find(key); } 108 right_const_iterator find(const B &key) const { return get<1>().find(key); } 109 110 left_const_iterator left_end() const { return get<0>().end(); } 111 right_const_iterator right_end() const { return get<1>().end(); } 112 113 const B &operator[](const A &key) const { return get<0>().find(key)->second; } 114 const A &operator[](const B &key) const { return get<1>().find(key)->first; } 115 }; ---code--- Thank you, Dwight Kelly Apago, Inc.

2 1

Serialization problem
by christophe loustaunau 20 Feb '08

20 Feb '08

Hello, I try to use serialization of boost. I have started with the exemple of bus stop in XML : http://www.boost.org/libs/serialization/example/demo_xml.cpp I have added to this example a map of 1000 bus_schedule. Then I save this map in a file. And I relaod this file in a new map. This works perfectly in XML, but if I try : binary_iarchive I got an boost::archive::archive_exception at memory location 0x00dfdb40.. I don't understand why It woks in XML and not in binary. I have attached to the mail the source file. If someone could help me It would be great ! Thanks. -- Christophe Loustaunau.

3 3