- Boost-users - lists.boost.org

[RFC] Preferred API for a CGI library proposal
by Darren Garvey 05 Apr '07

05 Apr '07

Hello all, This may seem off-topic, but I assure you it isn't. ;) This is an initial probe for ideas about an API for [any] upcoming CGI library submission. I have my own ideas, but I don't think putting them here would be particularly relevant. The reason I'm asking such an open-ended question is that I'm assuming lots of people here will have experience using CGI libraries (possibly in other languages, possibly your own) and may have specific ideas about what should or shouldn't move from them into a Boost.CGI library, if one were accepted. Whether or not you see the point of a C++ CGI library, your comments on APIs provided by other languages would still be very much appreciated! Cheers, Darren

2 1

[pool] Does alloc_size waste memory?
by Soren Soe 05 Apr '07

05 Apr '07

Hi, I am trying to replace an old implementation of a small object memory allocator with boost::pool. I have noticed that boost::pool uses quite a bit more memory for certain requested_size than the allocator I am trying to replace. After looking at the code and reading the Guaranteeing Alignment <http://www.boost.org/libs/pool/doc/implementation/alignment.html> document, I am wondering why boost::pool::alloc_size() have to be lcm(sizeof(void*), sizeof(size_type), requested_size)? Lots of bytes can be wasted in the chunks returned to user when requested_size is not a multiple of min_alloc_size (lcm(sizeof(void*), sizeof(size_type))). I would think the requested_size alignment could be the first multiple of min_alloc_size that is larger than requested_size. For example, the convoluted example in the alignment document has requested_size==7, sizeof(void*)==3, and sizeof(size_type)==5. Currently alloc_size() returns lcm(3,5,7)==105, but since min_alloc_size is lcm(3,5)==15, I think requested_size alignment could be 15 bytes, rather than the 105 bytes. Another way to look at this. If requested_size was changed to 15, then the current implementation of alloc_size() would return 15 bytes (lcm(3,5,15)==15), so why the huge penalty when requested size is 7? Am I missing something? Thank you. Soren

1 0

boost::filesystem and an odd linker error.
by Matt Handley 05 Apr '07

05 Apr '07

Hi. I'm having difficulty using the boost::filesystem library. I have checked-out the source from the CVS repository, and ran: /usr/local/bin/bjam --v2 --toolset=darwin --with-regex --with- serialization --with-filesystem --layout=system --prefix=SOMEDIR variant=release threading=single link=shared install Thinking that it would install headers in SOMEDIR/include and SOMEDIR/ lib. That it did. However, when I try to compile the simple_ls.cpp example, I get the following errors: $ g++ -o simple_ls simple_ls.cpp -Llibraries/lib -lboost_filesystem - lboost_system /usr/bin/ld: Undefined symbols: boost::filesystem::initial_path() boost::filesystem::is_directory(boost::filesystem::path const&) boost::filesystem::system_complete(boost::filesystem::path const&) boost::filesystem::path::path(char const*, bool (*) (std::basic_string<char, std::char_traits<char>, std::allocator<char> > const&)) boost::filesystem::detail::dir_itr_init (boost::shared_ptr<boost::filesystem::detail::dir_itr_imp>&, boost::filesystem::path const&) boost::filesystem::detail::dir_itr_increment (boost::shared_ptr<boost::filesystem::detail::dir_itr_imp>&) boost::filesystem::detail::dir_itr_dereference (boost::shared_ptr<boost::filesystem::detail::dir_itr_imp> const&) boost::filesystem::exists(boost::filesystem::path const&) boost::filesystem::path::native_file_string() const boost::filesystem::path::native_directory_string() const boost::filesystem::path::leaf() const collect2: ld returned 1 exit status Which is a bit odd. The contents of the libraries/lib directory: $ ls libraries/lib/libboost* libraries/lib/libboost_filesystem.dylib libraries/lib/ libboost_serialization.dylib libraries/lib/ libboost_wserialization.dylib libraries/lib/libboost_regex.dylib libraries/lib/ libboost_system.dylib Why is this simple program not linking? Matt

1 0

[guid] Linker Error when using precompiled headers
by Christian Henning 05 Apr '07

05 Apr '07

Hi there, there is a problem with the free flowing functions random_byte(...) and swap(...). stdafx.obj : error LNK2005: "unsigned char __cdecl boost::detail::random_byte(void)" (?random_byte@detail@boost@@YAEXZ) already defined in database.obj stdafx.obj : error LNK2005: "void __cdecl boost::swap(class boost::guid &,class boost::guid &)" (?swap@boost@@YAXAAVguid@1@0@Z) already defined in database.obj To solve that one would need to move the functions inside a structure. For example: namespace boost { namespace detail { struct utilities { static uint8_t random_byte() { ... } }; }} Christian

2 3

[FileSystem] can't use path class
by Alexander Fradiani 05 Apr '07

05 Apr '07

Hello. I can't initialize the path class with a char* argument. I'm using cygwin, with the boost libraries installed as part of the cygwin installation packages. this is the code: #include "boost/filesystem/operations.hpp" #include "boost/filesystem/path.hpp" #include <iostream> #include <string> using namespace boost::filesystem; using namespace std; int main() { path mypath("/code/fileMan.txt"); return 0; } and the error i get when trying to compile with g++ is: "undefined reference to boost::filesystem::path::path(char const*)" I dont know exactly what is wrong, this is the first time i use the boost libraries so i really appreciate some help, thanks _________________________________________________________________ Charla con tus amigos en línea mediante MSN Messenger: http://messenger.latam.msn.com/

2 1

Re: [Boost-users] [bind] binding a boost::reference_wrapper to a function template
by Barchenkov, Alexei (IT) 05 Apr '07

05 Apr '07

>The problem is that tr is ref( t ) and is interpreted by bind as a request to store a reference to t. >boost::bind( &foo< Test& >, tr ) should work, if this helps. Peter, thank you very much. Could you please clarify what you mean however, because my understanding was that on construction, boost::bind internally just stores a copy of the bound argument without regard to the signature of the function to be invoked. The conversion of the argument to the type appropriate to the function invocation occurs when the functor is actually called. Is my understanding incorrect? Thank you. -------------------------------------------------------- NOTICE: If received in error, please destroy and notify sender. Sender does not intend to waive confidentiality or privilege. Use of this email is prohibited when received in error.

2 1

bind visit_each and obtaining argument values from bind
by Braddock Gaskill 05 Apr '07

05 Apr '07

Am I correct that the bind visit_each functionality briefly mentioned in the boost::bind documentations will allow me to get at actual argument values passed to a bind statement? That would be VERY useful to me at the moment. Is there a working example? The bind_visitor.cpp example seems to be broken(?). thanks, braddock

2 3

ERROR
by Lucas de Sales Joner 05 Apr '07

05 Apr '07

C:\Arquivos de programas\Dev-Cpp\include\boost\numeric\ublas\exception.hpp exception handling disabled, use -fexceptions to enable Somebody Know what is this and how I do that de error say? _________________________________________________________________ Veja só alguns dos novos serviços online no Windows Live Ideas — são tão novos que ainda não foram disponibilizados oficialmente. http://ideas.live.com

1 0

Plot
by Lucas de Sales Joner 05 Apr '07

05 Apr '07

Somebody know how I plot a vector boost? _________________________________________________________________ Obtenha o novo Windows Live Messenger! http://get.live.com/messenger/overview

1 0

pointers to static structures
by Scott Finley 05 Apr '07

05 Apr '07

I am wondering about best practices for the use of pointers to static structures. I have a class that will essentially replace an enumeration and represent a set of options. The options will be static members of the class. I would also like to have a vector of pointers to all the static members in order to be able to query all possible values at run time. This vector would be initialized explicitly by the constructor. Is there any reason that the vector should not hold normal raw pointers? Obviously I want to use these options without copying them around all over the place, so pointers seem to be the answer, but I will never be creating an option dynamically, so I don't really have any risk of memory leaks. Is there any reason to use shared_ptr with a custom destructor that does nothing? My use case would be something like: class myOption; class myOptionSet { public: myOptionSet() { allOptions.push_back( &OPTION_1 ); allOptions.push_back( &OPTION_2 ); } vector<myOption> const &getAllOptions() { return allOptions; } static myOption const OPTION_1; // initialized in .cpp static myOption const OPTION_2; private: static vector<myOption const *> allOptions; } class OptionUsingClass { OptionUsingClass( vector<myOption> const &possibleOptions ); // This can now check that the option is a possibleOption void doSomething( myOption const &option ); // or maybe this is better... void doSomething( myOption const *option ); }; Is this a good way to do things? Should I be using some sort of pointer wrapper instead? -- /************************************** * Scott Finley * Research Software Engineer * TomoTherapy Inc. * Email: sfinley(a)tomotherapy.com * Phone: 608.824.3381 ***/

1 0