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

Compile problem with shared_ptr on unusual platform
by Allen Gooch 07 Nov '05

07 Nov '05

I'm trying to use shared_ptr<> from boost 1.33.0 on the Metrowerks CodeWarrior for PSP 1.2 (that's Playstation Portable) and am running into a problem, which I believe may be related to the boost config system and it not being prepared for my admittedly unusual toolchain. I have successfully used this on many other platforms in the past, and have thoroughly enjoyed the boost in productivity and safety through its use. Here's my problem, a simple "hello world" app using boost::shared_ptr< std::string > using the top-level <boost/shared_ptr.hpp> include yields the following link error: Link Error : Undefined : "boost::throw_exception(const std::exception&)" Referenced from "boost::detail::shared_count::!<std::basic_string<char, std::char_traits<char>, std::allocator<char>>>(std::basic_string<char, std::char_traits<char>, std::allocator<char>>*)" in main.cpp Link failed. Any insight into this problem, or the configuration system and how I might adapt it to this compiler, are greatly appreciated. -allen

4 3

Missing Files in Boost 1.33.0 Download
by Merrill Cornish 07 Nov '05

07 Nov '05

Bug? As nearly as I can tell, boost_1_33_0/doc/html/thread created by running the boost_1_33_0.exe download is supposed to contain the same files as http://www.boost.org/doc/html/thread.html, which you get to from libraries | documentation on the home page. However, in my download (and for 1.32.0 also) there are only the acknowledgements, build, and configuration HTML files. Could someone check to see if the other files are being dropped in the download packaging step? Merrill

1 0

Problem with dll and BOOST_STATIC_ASSERT
by Frank 07 Nov '05

07 Nov '05

I have the following code ( very simplified but compilable) // =========== start snip =============================== /// ======= in hpp ============ #include <boost/static_assert.hpp> template <typename T, unsigned short DIM=3> class Vector { public: Vector(void) {/*init code omitted*/} Vector(const Vector& v) {/* ctor code omitted */} //===================================================== // some ctors for small DIM's // will cause compile time errors if DIM != # of args //===================================================== #ifdef enable_small_dim_ctor explicit Vector( T x ) { BOOST_STATIC_ASSERT(DIM==1); data_[0]=x; } Vector( T x, T y ) { BOOST_STATIC_ASSERT(DIM==2); data_[0]=x; data_[1]=y; } Vector( T x, T y, T z ) { BOOST_STATIC_ASSERT(DIM==3); data_[0]=x; data_[1]=y; data_[2]=z; } #endif private: T data_[DIM]; }; template <typename T> class Vector3 : public Vector<T,3> { public: Vector3(){}; Vector3 Normal() const; }; //==================== in cpp ================= // provide implementation of Normal template <typename T> Vector3<T> Vector3<T>::Normal() const { Vector3 n; return n; } // explicit instantiation of the normal method #ifdef try_export // causes error STATIC_ASSERTION_FAILURE template class __declspec(dllexport) Vector3<double> Vector3<double>::Normal() const; #else //OK template class Vector3<double> Vector3<double>::Normal() const; #endif // ============== end snip =============================== I 'm trying to put the implementation into a dll and mark the explicit instantiation of Normal as exported with the following results: 1) With the small_dim_ctor's enabled I get the static assertion when marked for export 2) With the small_dim_ctor's enabled everything is ok when not marked for export 3) With the small_dim_ctor's disabled everything is ok regardless of export marking 4) If I comment out the BOOST_STATIC_ASSERT and build my app with export enabled I can use the resulting dll without any problems I'm thinking that marking for export is somewhow generating an implicit conversion somewhere but I don't know how. Using VStudio .Net 2003. Any help is appreciated. Thanks Frank

3 8

[iostreams] problem with gzip_compressor
by Tiago de Paula Peixoto 07 Nov '05

07 Nov '05

Hello. I've been experiencing the following problem with the gzip_compressor filter: It always writes data down the stream, regardless if any data has been written to the filtering_stream. Consider the following program: -------------------------------------------------------- #include <iostream> #include <fstream> #include <boost/iostreams/filtering_stream.hpp> #include <boost/iostreams/filter/gzip.hpp> using namespace std; using namespace boost; int main() { ofstream file; file.open("hello.gz", ios_base::out | ios_base::app | ios_base::binary); iostreams::filtering_stream<boost::iostreams::output> stream; stream.push(iostreams::gzip_compressor()); stream.push(file); return 0; } -------------------------------------------------------- If compiled as "test", the following behavior is observed: $ ./test; ls -l hello.gz -rw-r--r-- 1 count0 users 8 2005-10-30 19:51 hello.gz $ ./test; ls -l hello.gz -rw-r--r-- 1 count0 users 16 2005-10-30 19:51 hello.gz $ ./test; ls -l hello.gz -rw-r--r-- 1 count0 users 24 2005-10-30 19:51 hello.gz That is, it always adds 8 bytes to the file, despite the fact that nothing was written to the stream. To top things off, the data inside the file isn't even recognizable by gzip: $ gunzip hello.gz gunzip: hello.gz: not in gzip format When bzip2_compressor is used, this strange behavior doesn't happen, and the file remains with 0 bytes. I guess this is a bug, right? I'm using boost 1.33.0 with GCC 4.0.1 in a GNU/Linux system. Thanks for any help. -- Tiago de Paula Peixoto <tiago(a)forked.de>

2 6

linker problem with regex library
by Oliver.Kowalke＠infineon.com 07 Nov '05

07 Nov '05

Hello, I get the linker error (see below) if I try to compile my test app. What is the problem? /usr/ccs/bin/ld: Invalid loader fixup in text space needed in output file for symbol "__gxx_personality_sj0" in input file "/opt/boost/lib/libboost_regex-gcc. Best regards, Oliver

2 1

Re: [Boost-users] linker problem with regex library
by Oliver.Kowalke＠infineon.com 07 Nov '05

07 Nov '05

Sorry - I forget: os. HP/UX 11.11, gcc 3.4.2 >Hello, >I get the linker error (see below) if I try to compile my test app. >What is the problem? > >/usr/ccs/bin/ld: Invalid loader fixup in text space needed in output >file for symbol "__gxx_personality_sj0" in input file >"/opt/boost/lib/libboost_regex-gcc. > >Best regards, >Oliver > >_______________________________________________ >Boost-users mailing list >Boost-users(a)lists.boost.org >http://lists.boost.org/mailman/listinfo.cgi/boost-users

1 0

Can't Compile store() in Program Options
by Merrill Cornish 07 Nov '05

07 Nov '05

I am trying to use the Boost 1.32.0 program_options library for the first time with MGWin. Since what I need is identical to what is in the Getting Started section of the program_options Tutorial, I copied that. Here is the code from the tutorial: po::options_description desc("Allowed options"); desc.add_options() ("help", "produce help message") ("compression", po::value<int>(), "set compression level") ; po::variables_map vm; po::store(po::parse_command_line(ac, av, desc), vm); po::notify(vm); The problem occurs in the store() function. The compiler says it doesn't have a store() to match that call and suggests two possibilities. Examining the .hpp files, it appears that the two argument version of store() defaults to using wchar_t, which doesn't seem right. So, I tried po::store(po::parse_command_line(ac, av, desc), vm, true); to get UTF8 (simple char). The next complaint is apparently about the parse_command_line() call as the first argument. If I pull parse_command_line() out into a separate assignment statement to an intermediate variable, then put its results into store(), that could/would work except there now seems to be a conflict between the template types of char versus const char. I can get the assignment of parse_command_line() to an intermediate result to compile, but store() fails because of mismatched type problems. Or, I can get store() to compile, but the assignment fails for mismatched type problems. I apologize for not having the specific error messages here since the machine I'm compiling on is not available right now. Merrill P.S. About the Documentation: The program_options library document looks slick and very professional. Furthermore, it obviously took some time to create with all of the examples. However, I've never seen programming library documentation that didn't actually list the methods and functions in the library. As it is, if you are lucky, the method you need to mentioned, in passing, in an example. However, looking in the hpp and cpp files shows there is a lot more there.

2 1

Problem with program_options store()
by Merrill Cornish 07 Nov '05

07 Nov '05

Given this program fragment (based on the Getting Started example) using Boost::program_options compiled with MinGW: options_description desc("Program options"); ... variables_map vm; store(parse_command_line(argc, argv, desc), vm); gave the following error (reformatted for readability): error: no matching function for call to `store(boost::program_options::basic_parsed_options<const char>, boost::program_options::variables_map&)' candidates are: store(const boost::program_options::basic_parsed_options<char>&, boost::program_options::variables_map&, bool) store(const boost::program_options::basic_parsed_options<wchar_t>&, boost::program_options::variables_map&) I can add a "true" as a third argument to store(), but the mismatch still remains in the first argument: <const char> versus <char>&. If I assigned the result of parse_command_line() to an intermediate variable, I can match the &, but the const char versus char still remains. What is the problem way of calling store()? Merrill

2 1

Profiling Boost with gprof
by Dave 07 Nov '05

07 Nov '05

Hello All, I'm trying to profile Boost Graph examples using gprof, but can't get any reasonable results. My system is Mac OS X 10.3.9 and gcc version 3.3. I compile programs with -pg option, but the profile does not seem to contain any useful information. It only shows "no time accumulated", or 100% for main. What am I doing wrong? Thanks for your help! Dave __________________________________ Yahoo! Mail - PC Magazine Editors' Choice 2005 http://mail.yahoo.com

3 2

library planed for expression evaluation?
by Oliver.Kowalke＠infineon.com 07 Nov '05

07 Nov '05

Hello, does some plans for an library for expression evaluation exist? The user should be able to define its own operators (unary and binary with priority/affinity) and operands. The user can then store a sequence of its operators and operands in a vector in infix order. A special function creates from the vector an expression tree and evaluates the tree (calling the implemented action of each operator/operand -> the action can depend of the subsequent/child operators). With regards, Oliver

1 0