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

tokenizer: exceptions not documented
by Keith MacDonald 25 Apr '04

25 Apr '04

I discovered the hard way that tokenizer (from CVS head) can throw escaped_list_error. It would help if that was documented. Keith MacDonald

1 0

Iterators
by Witz 24 Apr '04

24 Apr '04

just a quick question - which maybe should have been directed at the developers list but ... would it be possible in the iterator adaptors to have an operator== defined that takes the base iterator as the right hand side? it would certainly cut down on typing and also obviate the need to create an end iterator that is only used as an 'end of container' sentinel value if it was possible it would allow us to write, for example: copy( make_filter_iterator<is_positive_number>(numbers, numbers + N), numbers + N, ostream_iterator<int>(cout, " ")); rather than: copy( make_filter_iterator<is_positive_number>(numbers, numbers + N), make_filter_iterator<is_positive_number>(numbers + N, numbers + N), ostream_iterator<int>(cout, " ")); of course i am aware that i might be missing something obvious ... ian whittley

3 5

boost format problems
by Navneet Dalal 23 Apr '04

23 Apr '04

Hi I decided to try boost::format for more compact and typesafe formatting of text. However, I am running into some problems. Problem 1 ============== The following line throws boost:bad_format_string exception. std::cout << format("%|1| %|2$3|") % "Hello" % 3 << std::endl; However, this one works std::cout << format("%|1|") % "Hello" << std::endl; All I want to do is use notation for printing character streams i.e. %|1| instead of %1% Is this a bug or I am doing something wrong? Problem 2 ============= How can I print a float such that it appends 0 for remaining precision digits. For e.g. let us say I want to print 0.2 with 5 characters of overall length and till 3 digits of precision. Kind of output I am looking for is 0.200 One would expect that following command may work (as I am indicating left print, fill any remaining characters and till 3 digits of precision after decimal) std::cout << format("#%|1$-05.03|#") % 0.2 << std::endl; // however it prints #0.2 # This command prints something like std::cout << format("#%|1$05.03|#") % 0.2 << std::endl; // it prints #000.2# with C++ cout, I can print it like std::cout << '#' << std::left << std::setfill('0') << std::setw(5) << std::setprecision(3) << 0.2 << '#' << std::endl; so on similar lines I wrote following line. it works, but it seems ugly to use group only for this std::cout << format("#%|1$-05.3|#") % group(setfill('0'), 0.2) << std::endl; // prints #0.200# Does any one has any other suggestions? regards navneet

2 1

Compilation problem with BOOST::filesystem lib
by Peter Koerber 23 Apr '04

23 Apr '04

Hello, i try to compile the *.cpp files of boost::filesystem within the following environment: os: Windows XP compiler: Visual C++ Toolkit 2003 (this is the free version of MS Visual C++ V7.1) boost: Version 1.31.0 If BOOST_WINDOWS is defined, the file windows.h is included in the sources: * exception.cpp * operations_posix_windows.cpp. Otherwise the file errno.h is included. BOOST_WINDOWS is defined within the environment mentioned above, but the compiler does not have a windows.h include file. Hence compilation fails. OTOH, if BOOST_WINDOWS is "manually" undefined, the errno constant: ETXTBSY is not defined in the file errno.h which comes with the Visual C++ Toolkit 2003. Kind regards Peter

2 1

BGL: INTERNAL COMPILER ERROR in VC7.1
by Nicholas Burlett 22 Apr '04

22 Apr '04

I'm rather unsure about what is causing this, but I get an internal compiler error in Visual C++ 7.1 when using BGL in some specific circumstances. I have a graph defined in class Graph defined as such: typedef property<edge_feedback_t, bool> EdgeFeedback; typedef property<edge_sourcejack_t, IAudioOutputJack*, EdgeFeedback> EdgeSourceJack; typedef property<edge_destinationjack_t, IAudioInputJack*, EdgeSourceJack> EdgeDestinationJack; typedef EdgeDestinationJack EdgeProperty; typedef boost::property<boost::vertex_index_t, int> VertexIndexProp; typedef property<vertex_name_t, nsCOMPtr<IAudioModule>, VertexIndexProp> VertexProperty; typedef adjacency_list<listS, listS, bidirectionalS, VertexProperty, EdgeProperty> graphdata; typedef graph_traits < graphdata >::vertex_descriptor vertdsc; This definition works fine in the graph.cpp file that implements the class Graph. However, when I use graph.h in another cpp file, I get bizarre compiler errors: c:\Boost\include\boost -1_31\boost\mpl\aux_\preprocessed\plain\apply.hpp(50) : fatal error C1001: INTERNAL COMPILER ERROR After some trial and error, I narrowed the problem down to needing to have a != between two oedge_iterators _somewhere_ in the cpp file including graph.h. Simply putting the following function into my other files will cause them to compile just fine: void makeVChappy() { typedef graph_traits< Graph::graphdata >::out_edge_iterator oedge_iterator; oedge_iterator oiter, oend; oiter != oend; } I'm a a complete loss as to how this could be happening, but since I have a work-around for now, I'm not going to push the issue too much further at the moment --- I just thought I'd bring this to your attention, and see if anyone knows of a more elegant solution.

5 5

RE: [Boost-users] Compilation problem with BOOST::filesystem lib
by Ben Hutchings 22 Apr '04

22 Apr '04

Peter Koerber <pk(a)ExperTeam.de> wrote: > i try to compile the *.cpp files of boost::filesystem within the > following environment: > > os: Windows XP > compiler: Visual C++ Toolkit 2003 (this is the free version > of MS Visual C++ V7.1) > boost: Version 1.31.0 > > If BOOST_WINDOWS is defined, the file windows.h is included > in the sources: > * exception.cpp > * operations_posix_windows.cpp. > > Otherwise the file errno.h is included. > > BOOST_WINDOWS is defined within the environment mentioned above, > but the compiler does not have a windows.h include file. Hence > compilation fails. <snip> You need to install the Windows Platform SDK as well as the Toolkit.

1 0

problems using VC 7.1 and boost regex library
by Keith Smith 22 Apr '04

22 Apr '04

Sorry if this is a resend, but not sure if the original got sent! I tried your suggestion, and it appears that I still have the same errors. Anyone have any other ideas? (See Attached) -----Original Message----- From: boost-users-bounces(a)lists.boost.org [mailto:boost-users-bounces@lists.boost.org] On Behalf Of John Maddock Sent: Thursday, April 15, 2004 5:45 AM To: boost-users(a)lists.boost.org Subject: Re: [Boost-users] problems using VC 7.1 and boost regex library > I have been trying to compile the following code with VC 7.1 ( which > compiled fine with VC 6 ) with no luck. > > std::string m_rBuffer; > boost::regex m_rStartExp; > > string::const_iterator rRawDataBegin = m_rBuffer.begin(); > string::const_iterator rRawDataEnd = m_rBuffer.end(); > > boost::cmatch rStartMessage; > > unsigned int flags = boost::match_default; > while ( boost::regex_search(rRawDataBegin, > rRawDataEnd, > rStartMessage, > m_rStartExp, > flags) ) > > Included is the compiler error report. The match-flags area distinct type, so use: boost::match_flag_type flags = boost::match_default; John. _______________________________________________ Boost-users mailing list Boost-users(a)lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users

2 1

Problems with installation of boost unit_test_framework
by Matthew Burgess 21 Apr '04

21 Apr '04

Hi folks, http://www.boost.org/libs/test/doc/components/unit_test_framework/compilati… states that I can include <boost/test/included/unit_test_framework.hpp> in order to avoid having to link against the unit_test_framework library. However, the files it #includes are located in the boost test library source directory which doesn't appear to be installed by the default bjam procedure (I just get '/opt/boost/lib' &'/opt/boost/include' directories after installing boost). If it helps to clarify things here's some output: --- bash-2.05b$ cat example1.cpp #include <boost/test/included/unit_test_framework.hpp> using boost::unit_test_framework::test_suite; void test_function() { BOOST_CHECK(2 == 1); // intentional failure } test_suite* init_unit_test_suite(int, char* []) { test_suite* test = BOOST_TEST_SUITE("Unit test example 1"); test->add(BOOST_TEST_CASE(&test_function)); return test; } bash-2.05b$ g++ -I /opt/boost/include/boost-1_31/ -L /opt/boost/lib/ -o test example1.cpp In file included from example1.cpp:2: /opt/boost/include/boost-1_31/boost/test/included/unit_test_framework.h pp:18:47: libs/test/src/execution_monitor.cpp: No such file or directory ... --- The only place I find the required files is in the original location I extracted the boost source tarball to. I can (and have) copied the 'libs' directory to sit inside my 'include' directory and verified this fixes things, but I doubt very much this is the correct solution. I'm particularly concerned that if someone else wants to compile my toy code, simply telling them to "install boost" might not be enough. Related to all this is the issue of best-practice. Should I make efforts not to have to link to a library, what are the trade-offs? Can I simply take a copy of the required boost headers and place them directly in my project's directory structure? If not, is anyone aware of an autoconf style macro that can be used to detect the presence of the boost headers and/or libraries? I took a quick glance at the autoconf-archive but didn't see anything immediately. Thanks for anyone who can hit me with a cluebat (or FAQ pointer). Matt Burgess

1 0

problems using VC 7.1 and boost regex library ( this time with attachment )
by Keith Smith 21 Apr '04

21 Apr '04

1 0

problems using VC 7.1 and boost regex library
by Keith Smith 21 Apr '04

21 Apr '04

I tried your suggestion, and it appears that I still have the same errors. Anyone have any other ideas? (See Attached) -----Original Message----- From: boost-users-bounces(a)lists.boost.org [mailto:boost-users-bounces@lists.boost.org] On Behalf Of John Maddock Sent: Thursday, April 15, 2004 5:45 AM To: boost-users(a)lists.boost.org Subject: Re: [Boost-users] problems using VC 7.1 and boost regex library > I have been trying to compile the following code with VC 7.1 ( which > compiled fine with VC 6 ) with no luck. > > std::string m_rBuffer; > boost::regex m_rStartExp; > > string::const_iterator rRawDataBegin = m_rBuffer.begin(); > string::const_iterator rRawDataEnd = m_rBuffer.end(); > > boost::cmatch rStartMessage; > > unsigned int flags = boost::match_default; > while ( boost::regex_search(rRawDataBegin, > rRawDataEnd, > rStartMessage, > m_rStartExp, > flags) ) > > Included is the compiler error report. The match-flags area distinct type, so use: boost::match_flag_type flags = boost::match_default; John. _______________________________________________ Boost-users mailing list Boost-users(a)lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users

1 0

Selection scheme for boost libraries
by Uli Tuerk 19 Apr '04

19 Apr '04

Hi all! I have a very basic problem with the boost libraries: I used bjam to build all boost libraries for Linux (using "gcc" as toolset and the call 'bjam "-sTOOLS=gcc" install'). All went fine, however I wanted to do the regression tests for boost_regex using the gcc.mak Makefile in <boost_main>/libs/regex/test/regress which uses "-lboost_regex" for linking. The linker reported an error because there is no "libboost_regex.so" in my lib folder unless I created a link with this name to one of the various versions of libboost_regex. A comment for these Makefiles would be helpful for the case when having build all with bjam. When compiling the regex part alone, the library is created in build/gcc with the correct name (libboost_regex.so). I'm new to C++, so maybe I haven't got the point with the sophisticated naming scheme of the libraries, but: for the library "libboost_regex-gcc-mt-d-1_31.so" e.g. are the tool specifier ("gcc") and the version ("1_31") really necessary for a platform using version numbering of the libraries anyway in the ".so" extension? OK, there are still the categories with/without debug and with/without threading. How would you guys select the type of library? In the Makefile? Thanks in advance! Uli ************************************************************************ Uli Türk

2 1

BGL assertion
by Val Samko 18 Apr '04

18 Apr '04

Something triggers the assert line in push_relabel_max_flow, assert(algo.is_flow()); is that internal push_relabel_max_flow problem, or this means I supply invalid data to this function? Val Samko http://val.digiways.com

1 0

Problems compiling boost.python with Boost 1.31.0
by Markus Sch�pflin 17 Apr '04

17 Apr '04

I'm trying to compile boost.python from Boost 1.31.0 on Tru64 5.1 with Compaq C++ V6.5-033. Python has been built like this: ./configure --with-cxx=cxx --prefix=/opt --exec-prefix=/opt/exec/OSF1-V5.1-a lpha --without-gcc The command line used to build boost.python looks like this: bjam -sTOOLS=tru64cxx65 -sPYTHON_ROOT="/opt" -sPYTHON_VERSION="2.3" stage When build boost.python I get the following error message: tru64cxx65-C++-action bin/boost/libs/python/build/libboost_python.so/tru64cxx65/debug/numeric.o cxx: Error: /net/camos/opt/include/python2.3/Python.h, line 8: #5 could not open source file "pyconfig.h" #include "pyconfig.h" ---------------------^ It looks like the exec prefix is not taken into account because the file is located here: /opt/exec/OSF1-V5.1-alpha/include/python2.3/pyconfig.h Is this not supported by the current boost build process or is there any option I missed? Or is there anything else I'm doing wrong? Thanks, Markus

3 3

broken link in boost.org FAQ
by Mike Morearty 17 Apr '04

17 Apr '04

Minor problem: on the FAQ at http://www.boost.org/more/faq.htm , the very first question is, "How do I download the libraries?" and the answer is, "See Download instructions." But "Download" is a broken link. Points to http://www.boost.org/libs/libraries.htm#Download ; although http://www.boost.org/libs/libraries.htm does exist, it does not contain an anchor called "Download". The FAQ should probably be changed to point to http://www.boost.org/more/getting_started.html , or maybe to http://sourceforge.net/project/showfiles.php?group_id=7586 - Mike Morearty

1 0

[MPL] Runtime Interaction
by Debian User 16 Apr '04

16 Apr '04

[Sent via gmane but it never appeared] Hi gang, Suppose I have a number of types that each have a "value" enum, much like MPL's constant classes. Now suppose that I want to construct a runtime-indexable array of those values, according to the position of each type in an MPL vector. For example: int values[3]; typedef boost::mpl::vector_c<int,10, 11, 12> values; typedef some_magic_metafunction<values>::type runtime_function; // values[0] = 10, values[1] = 11, values[2] = 12 runtime_function::fill_array(values); The idea is that some_magic_metafunction<values> would generate a type kind of like this: template<typename TypeValueList, typename Index> struct magic { static void fill_array(int *v) { // Non-MPL type vector indexing -- could use an // MPL iterator here v[Index::value] = TypeValueList<Index::value>::value; if (Index::value < distance<begin<TypeValueList>::type, end<TypeValueList>::type>::type::value) { magic<TypeValueList, Index::next>::fill_array(v); } }; }; Get it? The problem is that I don't know how to make MPL transition from strictly compile-time to something in-between. It can statically know the size of the value array and the values themselves but it has to fill the array "values" at runtime so that the values may be indexed by runtime computations. Any guidance? Thanks! -Dave

2 1

problems using VC 7.1 and boost regex library
by Keith Smith 15 Apr '04

15 Apr '04

I have been trying to compile the following code with VC 7.1 ( which compiled fine with VC 6 ) with no luck. std::string m_rBuffer; boost::regex m_rStartExp; string::const_iterator rRawDataBegin = m_rBuffer.begin(); string::const_iterator rRawDataEnd = m_rBuffer.end(); boost::cmatch rStartMessage; unsigned int flags = boost::match_default; while ( boost::regex_search(rRawDataBegin, rRawDataEnd, rStartMessage, m_rStartExp, flags) ) Included is the compiler error report. Please throw me a bone.

2 1

BGL linking on Win32
by maxx＠dgap.mipt.ru 15 Apr '04

15 Apr '04

Hi! whenever I use libbgl-viz.lib to read graphviz files, I get a linker errors like this: msvcprt.lib(MSVCP71.dll) : error LNK2005: "public: __thiscall std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> >::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> > const &)" (??0?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@QAE@ABV01@@Z) already defined in Graphviz.obj It was Ok to add "/NODEFAULTLIB:msvcrt" to the compiler options, but now I need that library! So if I don't add this option, I get errors like the one above, if I do add it, I get errors "unresolved symbols"... What is probably(or definitely) wrong? Thanks in advance, -- Maxim

2 1

RE: [Boost-users] Changing the heap allocator
by Jim.Hyslop 14 Apr '04

14 Apr '04

David Barrett-Lennard wrote: > Yes HeapSetInformation() will do it once you have the heap handle. > However, the CRT creates its own heap with a call to HeapCreate(), and > stores the heap handle in a global variable called _crtheap > that is not > exported by the library. Right, of course - I thought that GetProcessHeap() would return the correct heap, but it does not. -- Jim Hyslop Senior Software Designer Leitch Technology International Inc. (http://www.leitch.com) Columnist, C/C++ Users Journal (http://www.cuj.com/experts)

1 0

RE: [Boost-users] Changing the heap allocator
by Jim.Hyslop 14 Apr '04

14 Apr '04

David Barrett-Lennard wrote: > Does anyone know how to enable the LFH (low fragmentation > heap) for the > heap created by the VC CRT? Have you tried HeapSetInformation()? -- Jim Hyslop Senior Software Designer Leitch Technology International Inc. (http://www.leitch.com) Columnist, C/C++ Users Journal (http://www.cuj.com/experts)

2 1

Changing the heap allocator
by David Barrett-Lennard 13 Apr '04

13 Apr '04

Hi, I'm working on a project where I've replaced Microsoft's (terrible) heap allocator with the Doug Lea allocator. There seems to be some cases where the Windows XP allocator is 500 times slower than it should be. This made my program take minutes rather than seconds to run. To try to understand what kind of allocation patterns cause the behaviour I wrote a small console application (see below). After playing around with different allocation patterns for half an hour I found a case where the pathological behaviour appeared. I'm none the wiser about the problem. For example, moving the malloc(301) statement before the 7 x malloc(1000) makes the problem go away! However, given the ease with which I rediscovered the problem, I wonder why there aren't more complaints about Microsoft's heap allocator. With some allocation patterns I have seen the allocator fall below 1kHz on a Pentium IV 2GHz machine! I've created a library that exports alternative versions of operator new and operator delete, by using a def file as follows EXPORTS ??2@YAPAXI@Z ??3@YAXPAX@Z I've had to do some hacky things to make it work with the STL (because of msvcp60.dll) , but that's another story! Now my question : I want to use boost, but need to work out how to make it use a different allocator. Will I need to rebuild the boost dlls (such as boost_python.dll), against my static library? Regards, David Barrett-Lennard /////////////////////// heap allocation test /////////////////////////// #include <iostream> #include <time.h> int main(int argc, char* argv[]) { for (int count=1 ; count <= 5 ; ++count) { int n1 = 0, n2 = 0; clock_t start = clock(); { const int NUM = 10000; void* L[NUM*7]; for (int i=0 ; i < NUM ; ++i) { for (int j=0 ; j < 7 ; ++j) L[n1++] = malloc(1000); void* leak = malloc(301); ++n2; } for (int k=n1-1 ; k >= 0 ; --k) free(L[k]); } clock_t elapsed = clock()-start; if (elapsed) { std::cout << "Time = " << elapsed << " ms" << " Rate = " << 1000 * (n1 + n2) / elapsed << " Hz" << std::endl; } } return 0; }

3 4

Using boost::regex 1.31.0 on vc6
by Chris Branch 10 Apr '04

10 Apr '04

Hi, I'm trying to use boost::regex 1.31.0 (the v4 source code) on MSVC++6 (sp5), but whenever I compile a project using the new code I get an internal compiler error in perl_matcher_common.hpp and no object is built. I want to use the new v4 code (the older v3 code works, and I'll use that if I have to) but this file is quite important... is there something I can do, or does it not work with VC6? Thanks, Chris

2 2

boost::thread and boost::bind
by GoochRules! 10 Apr '04

10 Apr '04

I am attempting to do this in vs.net 2003: int f(HANDLE, HANDLE, HANDLE); int main(int argc, char* argv[]) { HANDLE x; HANDLE y; HANDLE z; boost::thread* p = new boost::thread(boost::bind(f, x, y, z); p->join(); delete p; } As a contrived example it works. In my project it does not, spewing out 5 errors: 1) c:\src\include\boost\bind.hpp(62) : error C2825: 'F::result_type': cannot form a qualified name c:\src\include\boost\bind\bind_template.hpp(16) : see reference to class template instantiation 'boost::_bi::result_traits<R,F>' being compiled with [ R=boost::_bi::unspecified, F=int (__cdecl *)(HANDLE,HANDLE) ] c:\src\ansa-1.0\src\main\main.cpp(369) : see reference to class template instantiation 'boost::_bi::bind_t<R,F,L>' being compiled with [ R=boost::_bi::unspecified, F=int (__cdecl *)(HANDLE,HANDLE), L=boost::_bi::list3<boost::_bi::list_av_3<fd_xpt,fd_xpt,fd_xpt>:: B1,boost::_bi::list_av_3<fd_xpt,fd_xpt,fd_xpt>::B2,boost::_bi:: list_av_3<fd_xpt,fd_xpt,fd_xpt>::B3> ] 2) c:\src\include\boost\bind.hpp(62) : error C2039: 'result_type' : is not a member of 'operator``global namespace''' 3) c:\src\include\boost\bind.hpp(62) : error C2146: syntax error : missing ';' before identifier 'type' 4) c:\src\include\boost\bind.hpp(62) : error C2955: 'boost::_bi::type' : use of class template requires template argument list c:\src\include\boost\bind.hpp(94) : see declaration of 'boost::_bi::type' 5) c:\src\include\boost\bind.hpp(62) : fatal error C1903: unable to recover from previous error(s); stopping compilation note: fd_xpt is declared as "typedef HANDLE fd_xpt;", and c:\src\ansa-1.0\src\main\main.cpp(369) is similar to "boost::thread* p = new boost::thread(boost::bind(f, x, y, z);" in the example. It appears that the compiler thinks my first argument is a "int (__cdecl *)(HANDLE,HANDLE)" when it is most certainly a "int (__cdecl *)(HANDLE,HANDLE,HANDLE)". It used to take two arguments, but I've since added a third (it was at this very point that these errors crashed the party). I am sure that there is no declaration that specifies only two arguments: I've had two other developers verify this. I have cleaned my project multiple times, even by hand, as well as getting a clean source tree from cvs, and yet I still get these errors. Any suggestions on how to silence these errors are greatly appreciated. --mpeltzer goochrules(a)hotmail.com

3 3

Re: [boost] Re: [python] external references to python details
by Domenico Andreoli 10 Apr '04

10 Apr '04

hi, Lars Gullik Bj�nnes writes: > David Abrahams <dave(a)boost-consulting.com> writes: > > | larsbj(a)gullik.net (Lars Gullik Bj�nnes) writes: > > > >> I just discovered that boost libraries (or only one...) depend on > >> files in python/detail. That should probably be fixed? > > > | Got a suggested patch? > > Not really. > > With the subset I use from boost moving indirect_traits.hpp and > is_xxx.hpp from boost/python/detail/ to boost/detail/ (and fixup > #includes) seems like the simple fix. here at Debian we have the same problem, please see bug #242801 [0]. this suggestion could would fix also our situation. any news? thanks domenico [0] http://bugs.debian.org/242801 -----[ Domenico Andreoli, aka cavok --[ http://filibusta.crema.unimi.it/~cavok/gpgkey.asc ---[ 3A0F 2F80 F79C 678A 8936 4FEE 0677 9033 A20E BC50

1 0

template
by adel.essafi 08 Apr '04

08 Apr '04

hi list he is the source code of my project. it is an algorithms for scheduling DAG taks on machines. there is a template function that does not work well. please , can you guess where is the fault. thanks Accédez au courrier électronique de La Poste : www.laposte.net ; 3615 LAPOSTENET (0,34/mn) ; tél : 08 92 68 13 50 (0,34/mn)

1 0

template
by adel.essafi 08 Apr '04

08 Apr '04

hi list I am developping a scheduling algorithm. here you have the source code of my project. please , can you see what is wrong with it. thanks Accédez au courrier électronique de La Poste : www.laposte.net ; 3615 LAPOSTENET (0,34/mn) ; tél : 08 92 68 13 50 (0,34/mn)

1 0