June 2003 - Boost-users - lists.preview.boost.org

MSVC 6.0 SP5: INTERNAL COMPILER ERROR
by Peter 12 Jun '03

12 Jun '03

Arghhhh.... c:\Development\Libraries\boost\boost_1_30_0\boost\iterator_adaptors.hpp(872) : fatal error C1001: INTERNAL COMPILER ERROR (compiler file 'msc1.cpp', line 1794) Please choose the Technical Support command on the Visual C++ Help menu, or open the Technical Support help file for more information c:\Development\Libraries\boost\boost_1_30_0\boost\date_time\date_parsing.hpp (27) : see reference to class template instantiation 'boost::iterator_adaptor<class boost::detail::token_iterator_base<char const *>,class boost::detail::tokenizer_policy<class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> >,class boost::char_delimiters_separator<char,struct std::char_traits<char> > >,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> > const &,class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> > const *,struct std::forward_iterator_tag,int>' being compiled

3 2

boost::condition and lock concept
by db10www 12 Jun '03

12 Jun '03

I have created a mutex class which is a model of the mutex concept. I also have lock classes for this mutex which models the lock concept (e.g. called CMyLock). (The reason for this is to implement a read/write mutex and associated read and write locks) According to the boost documentation I should be able to use this lock with the boost::condition but I get a compile time error:- 'm_mutex' : is not a member of 'CMyLock' This is due to the condition::wait method defined in condition.hpp making a reference to lock.m_mutex Surely boost::condition only needs a lock object to be an instance of a class which is a model of lock. The lock concept makes no reference to a member variable called m_mutex. Is this a bug of boost::condition or am I doing something wrong. Any ideas? Many thanks.

2 1

Struggling with bind
by Paul Thompson 12 Jun '03

12 Jun '03

Hi all. I've been using boost::bind to work around deficiencies with std::mem_fun_ref and std::bind2nd under MSVC6. I have a class called CDocumentAttachments which had a member function thus: HINSTANCE Execute(LPCTSTR verb) const; I was using bind with this as follows: for_each(filenames.begin(), filenames.end(), bind(&CDocumentAttachments::Execute, _1, _T("print"))); "filenames" is a const vector of CDocumentAttachments objects. This worked fine. A little later on in development, I changed the function signature for CDocumentAttachments::Execute to: HINSTANCE Execute(LPCTSTR verb, bool visible) const; How will I pass the extra bool to bind? I can't find any examples of passing more than one parameter to a member function. I thought about (among many others) for_each(filenames.begin(), filenames.end(), bind(&CDocumentAttachments::Execute, _2, _T("print"), false)); but the compiler (VC6) complains with two of the following: k:\vc\include\boost\bind.hpp(760) : error C2679: binary '[' : no operator defined which takes a right-hand operand of type 'const class boost::arg<2>' (or there is no acceptable conversion) k:\vc\include\boost\bind\bind_template.hpp(33) : see reference to function template instantiation 'struct HINSTANCE__ *__cdecl boost::_bi::evaluator3<struct HINSTANCE__ *>::eval(const class boost::_bi::list3<class boost::arg<2>,class boost:: _bi::value<char *>,class boost::_bi::value<bool> > &,struct boost::_mfi::cmf2<struct HINSTANCE__ *,class CDocumentAttachments,char const *,bool>,class boost::_bi::list1<class CDocumentAttachments const &> &)' being compiled I know I'm being really thick here, and I'm sure there's something really obvious that I'm missing, but reading through the documentation hasn't helped me. Thank you very much. Paul -- Paul Thompson Programmer McLellan Software Design Ltd.

2 3

Re: [crc] Simple example please
by Daryle Walker 12 Jun '03

12 Jun '03

On Tuesday, June 10, 2003, at 9:08 AM, Vladimir Prus wrote: [SNIP] > Here's the function that I use for the same purpose. Might not be > optimal, but works: > > unsigned file_crc(const string& name) > { > #if defined(__GNUC__) && __GNUC__ < 3 > ifstream ifs(name.c_str(), ios::binary); > #else > ifstream ifs(name.c_str(), ios_base::binary); > #endif > if (!ifs) > return 0; > else { > using namespace boost; > > crc_32_type crc32; > int c; > while( (c = ifs.rdbuf()->sbumpc()) != -1) > crc32.process_byte(char(c)); > return crc32.checksum(); > } > } > > > I believe some example of this kind should be included... Daryle, what > do you think? I'll consider it. [A few hours pass...] How about something like this: //================================================================ #include <cstdlib> #include <exception> #include <fstream> #include <ios> #include <iostream> #include <ostream> #include <boost/crc.hpp> #ifndef BLOCK_SIZE #define BLOCK_SIZE 1024 #endif int main ( int argc, char const * argv[] ) try { boost::crc_32_type result; // Loop over each file argument for ( int i = 1 ; i < argc ; ++i ) { std::ifstream ifs( argv[i], std::ios_base::binary ); if ( ifs/*std::ifstream ifs(argv[i], std::ios_base::binary)*/ ) { /*char buffer[ BLOCK_SIZE ]; std::streamsize len; while ( 0 < (len = ifs.readsome( buffer, BLOCK_SIZE )) ) { result.process_bytes( buffer, len ); }*/ char c; while ( ifs.get(c) ) { result.process_byte( c ); } } else { std::cerr << "Failed to open file '" << argv[i] << "'." << std::endl; } } std::cout << std::hex << result.checksum() << std::endl; return EXIT_SUCCESS; } catch ( std::exception &e ) { std::cerr << "Found an exception with '" << e.what() << "'." << std::endl; return EXIT_FAILURE; } catch ( ... ) { std::cerr << "Found an unknown exception." << std::endl; return EXIT_FAILURE; } //================================================================ Note in the "if" statement, I had the construction in it. However, neither one of my compilers took it! Moving the constructor call worked, but I thought my original method should be accepted. One compiler finished after the above 'fix'. The other (a variant of GCC 3.1) complained about missing symbols at link-time; it could not create the global objects hidden in Boost.CRC. (I haven't worked on the code in a long time; there could be quirks on compilers written after Boost.CRC.) When I ran the program under the compiler that worked, the "readsome" function always returned zero! There were no suitable substitutes; the "read" function doesn't tell you how many characters it read (bad if the last block isn't full) and the multi-character "get" functions use delimiters. So this program is stuck at single-character reads at the moment. Maybe I should switch to the stream-buffer method you used. Can you try this out to make sure the problems aren't just mine? Daryle

2 2

RegEx Windows vs Linux
by Mark Sizer 12 Jun '03

12 Jun '03

I compiling the same code on both platforms (same my code, same boost code). Under Windows (MSVC++ 6 w/SP something), this compiles: // these are constructed w/values, it's just ugly here std::string sRegEx(); std::string sMe(); boost::regex regexpPattern( sRegEx ); boost::cmatch mrFound; bool bMatch = boost::regex_search(sMe, mrFound, regexpPattern); Under Linux (gcc 3.3) I get an error message (I added the whitespace): error: could not convert `mrFound' to `boost::match_results< std::__normal_iterator< const char*, std::basic_string< char, std::char_traits<char>, std::allocator<char> > >, std::allocator<char> >&' Digging around a little bit, I found this: template <class iterator, class Allocator> class match_results : yadda yadda and typedef match_results<const char*> cmatch; This certainly SEEMS much simpler than what the error message claims to be looking for. Can anyone help, please? Thanks, - Mark

1 0

Re: [crc] Simple example please
by Daryle Walker 12 Jun '03

12 Jun '03

On Monday, June 9, 2003, at 5:48 PM, Alex Henderson wrote: > I'd like to calculate a value for a cyclic redundancy check (crc) for > a number of files. > > This isn't going to be transmitted anywhere, I simply want to compare > the values to see whether the files are exactly the same. I know I can > simply check byte for byte, but I've got a few thousand files and it's > easier to calculate the crc for each and see whether any are the same > first. > > I've looked at the crc library, but can't get my head round the > example. It seems to test the library rather than show how its used. Yes, the sample program is meant as a test of the library. You can still get an idea of usage from it. There are also short examples in the HTML documentation. > I'm after something like > > std::vector<char> MyStream1; // the whole file (about 500kb) > > boost::SomeCRCTypeThatIsOptimisedForSafetyAndLargeishBlocks<???> x; > > long = x.checkme(MyStream1); > > Could someone please point me to some code where a simple example is > detailed? What's in the documentation is it. Do you have a specific CRC algorithm in mind? If so, you just pick the appropriate template parameters. If not, you can choose one of the samples I provided, like boost::crc_32_type. The optimized CRC function objects can work with a single byte at a time or an array-segment of bytes. By putting your file's data in a std::vector as you have in your example, you limit yourself to just the single-byte computing call, as the elements of a vector may not be contiguous. Just allocate a simple char array and use one of the CRC object's array computing calls or use one of the CRC functions. Based on what you said, I don't think you need the object-wrapping that std::vector gives. Daryle

2 4

Boost::Test v. CppUnit
by Jon Stewart 11 Jun '03

11 Jun '03

Howdy - I've been using CppUnit and various parts of Boost for a little over a year. The 1.3.0 rlease includes support for unit testing, a la CppUnit, JUnit, etc. At first glance, I can't find any fundamental differences between CppUnit and Boost's unit testing framework. I'm sure that CppUnit could benefit from a code review or two by Boost's luminaries, but I'm still surprised that a new framework was created instead of trying to improve CppUnit. Can anyone state a few areas where CppUnit and Boost::test differ? In general, I tend to prefer Boost's solutions over others', but I don't see a clear cut case to be made for Boost over CppUnit, vis a vis unit testing. I'd be happy to change my mind, however. :-) Jon -- Jon Stewart stew1(a)ellipsis.cx

7 8

RE: [Boost-Users] first time boost
by Jim.Hyslop 11 Jun '03

11 Jun '03

G Keshav wrote: > I want to use boost in my MS VIsual Studio project. > Can you please advise stepwise about the procedure. 1) Decide which part of Boost you want to use. 2) download and extract the source from boost.org (follow the instructions there) 3) include the proper headers 4) write your code -- Jim

1 0

first time boost
by G Keshav 11 Jun '03

11 Jun '03

Hi there, I want to use boost in my MS VIsual Studio project. Can you please advise stepwise about the procedure. Thanks and regards, Gagnesh --------------------------------- Do you Yahoo!? Free online calendar with sync to Outlook(TM). [Non-text portions of this message have been removed]

1 0

using shared pointer in a linked data structure
by Zumichu 11 Jun '03

11 Jun '03

Basically I'm having trouble understanding exactly how to use shared_ptr . If I have a class: GraphNode { public: AddChild(shared_ptr<GraphNode> n) { m_Children.push_back(n); shared_ptr<GraphNode> thisptr(this); n->m_Parents.push_back(thisptr); } DeleteChild(shared_ptr<GraphNode> n); private: vector<shared_ptr<GraphNode> > m_Children; vector<shared_ptr<GraphNode> > m_Parents; }; main() { shared_ptr<GraphNode> n1(new GraphNode()); shared_ptr<GraphNode> n2(new GraphNode()); n1->AddChild(n2); } doing it this way gives me a run-time error that seems to have to do with updating the child node's parent vector with the parent node's this pointer. Am I not using the shared_ptr right?

2 1