- Boost-users - lists.boost.org

Unknown program_options bomb
by Merrill Cornish 21 Nov '05

21 Nov '05

I am using Boost 1.33.0 with MinGW on Windows XP. I am using the following program_optins code to collect command line arguments: po::options_description desc("Options"); desc.add_options() ...[everyting defined in here works as advertised]... ;//add-options() // specify the single, required graph file positional parameter po::positional_options_description pd; pd.add("graph_file", 1); po::variables_map vm; po::store(parse_command_line(argc, argv, desc), vm); po::notify(vm); ... The above code is shamelessly lifted from the tutorial. Everything defined in desc.add_options() works including the "-h" alternative to "--help". However, two things fail. * vm.count("graph_file") returns 0 regardless of whether there are positional argument(s). * specifying an -x, --xx, --xx=n option where "x" or "xx" is not included in desc.add_options() causes store() to not return (the program simply stops). Am I missing something? Merrill

3 2

[wave] question about preprocessing hooks
by Alexey Bakhvalov 20 Nov '05

20 Nov '05

Hello, What is a correct way to implement 'defined_macro' hook? For some reason I always get empty 'definition' list. My hook implementation looks like this: class hook : public boost::wave::context_policies::default_preprocessing_hooks { public: template <typename TokenT, typename ParametersT, typename DefinitionT> void defined_macro(TokenT const &macro_name, bool is_functionlike, ParametersT const &parameters, DefinitionT const &definition, bool is_predefined) { if (!definition.empty()) { // never gets here } } }; I am using Boost 1.33 and VC .NET 2005 compiler. Alexey Bakhvalov

2 1

Re: [Boost-users] Wave library fails to compile with VC7.1 - Bufferoverrun in cl.exe
by Hartmut Kaiser 20 Nov '05

20 Nov '05

> Didier Malenfant wrote: > > > It seems like this was fixed by turning off the rtti for > > VC7.1 and I verified that both V1 and V2 jamfiles have the > workaround > > line in them (not sure which one is getting used as I'm not passing > > any parameters for that) > > V1 is used by default. Just another thought. Since you're using the 'vc-7_1-stlport' toolset, you'll need to add a <vc-7_1-stlport><*><rtti>off # workaround for compiler bug line to the Jamfiles. Could you please try this and report back? Regards Hartmut

1 0

Documentation Typo
by Merrill Cornish 20 Nov '05

20 Nov '05

FYI: The documentation page for the ptr_list<> class, http://www.boost.org/libs/ptr_container/doc/ptr_list.html, has a heading the says "Class ptr_vector". The headings are correct for ptr_vector<>, ptr_array<>, and ptr_deque. Merrill

2 1

Thread sheduling priority
by Leon Mergen 19 Nov '05

19 Nov '05

Hello, If been lookup up possible planned functonality for adjusting a thread's priority in the boost thread library. In my search, i found this email: http://aspn.activestate.com/ASPN/Mail/Message/boost/1690190 In that message, sent 2 years ago, William E. Kempf states that they are implemented but are still undergoing design changes. At the moment, I still don't see that functionality in the main boost library - is there any idea when this is going to be part of the main boost library ? Thanks in advance. -- Leon Mergen

1 0

[Iostreams] file_descriptor::seek() bug
by Takeshi Mouri 19 Nov '05

19 Nov '05

Hello, I try file_descriptor::seek() for a big file over 4GB on Win32, but the result was not correct. file_descriptor.cpp(185) > LONG lDistanceToMoveHigh = > off < 0xffffffff ? > static_cast<LONG>(off >> 32) : > 0; I think that lDistanceToMoveHigh always becomes 0. The condition above should be "off > 0xffffffff". Regards, Takeshi Mouri

2 1

build times
by gast128 19 Nov '05

19 Nov '05

Dear all, we use boost a lot in our code, but unfortunately this slows down the build times enormously. Although we use the precompiled header in Visual Studio, tracing the includes (with the /showincludes option), it seems that even relatively basic libraries (e.g. tuples, boost\utility.hpp, boost function and boost bind) end up in a massive include of mpl and type traits header files. So I begin to wonder if this wonderful template and preprocessor stuff is really such a good idea for those basic library. Besides that it doesn't improve the readability of the source, my guess is that it has a negative effect on short build times. Can a boost guru comment on this? Note that this is not a criticism, only an observation. I always like to write a boost bind construction, especailly with the == addition in 1.33. Wkr, me

14 91

[iostreams] Assertion in detail/optional.hpp, line 54
by Daniel Lidström 18 Nov '05

18 Nov '05

Hello! when compiled in debug mode, the following program causes an exception in the file p:\boost_1_33_0\include\boost-1_33\boost\iostreams\detail\ optional.hpp, line 54 (I'm using VC 7.1.3088, Windows XP): #include <iostream> #include <boost/iostreams/concepts.hpp> // sink #include <boost/iostreams/stream.hpp> // stream using namespace std; class IOutput { public: virtual void print(int, const char*) const = 0; }; class Output : public IOutput { public: void print(int id, const char* a) const { cout << a; } }; static IOutput* gpOutput; template<int id> class my_sink : public boost::iostreams::sink { public: std::streamsize write(const char* s, std::streamsize n) { gpOutput->print(id, s); std::streamsize sz = 0; for( const char* p=s; *p; ++p ) sz++; return sz; } }; typedef boost::iostreams::stream<my_sink<5> > output; int main() { gpOutput = new Output; output stream; stream << "Daniel" << endl; return 0; } When I compile in release mode the program prints `Daniel', just as expected. Is there a problem with my program? Thanks in advance! Hälsningar, Daniel

2 1

typo in Boost.Function docs
by Weston Markham 18 Nov '05

18 Nov '05

The following is from the documentation published at www.boost.org. Surely that "!" in the "throws" description should not be there, right? Weston functionN invocation result_type operator()(arg1_type a1, arg2_type a2, ... , argN_type aN) const; Effects: f(a1, a2, ..., aN), where f is the target of *this. Returns: if R is void, nothing is returned; otherwise, the return value of the call to f is returned. Throws: bad_function_call if !this->empty(). Otherwise, may through any exception thrown by the target function f.

2 1

calling on the template gods
by Doug Henry 18 Nov '05

18 Nov '05

I have been attempting to hide the boost interface when using the boost smart pointers. I basically want to get rid of boost declarations in my API's. This basically involves typedef-ing template classes, and the best solution I have been able to come up with is: template <typename T> struct pointer { typedef boost::scoped_ptr<T> obj; typedef boost::scoped_array<T> arr; typedef boost::shared_ptr<T> sobj; typedef boost::shared_array<T> sarr; }; so I can do things like pointer<int>::arr var(new int[10]). This works fine, but if I try to get tricky it falls apart. Another example: template <typename T> class Stuff { private: pointer<T>::arr some_stuff; }; This does not work (I can't even start to tell you why), but I have hit the point where I have to start second guessing my approach to this problem. Are there other (possibly terrible) tricks to get around the typedef templates problem? Any input will probably help. -thanks

5 4