October 2005 - Boost-users - lists.preview.boost.org

using uBLAS with quantity<> (C++TMP)
by Malte Clasen 17 Oct '05

17 Oct '05

Hi, the book "C++ Template Metaprogramming" by Abrahams and Gurtovoy describes a dimensional analyis framework that enables compile-time checking of physical units. For example dividing a length quantity by a time quantity results in a speed quantity. Now it's tempting to use boost uBLAS with these quantities: A position in space can be expressed as a vector of length quantities etc. But there's one problem: "// uBLAS assumes a common return type for all binary arithmetic operators" (traits.hpp, line 42, boost release 1.33) The promote_traits class relies on this when computing the following: static const std::size_t size = sizeof ( type_deduction_detail::test< typename base_type::x_type , typename base_type::y_type >(x + y) // Use x+y to stand of all the arithmetic actions ); This fails for most quantity operations since dividing a length by a length results in a different type than adding them. As a result Visual C++ 7.1 complains about not being able to deduce template arguments. Is there a way to fix this? I'm not that familiar with the internals of uBLAS, I don't even know the purpose of promote_traits. Malte

2 1

date_time: Print Holidays Example
by __PPS__ 17 Oct '05

17 Oct '05

Aren't you supposed to delete new'd objects, or you simply rely that nobody is gonna look/use/copy this piece of code? http://tinyurl.com/8wanl

2 2

Failing to compile Boost.Threads 1.33.0 with Borland C++
by Jarl Lindrud 16 Oct '05

16 Oct '05

Hi, I've just upgraded from 1.32.0 to 1.33.0, and now when I try to build Boost. Threads with Borland C++, I just get a lot of errors, as shown below. I've compiled with several other compilers, and it all went fine. I'm using Borland C++ 5.6, the one that comes with Borland C++Builder 6, and I have no problems building Boost.Threads from the 1.32.0 distribution . The regression test table seems to indicate it should work, so what am I missing? Thanks. ************************************************** E:\Development\boost_1_33_0\libs\thread\build>..\..\..\..\bjam.1.33.0.exe -sTOOL S=borland ...found 556 targets... ...updating 20 targets... borland-C++-action ..\..\..\bin\boost\libs\thread\build\boost_thread.dll\borland \debug\threading-multi\exceptions.obj ..\src\exceptions.cpp: Error E2090 ..\src\exceptions.cpp 71: Qualifier 'thread_exception' is not a clas s or namespace name Error E2040 ..\src\exceptions.cpp 71: Declaration terminated incorrectly Error E2190 ..\src\exceptions.cpp 187: Unexpected } *** 3 errors in Compile *** "H:\Program\Borland\CBuilder6\bin\bcc32" -j5 -g255 -q -c -P -w -Ve -Vx -a8 -b- -DBOOST_THREAD_LIB_NAME=boost_thread -DBOOST_THREAD_BUILD_DLL=1 -v -Od -vi - -tWR -tWD -WM- -tWM -w-8004 -w-8057 -w-8001 -I"..\..\..\bin\boost\libs\threa d\build" -I"H:\Program\Borland\CBuilder6\include" -I"E:\Development\boost_1_33 _0" -o"..\..\..\bin\boost\libs\thread\build\boost_thread.dll\borland\debug\thre ading-multi\exceptions.obj" "..\src\exceptions.cpp" ...failed borland-C++-action ..\..\..\bin\boost\libs\thread\build\boost_thread.d ll\borland\debug\threading-multi\exceptions.obj... borland-C++-action ..\..\..\bin\boost\libs\thread\build\boost_thread.dll\borland \debug\threading-multi\mutex.obj ..\src\mutex.cpp: Error E2312 ..\src\mutex.cpp 44: 'm_critical_section' is not an unambiguous base class of 'mutex' in function mutex::mutex() Error E2451 ..\src\mutex.cpp 46: Undefined symbol 'm_critical_section' in functi on mutex::mutex() Error E2451 ..\src\mutex.cpp 55: Undefined symbol 'm_critical_section' in functi on mutex::~mutex() Error E2451 ..\src\mutex.cpp 63: Undefined symbol 'm_critical_section' in functi on mutex::do_lock() Error E2451 ..\src\mutex.cpp 71: Undefined symbol 'm_critical_section' in functi on mutex::do_unlock() Error E2228 ..\src\mutex.cpp 71: Too many error or warning messages in function mutex::do_unlock() *** 6 errors in Compile *** "H:\Program\Borland\CBuilder6\bin\bcc32" -j5 -g255 -q -c -P -w -Ve -Vx -a8 -b- -DBOOST_THREAD_LIB_NAME=boost_thread -DBOOST_THREAD_BUILD_DLL=1 -v -Od -vi - -tWR -tWD -WM- -tWM -w-8004 -w-8057 -w-8001 -I"..\..\..\bin\boost\libs\threa d\build" -I"H:\Program\Borland\CBuilder6\include" -I"E:\Development\boost_1_33 _0" -o"..\..\..\bin\boost\libs\thread\build\boost_thread.dll\borland\debug\thre ading-multi\mutex.obj" "..\src\mutex.cpp" ...failed borland-C++-action ..\..\..\bin\boost\libs\thread\build\boost_thread.d ll\borland\debug\threading-multi\mutex.obj... borland-C++-action ..\..\..\bin\boost\libs\thread\build\boost_thread.dll\borland \debug\threading-multi\recursive_mutex.obj ..\src\recursive_mutex.cpp: ...interrupted **************************************************

1 0

Which algorith should I use(Boost::reg_ex)?
by Aaron 16 Oct '05

16 Oct '05

I'm writting a program which could evaluate macros in a c++ header file, like this: #define CMD_IS_BASE 0x00001000 #define CMD_GET_ALL_DOMAINNAME (CMD_IS_BASE+0x1) /* Get all domain name from clients*/ After parsing is done, if you input a value(int), the program could tell you the symbol name, symbol meaning (if there is) corresponding to the valude. As to the above example, when you input 1025(or 0x00010001), the program should tell you it's correspond to symbol CMD_GET_ALL_DOMAINNAME, with it's detail comments "get all domain name from clients". I want also this program could parse the following paragraph too: // // MessageId: GL_CAT_SUMMARY // // MessageText: // // Summary%0 // #define GL_CAT_SUMMARY 0x00000002L It's easy to define a regex pattern, then call boost::regex_grep to got all these symbols and their meaning. But what if two formats exist in one header file? I wrote two grepper plug-ins(one is called singleline.dll, the other is McHearder.dll) to find these symbols, which re pattern contains more lines has the high pirioty to be called, then after it's called, the matched string should be erased or be filled with meaningless chars to prevent the later greppers to work on it. See example: The original string: // // MessageId: GL_CAT_SUMMARY // // MessageText: // // Summary%0 // #define GL_CAT_SUMMARY 0x00000002L #define CMD_IS_BASE 0x00001000 First we call McHeader.dll to work on this string, then it grepped: symble:GL_CAT_SUMMARY value: 2 comments:Summary Then the string should be only one lines, ie: #define CMD_IS_BASE 0x00001000 So, how could I achieve that? Could I still use boost::regex_grep? Thanks in advance!

2 1

Build boost using 64 bit flags on hp-ux 11.23 itanium
by Naeem ahmad 15 Oct '05

15 Oct '05

Do any body how to build boost 64 bit version I have tried to use this command to build, but it does not work bjam cflags="-m64" linkflags="-m64" cxxflags="-m64" "-sTOOLS=gcc" install the 32 bit version build ok by default. Naeem Ahmad

2 1

Re: [Boost-users] Errors compiling greg_month.cpp
by D. R. Evans 15 Oct '05

15 Oct '05

A couple of days ago I requested help (below) as to why greg_month.cpp in 1.33.0 was giving errors under VC++ 6.0. Seeing no response, I guess I'll just go back to 1.30.2, which was the version that has been working fine. Sorry to have troubled you. Doc ----- I just upgraded to 1.33.0, and am getting a bunch of compilation errors with the current version of greg_month.cpp: The list of errors starts with: c:\boost\boost_1_33_0\boost\date_time\gregorian\formatters.hpp(34) : error C2977: 'date_formatter' : too many template arguments c:\boost\boost_1_33_0\boost\date_time\date_formatting_limited.hpp(114) : see declaration of 'date_formatter' c:\boost\boost_1_33_0\boost\date_time\gregorian\formatters.hpp(40) : see reference to function template instantiation 'class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> > __cdecl boost::gregorian::to_simpl e There are 31 errors in all. Using VC++ 6.0 on XP Service Pack 2. I would appreciate any gentle pointers as to what I might be doing wrong. Doc

3 2

Exception Stack Trace Question.
by Piyush Kapadia 14 Oct '05

14 Oct '05

Though catch(.) statement allows catching exception not associated with any class, it leads to ambiguous condition as there is no way to have any error message with it. Does Boost provide any means to be able to print stack trace when exception is caught ? In spite of many deficiencies with Java, it let's you access stack trace using printStackTrace() and trace out as what happened and how it happened. This is where C++ std::exception is lacking support and one has to look for such improvements from Boost community.

4 3

[spirit] error removal
by Aaron Griffin 14 Oct '05

14 Oct '05

Hi all, This is a contrived example, but the solution should be the same. Assuming I have two rules, like so: section = identifier >> ch_p('{') >> *setting >> ch_p('}'); setting = identifier >> ch_p('=') >> value; // identifier and value are simple, but left off for brevity This should parse something like: somesettings { a = 12 b = "hello whirled" } Now, what I am looking to do is, if the parser fails parsing a value inside a "section", the parsing itself does not fail, but that section itself does. So if I also had: moresettings { this_is_valid = 12345 bigfaterror = 12"hi there"4.3"fewewfewfw" } Then the "somesettings" section would exist in the final AST, but "moresettings" would not. The error handling docs are a bit unclear to me, bit if I were to make a generic functor_parser which kept an error flag, how would I conditionally remove the section based on that value, seeing as it will already be mid-parse (I don't think something like if_p will work). Thanks in advance for any advice. On a side note, if I have a rule similar to (keyword >> identifier), the skip_parser will translate "somekeyword a" into "somekeyworda" and will parse it fine. However, what is to stop someone from removing the whitespace in the original code. With a quick glance, it appears the example c grammar from the spirit site will parse "voidfoo()" just fine due to this. Is there anyway I can tell the skip grammar to force a seperation? Thanks again, Aaron Griffin

3 4

Re: [Boost-users] Compilation errors with Intel Compiler 9.0 for Linux
by Olivier Delalleau 14 Oct '05

14 Oct '05

On 14 Oct 2005, boost-users-request(a)lists.boost.org wrote: > > There are many errors like this one. Again, if I remove the include > > command -I/opt/intel/compiler90/include/c++ > > it works fine. Also if I use the intel 8.0 includes it works fine (but > > not intel 8.1). > > > > Any help fixing these errors (actually, mainly the regex one, as this > > is the library I really need) will be much appreciated :) > > I believe this is fixed for 1.33.1, can you get hold of that from cvs (it's > on the RC_1_33_0 branch), or would you like a zip to test? I'd like to be > sure this really is fixed for the next release as you're not the only one > with the problem. Thank you :) I got the RC_1_33_0 release from the cvs, and regex is now compiling perfectly with the Intel 9.0 includes. Note that the iostreams library is still broken, but I don't really care that much since I don't plan to use it in the near future. Thanks again for the help! -- Olivier

1 0

Compilation errors with Intel Compiler 9.0 for Linux
by Olivier Delalleau 14 Oct '05

14 Oct '05

I'm getting some annoying errors when compiling some of the libraries of Boost 1.33.0 with Intel Compiler 9.0 (icpc (ICC) 9.0 20050811) under Linux. Here is what happens when I try to compile (again after a first try which failed to update all targets): ----------------------------------------------------------------------------------- ../boost-jam-3.1.11-1-linuxx86/bjam -sTOOLS=intel-linux --prefix=/RQusagers/delallea/local --builddir=/RQexec/delallea/boost_1330_90 -sPYTHON_VERSION=2.4 --with-python-root=/usr/network.ALTIX/python-2.4.1 -sINTEL_VERSION=90 -sINTEL_PATH=/opt/intel/compiler90 install Building Boost.Regex with the optional Unicode/ICU support disabled. Please refer to the Boost.Regex documentation for more information (and if you don't know what ICU is then you probably don't need it). ...found 14461 targets... ...updating 104 targets... intel-linux-C++-action /RQexec/delallea/boost_1330_90/bin/boost/libs/iostreams/build/libboost_iostreams.a/intel-linux/debug/file_descriptor.o /RQMNT/exec1/lisa/delallea/boost_src/boost_1_33_0/boost/iostreams/positioning.hpp(45): error: no suitable conversion function from "mbstate_t" to "std::streampos={std::streamoff={long}}" exists return std::streampos(std::mbstate_t(), off); ^ /RQMNT/exec1/lisa/delallea/boost_src/boost_1_33_0/boost/iostreams/positioning.hpp(45): error: expected a ")" return std::streampos(std::mbstate_t(), off); ^ /RQMNT/exec1/lisa/delallea/boost_src/boost_1_33_0/boost/iostreams/positioning.hpp(51): error: no suitable conversion function from "fpos_t" to "boost::iostreams::stream_offset={intmax_t={long}}" exists return pos; ^ /RQMNT/exec1/lisa/delallea/boost_src/boost_1_33_0/boost/iostreams/positioning.hpp(64): error: expression must have class type return fpos_t_to_offset(pos.seekpos()) + ^ /RQMNT/exec1/lisa/delallea/boost_src/boost_1_33_0/boost/iostreams/positioning.hpp(66): error: expression must have class type stream_offset(std::streamoff(pos.seekpos())); ^ compilation aborted for /RQMNT/exec1/lisa/delallea/boost_src/boost_1_33_0/libs/iostreams/build/../src/file_descriptor.cpp (code 2) . /opt/intel/compiler90/bin/iccvars.sh icpc -c -w1 -g -O0 -I"/RQexec/delallea/boost_1330_90/bin/boost/libs/iostreams/build" -I"/RQMNT/exec1/lisa/delallea/boost_src/boost_1_33_0" -I"/RQMNT/exec1/lisa/delallea/boost_src/boost_1_33_0" -o "/RQexec/delallea/boost_1330_90/bin/boost/libs/iostreams/build/libboost_iostreams.a/intel-linux/debug/file_descriptor.o" "/RQMNT/exec1/lisa/delallea/boost_src/boost_1_33_0/libs/iostreams/build/../src/file_descriptor.cpp" ...failed intel-linux-C++-action /RQexec/delallea/boost_1330_90/bin/boost/libs/iostreams/build/libboost_iostreams.a/intel-linux/debug/file_descriptor.o... ----------------------------------------------------------------------------------- Then other similar errors in iostreams... Now, the important detail: I have in my icpc.cfg the following line: -I/opt/intel/compiler90/include/c++ which forces the Intel compiler to (preferably) use its own includes. Otherwise it uses the g++ standard includes and (at least for the file above) compilation works fine. But I would really prefer to be able to compile Boost using only the Intel code (especially since I need this line in my icpc.cfg for some other non-Boost stuff to compile). Then when compiling the regex library I have other errors. Here is an example of what happens with a specific compile line (here I cleaned my icpc.cfg file): ----------------------------------------------------------------------------------- icpc -I/opt/intel/compiler90/include/c++ -c -w1 -DNDEBUG -DNDEBUG -O3 -openmp -I"/RQexec/delallea/boost_1330_90/bin/boost/libs/regex/build" -I"/RQMNT/exec1/lisa/delallea/boost_src/boost_1_33_0" -o "/RQexec/delallea/boost_1330_90/bin/boost/libs/regex/build/libboost_regex.a/intel-linux/release/threading-multi/instances.o" "/RQMNT/exec1/lisa/delallea/boost_src/boost_1_33_0/libs/regex/build/../src/instances.cpp" /RQMNT/exec1/lisa/delallea/boost_src/boost_1_33_0/boost/regex/v4/instances.hpp(84): error #803: function "boost::match_results<BidiIterator, Allocator>::match_results(const boost::match_results<BidiIterator, Allocator> &) [with BidiIterator=const char *, Allocator=std::vector<boost::sub_match<const char *>, std::allocator<boost::sub_match<const char *>>>::allocator_type]" explicitly instantiated more than once template class BOOST_REGEX_DECL match_results< std::basic_string<BOOST_REGEX_CHAR_T>::const_iterator >; ----------------------------------------------------------------------------------- There are many errors like this one. Again, if I remove the include command -I/opt/intel/compiler90/include/c++ it works fine. Also if I use the intel 8.0 includes it works fine (but not intel 8.1). Any help fixing these errors (actually, mainly the regex one, as this is the library I really need) will be much appreciated :) -- Olivier

2 1