July 2008 - Boost-users - lists.preview.boost.org

Operator Delete and Boost
by William Marié 09 Jul '08

09 Jul '08

Hi, I would like to know if the delete operator has been overridden in the boost library because i overrode it in my program and it causes some troubles with boost-python ? Thanks -- View this message in context: http://www.nabble.com/Operator-Delete-and-Boost-tp18363086p18363086.html Sent from the Boost - Users mailing list archive at Nabble.com.

1 0

[BGL] accessing vertex/edge by ID
by M. Siebeneicher 09 Jul '08

09 Jul '08

Hello list members, I've a little question: Is it possible and if so, how is it possible, to randomly access a vertex or edge by its id (according to vertex_index or edge_index proprty map)?, i.e. calling 'vertices(graph)[id]'. My problem is that I wan't to access a vertex or edge in an unique and graph lifespan-wide manner without iterate over all vertices or edges. With vertex_descriptor/edge_discriptor it's possible but only if I don't change the graph structur (not lifespan-wide). Kind regrads, Manuel -- GMX startet ShortView.de. Hier findest Du Leute mit Deinen Interessen! Jetzt dabei sein: http://www.shortview.de/wasistshortview.php?mc=sv_ext_mf@gmx

2 1

[BGL] direct access to compressed_sparse_row vectors?
by Zeljko Vrba 09 Jul '08

09 Jul '08

In the header file I find: // private: non-portable, requires friend templates --cut-- std::vector<EdgeIndex> m_rowstart; std::vector<Vertex> m_column; I already have a graph stored as a BGL adjacency_list and now I want to use a 3rd-party library (SCOTCH) which takes its input also in the CSR format. As I can see from the docs, there is no standardized interface to retrieve the raw CSR representation. This way, I either have to rely on the above members not being private, or myself effectively duplicating the already existing code for constructing the CSR representation. Am I overlooking something?

3 4

WinSock.h has already been included
by Jean-Sebastien Stoezel 09 Jul '08

09 Jul '08

Hi, Why trying to use the asio library I'm getting the following error (I believe this happens I include <boost/asio.hpp> once): WinSock.h has already been included c:\program files\boost \boost_1_35_0\boost\asio\detail\socket_types.hpp 27 I'm compiling using VS 2008 on WinXP. Any idea how to deal with this? Thanks, Jean

5 7

non reentrant standard libc with mult-threads
by frederic.bron＠alcan.com 09 Jul '08

09 Jul '08

I try to use the standard function "int system(const char*)" defined in <cstdlib> in parallel threads. We experience troubles under cygwin environment. Is there anything in boost or in the vault where I could find a thread-safe version of the system function? The list of non reentrant functions of the standard C library is pretty long. From info libc I get: errno localeconv setlocale stdin stdout stderr fdopen perror tempnam fopen putchar tmpnam getchar puts tmpfile gets emove vfprintf iprintf ename vsnprintf mkstemp snprintf vsprintf mktemp_t sprintf init_signal signal kill sigtramp aise calloc mblen setenv dtoa mbstowcs srand free mbtowc strtod getenv memalign strtol mallinfo mstats strtoul malloc putenv system malloc and wcstombs malloc_stats ealloc wctomb strdup strtok close link unlink execve lseek wait fcntl open write fork ead fstat sbrk gettimeofday stat getpid times fstat64 lseek64 open64 asctime. What would be the best portable way to deal with that? Regards, F. Bron Avis : Ce message et toute pièce jointe sont la propriété d'Alcan et sont destinés seulement aux personnes ou à l'entité à qui le message est adressé. Si vous avez reçu ce message par erreur, veuillez le détruire et en aviser l'expéditeur par courriel. Si vous n'êtes pas le destinataire du message, vous n'êtes pas autorisé à utiliser, à copier ou à divulguer le contenu du message ou ses pièces jointes en tout ou en partie. Notice: This message and any attachments are the property of Alcan and are intended solely for the named recipients or entity to whom this message is addressed. If you have received this message in error please inform the sender via e-mail and destroy the message. If you are not the intended recipient you are not allowed to use, copy or disclose the contents or attachments in whole or in part.

1 0

Boost.Asio async read/write callback weirdness/
by James Vanns 09 Jul '08

09 Jul '08

I am using this as my main event loop (so I can call it within another event loop that does additional processing): wait () { m_io.run_one(); // Executes at most 1 m_io.reset(); // Apparently I must do this m_io.poll(); // Executes 0 or more handlers but doesn't block m_io.reset(); // Apparently I must do this } So that run_one blocks until at least 1 handler has been executed and then poll doesn't block but executes the rest... I register a callback with async_read and async_write. On the client side the write callback is executed and the buffer is emptied (I'm using asio::streambuf). However, on the receiving endpoint (the server) the io_service reports that handlers (the correct amount for a send/write from the client side - e.g. they tally) are ready or have been executed - but no callback is actually registered. Running wireshark (ethereal) whilst the send/recvs are supposed to be happening reports that no data is sent!! That is, until I quite the client app (which closes the socket, stops the io_service etc.) then wireshark reports all the data that should have already been sent! I am really confused!! The callbacks report that the data has been sent (or at least that is what I understand) and the callbacks are even executed at the other end (or rather they're not but the io_service reports that handlers have been executed!). Can anyone think of why this might be?? Regards, Jim PS: Some more code: async_write (m_socket, m_outbound, transfer_at_least(p_max), bind(&base_io::write_callback, shared_from_this(), placeholders::error, placeholders::bytes_transferred)); async_read (m_socket, m_inbound, bind(&base_io::read_callback, shared_from_this(), placeholders::error, placeholders::bytes_transferred)); -- James Vanns Systems Programmer Framestore CFC Ltd.

1 0

Compilation problems using boost::regex in Objective-C on OSX
by Helge Gudmundsen 09 Jul '08

09 Jul '08

I have installed and built Boost (1.35.0) on my Mac, and things are working beautifully when I create a "pure" C++ project. My test source file looks like this: #include <iostream> #include <boost/regex.hpp> #include <string> int main (int argc, char * const argv[]) { std::string expr = std::string("(Colo)(u)(r)"); std::string textToMatch = std::string("Colour, colorize, colours, color"); boost::regex reg(expr, boost::regex::icase | boost::regex::perl); std::string result = regex_replace(textToMatch, reg, "$1$3"); std::cout << result; return 0; } It runs and produces the expected output. However, my intent is to use Boost in a project combined with Objective-C code. When I use the exact same code in a file called Test.mm and try to compile in XCode, I get compiler errors in cpp_regex_traits.hpp end perl_matcher_hpp: error: cannot find protocol declaration for "" - in line 799 of cpp_regex_traits error: expected unqualified-id before '=' token - in lines 259 and 265 of perl_matcher.hpp I am sure it is just some compiler flag I need to set somewhere, but at the moment I am stumped. Has anyone reading this list successfully mixed objective-c with c++ and the boost library? Sincerely, Helge

1 0

BGL Newbie
by Eric B 09 Jul '08

09 Jul '08

Hi this message may have been sent twice as i got an error on my previous post, if so I apologize I'd like to build a boost graph from my own graph representation which is made of: std::vector<tNode*> => vertices std::set<std::pair<tNode*,tNode*> > => edges (oriented) tNode is a custom struct which contains various data and 2 integer attribtues for coordinates x,y I need to compute XY for each node in order to display them I have done that with another lib but as BGL seems more recent I d like to evaluate it. I'm not a boost expert neither a template guru => therefore I have some problems (even after having read lot of examples). I'd like to have an example to understand how BGL works. 1) how to build a boost representation of my own graph (where boost nodes ID would be integer cast of tNode* so that I can feed back my own nodes with XY) 2) how to run a layout algorithm on the boost representation (kawai or any other else). Than you

2 3

Boost.Threads for Fedora/Wine
by most＠museresearch.com 09 Jul '08

09 Jul '08

Hi, I would like to use Boost.Thread in my 'winelib' application (a Linux executable that can also use Windows APIs via Wine). My application hosts, and shares threads with, 3rd party Windows DLL binaries. I've been told that to interact well those threads I must use the Windows threading APIs in my application. But by default Boost for Fedora (naturally) uses the pthread APIs. So I would need to use the Win32 variant of Boost.Thread. Any suggestions for how to use the Win32 thread APIs in Boost.Thread, but use the Linux variants of the rest of the libraries like Boost.Filesystem etc? I looked in to Boost.Build and bjam, but wasn't sure how to proceed. I got stuck trying to configure it to use winegcc or how to set BOOST_THREAD_PLATFORM_WIN32. Is that the right tack? Thanks for any tips... Michael Ost PS: FYI the Wine project provides the 'winegcc' tool which can build sources with Windows APIs into winelib shared libraries. That is, 'winegcc' could build the Win32 versions of the Boost.Thread source into Linux shared libraries. It is command line compatible with gcc.

2 1

[mpl] minimizing template nesting depth
by Brian Simpson 09 Jul '08

09 Jul '08

I've got some code that makes heavy use of mpl. I developed and successfully testid it with msvc8.0 . Then I #included it in a WinForms app (meaning that the compiler is now running with one of the /clr options). The compiler stops with an error explaining that it has "run out of stack space". This could be simply a difference in compiler implementations, but I thought I remembered that mpl was implemented with some builtin recursion unrolling classes to mitigate deeply recursive typedef issues. If so, it occurred to me that maybe the macros mpl is depending on to trigger the level of recursion unrolling aren't supplied when msvc is compiling for CLI. I've perused the code a bit to try to find where it is setting such a thing, but haven't been successful. So I thought I'd post here in hopes of a quick answer. Does mpl implement such a thing? And if so, where is it? I'd like to see if it isn't getting used for cli compilation. Thanks! --Brian

2 1