September 2009 - Boost-users - lists.preview.boost.org

Solaris Container
by Bernhard Duebi 10 Sep '09

10 Sep '09

Hello, I'm not a programmer, I'm only a system engineer, so please be gentle. One of our code producers compiled boost in a Solaris Zone/Container using SUN Studio. For testing he used a program called test. I guess it is part of the distribution. It tests if 2+2==4 with different calls to boost. The test program runs on real machines, but not in Containers. Looking at the output of truss I found a probable cause. The program climbs up the process tree using the proc filesystem until it reaches the root of the tree. In a real machine the init process has PID 1 and this seems the sign for the program to stop climbing. In a Container the zsched process has a random PID and its PPID is the same as the PID. This leads to an endless loop. In the test.cpp file there is not much code besides the add function and some calls to the boost library. So, I guess the process tree walk is done somewhere in the boost library. My fellow co-worker denies this strictly. As we both have no clue about boost, I want to ask you to sched some light on this incident. - Is it possible that boost does that process tree walk ? - Is this a known problem ? - Are there any hints how to solve this problem ? I talked to SUN and they say it's not their problem. There is no official whitepaper that requires that the init process always has PID 1. Any help is appreciated Sincerely Bernhard -- Minds are like parachutes They only function when open

6 8

how to suppress warning messages in boost file system
by Avi Bahra 10 Sep '09

10 Sep '09

Does anyone know how to suppress warning messages made by HP-UX aCC compiler. In particular: "/scratch/ma/ma0/hpia64/boost/boost_1_39_0/boost/filesystem/operations.hpp", line 579: warning #2236-D: controlling expression is constant && "boost::filesystem::complete() precondition not met" ); I get so many of these, its difficult to see the wood from the trees. Any help appreciated. Best regards, Ta, Avi

1 0

[asio] fails to compile on HP-UX aCC 11.23 due to ::pselect
by Avi Bahra 10 Sep '09

10 Sep '09

Boost 1.39/1.40 asio will not compile with acc A.06.20 on HPUX 11.23 : I have following error: "/scratch/ma/ma0/hpia64/boost/boost_1_39_0/boost/asio/detail/socket_ops.hpp", line 643: error #2282: the global scope has no "pselect" return error_wrapper(::pselect(nfds, readfds, The offending code in socket_ops.hpp is : inline int select(int nfds, fd_set* readfds, fd_set* writefds, fd_set* exceptfds, timeval* timeout, boost::system::error_code& ec) { ........... #if defined(__hpux) && defined(__HP_aCC) timespec ts; ts.tv_sec = timeout ? timeout->tv_sec : 0; ts.tv_nsec = timeout ? timeout->tv_usec * 1000 : 0; return error_wrapper(::pselect(nfds, readfds, writefds, exceptfds, timeout ? &ts : 0, 0), ec); #else int result = error_wrapper(::select(nfds, readfds, writefds, exceptfds, timeout), ec); if (result >= 0) clear_error(ec); return result; #endif } If I comment out the #ifdef with pselect in it, then everything appears to be ok. >From googling around it appears that pselect was only valid for certain OS versions, however #if can not detect this. Has anyone found a more reliable workaround ? Best regards, Ta, Avi

1 0

sending struct of int from Linux to Windows using boost::asio
by A.Agrawal＠patrick.com.au 10 Sep '09

10 Sep '09

Hello All, I am having trouble using boost:asio to send a simple struct message_header from server (linux) to the client (windows). typedef struct MessageHeader { uint32_t _data_length; //length of data attached uint8_t _message_type; //MESSAGE_TYPE uint8_t _segment_name; //SEGMENT_NAME uint16_t _request_id; //Request Id assigned by the client to the request to uniquely identify the request uint16_t _count_of_records; //Number of records affected by the operation, 0 means complete segment is affected. } MessageHeader; typedef struct Message { boost::array<char, header_length> _header_array; } Message Server side code (Linux): =============== MessageHeader request_header; request_header._data_length = 0; request_header._message_type = MSG_READ; request_header._segment_name = SEG_OOW; request_header._count_of_records = 0; request_header._request_id = 100; string _data_serialized = ""; uint8_t header[header_length]; //header_length is defined as 10 memcpy(header, (uint8_t*)(&(request_header)), header_length); std::vector<boost::asio::const_buffer> buffers; buffers.push_back(boost::asio::buffer(header, header_length)); buffers.push_back(boost::asio::buffer((_data_serialized).c_str(), _data_serialized.length())); boost::asio::async_write(_socket, buffers, boost::bind(&ClientSession::handle_write, shared_from_this(), boost::asio::placeholders::error)); Client Side code (Windows): ================ Message _read_message; MessageHeader _message_header; Handle_connect(....) { boost::asio::async_read(_socket, boost::asio::buffer(_read_msg._header_array, header_length), boost::bind(&DJClient::Handle_read_header, this, boost::asio::placeholders::error)); } Handle_read_header(...) { memcpy(&_message_header, read_msg._header_array.data(), header_length); } Surprisingly the same code works fine from "Windows to Windows" and "Windows to linux" but has problem when used on "Linux to Windows". When sent from Linux server, Windows server receives garbage data. Sending of normal ASCII characters from Linux to Windows work fine. Please help if anyone can. Thanks and Regards, Anil Agrawal

4 3

graph cut compilation error
by David Doria 10 Sep '09

10 Sep '09

I made a super simple example: A graph with 2 nodes and an edge between them. On trying to call kolmogorov_max_flow, I am getting: error: cannot convert 'boost::detail::error_property_not_found' to 'long int' in initialization Can anyone see what I am doing wrong? #include <iostream> // for std::cout #include <utility> // for std::pair #include <algorithm> // for std::for_each #include <boost/graph/graph_traits.hpp> #include <boost/graph/adjacency_list.hpp> #include <boost/graph/kolmogorov_max_flow.hpp> using namespace boost; // create a typedef for the Graph type typedef adjacency_list<vecS, vecS, bidirectionalS> Graph; int main(int,char*[]) { // declare a graph object Graph g(2); //a graph with 2 vertices // add the edges to the graph object for (int i = 0; i < 1; ++i) { add_edge(0, 1, g); //add an edge between node 0 and node 1 } //find min cut //typedef adjacency_list_traits<vecS, vecS, directedS> Traits; //if we want a directed graph typedef adjacency_list_traits<vecS, vecS, bidirectionalS> Traits; Traits::vertex_descriptor s, t; *//the error is on this line* long flow = kolmogorov_max_flow(g, s, t); // a list of sources will be returned in s, and a list of sinks will be returned in t std::cout << "Max flow is: " << flow << std::endl; return 0; } Also, I see that you can add edge weights like this: > const int num_nodes = 5; > E edges[] = { E(0,2), > E(1,1), E(1,3), E(1,4), > E(2,1), E(2,3), > E(3,4), > E(4,0), E(4,1) }; > int weights[] = { 1, 2, 1, 2, 7, 3, 1, 1, 1}; > > Graph G(edges + sizeof(edges) / sizeof(E), weights, num_nodes); > but if I want to add edges like this: for (int i = 0; i < 1; ++i) > { > add_edge(0, 1, g); //add an edge between node 0 and node 1 > } > How would I specify the edge weights? Thanks! David

2 10

[thread] Obtaining thread id on Windows
by Agustín K-ballo Bergé 09 Sep '09

09 Sep '09

I'm using Boost.Thread to create a thread that process the message queue on Windows. In order to send it a message via PostThreadMessage I need a thread id rather than a handle. Is there a way to obtain the native thread id from a boost::thread ? If not, any help on how to patch my local copy of the library? There is no documented way of getting a thread id from a thread handle in Windows XP. Thank you. Agustín K-ballo Bergé.-

6 8

[Spirit/Graph] New 1.40.0 compile errors in Spirit/Graph headers when using GCC with _GLIBCXX_DEBUG‏
by Tony Lewis 09 Sep '09

09 Sep '09

When I compile this script (saved as test.cpp) which includes a pair of Boost Graph headers : #include <boost/graph/adjacency_list.hpp> #include <boost/graph/depth_first_search.hpp> int main() { return 0; } ...using the compile command : g++ -I/usr/include/boost -D_GLIBCXX_DEBUG test.cpp -o test ...I get (this warning and) these errors from the Boost Spirit library: In file included from /usr/include/c++/4.3/backward/hash_set:64, from /usr/include/boost/graph/adjacency_list.hpp:25, from test.cpp:1: /usr/include/c++/4.3/backward/backward_warning.h:33:2: warning: #warning This file includes at least one deprecated or antiquated header which may be removed without further notice at a future date. Please use a non-deprecated interface with equivalent functionality instead. For a listing of replacement headers and interfaces, consult the file backward_warning.h. To disable this warning use -Wno-deprecated. In file included from /usr/include/boost/spirit/home/phoenix/stl/algorithm/detail/has_find.hpp:23, from /usr/include/boost/spirit/home/phoenix/stl/algorithm/querying.hpp:17, from /usr/include/boost/spirit/home/phoenix/algorithm.hpp:11, from /usr/include/boost/spirit/home/phoenix.hpp:10, from /usr/include/boost/graph/depth_first_search.hpp:24, from test.cpp:2: /usr/include/boost/spirit/home/phoenix/stl/algorithm/detail/./is_std_map.hpp:38: error: wrong number of template arguments (4, should be 1) /usr/include/boost/spirit/home/phoenix/stl/algorithm/detail/./is_std_map.hpp:28: error: provided for ‘template<class T> struct boost::is_std_map’ /usr/include/boost/spirit/home/phoenix/stl/algorithm/detail/./is_std_map.hpp:38: error: expected unqualified-id before ‘>’ token /usr/include/boost/spirit/home/phoenix/stl/algorithm/detail/./is_std_map.hpp:53: error: wrong number of template arguments (4, should be 1) /usr/include/boost/spirit/home/phoenix/stl/algorithm/detail/./is_std_map.hpp:43: error: provided for ‘template<class T> struct boost::is_std_multimap’ /usr/include/boost/spirit/home/phoenix/stl/algorithm/detail/./is_std_map.hpp:53: error: expected unqualified-id before ‘>’ token In file included from /usr/include/boost/spirit/home/phoenix/stl/algorithm/detail/has_find.hpp:24, from /usr/include/boost/spirit/home/phoenix/stl/algorithm/querying.hpp:17, from /usr/include/boost/spirit/home/phoenix/algorithm.hpp:11, from /usr/include/boost/spirit/home/phoenix.hpp:10, from /usr/include/boost/graph/depth_first_search.hpp:24, from test.cpp:2: /usr/include/boost/spirit/home/phoenix/stl/algorithm/detail/./is_std_set.hpp:37: error: wrong number of template arguments (3, should be 1) /usr/include/boost/spirit/home/phoenix/stl/algorithm/detail/./is_std_set.hpp:28: error: provided for ‘template<class T> struct boost::is_std_set’ /usr/include/boost/spirit/home/phoenix/stl/algorithm/detail/./is_std_set.hpp:37: error: expected unqualified-id before ‘>’ token /usr/include/boost/spirit/home/phoenix/stl/algorithm/detail/./is_std_set.hpp:51: error: wrong number of template arguments (3, should be 1) /usr/include/boost/spirit/home/phoenix/stl/algorithm/detail/./is_std_set.hpp:42: error: provided for ‘template<class T> struct boost::is_std_multiset’ /usr/include/boost/spirit/home/phoenix/stl/algorithm/detail/./is_std_set.hpp:51: error: expected unqualified-id before ‘>’ token In file included from /usr/include/boost/spirit/home/phoenix/stl/algorithm/detail/has_sort.hpp:22, from /usr/include/boost/spirit/home/phoenix/stl/algorithm/transformation.hpp:17, from /usr/include/boost/spirit/home/phoenix/algorithm.hpp:12, from /usr/include/boost/spirit/home/phoenix.hpp:10, from /usr/include/boost/graph/depth_first_search.hpp:24, from test.cpp:2: /usr/include/boost/spirit/home/phoenix/stl/algorithm/detail/./is_std_list.hpp:36: error: wrong number of template arguments (2, should be 1) /usr/include/boost/spirit/home/phoenix/stl/algorithm/detail/./is_std_list.hpp:28: error: provided for ‘template<class T> struct boost::is_std_list’ /usr/include/boost/spirit/home/phoenix/stl/algorithm/detail/./is_std_list.hpp:36: error: expected unqualified-id before ‘>’ token I don't get the errors if I use this compile command which doesn't define _GLIBCXX_DEBUG : g++ -I/usr/include/boost test.cpp -o test ...and I don't get the errors if I use this compile command which uses the previous 1.39 libraries : g++ -I/usr/include/boost-1_39 -D_GLIBCXX_DEBUG test.cpp -o test If I reverse the order of the headers when using the original compile command, I still get errors but different errors. My compiler is gcc 4.3.3 (Ubuntu 4.3.3-5ubuntu4), my OS is Ubuntu 9.04 and my kernel is 2.6.28-15-generic. Many thanks in advance for any help. _________________________________________________________________ View your other email accounts from your Hotmail inbox. Add them now. http://clk.atdmt.com/UKM/go/167688463/direct/01/

3 3

units use and weird cases....suggestions needed
by Noah Roberts 09 Sep '09

09 Sep '09

I'm using the units library to make sure values supplied to functions make sense, dimensionally, and are in the units expected. I've got a strange case I have to work with though. For the most part we've decided to standardize on the SI system. I write my functions to be system independent when possible, but most of what I'm working with is empirically based. For this I'll take the SI units expected and write the function only for that system (since other systems will use different equations entirely). The problem domain is fluid dynamics so enter the Kv and Cv values. Technically these are a measure of volumetric flow. For instance, Kv is defined as: the flow of water with temperature ranging 5 - 30 oC through a valve in cubic meters per hour (m3/h) with a pressure drop of 1 bar The problem is that engineers see this as a "unitless" value and treat it as having no units at all. They use Kv in SI based equations, not m3/s as we'd expect. Thus all equations being fed to us use Kv as if it were a natural part of the equation. Most of the time these dimensions do not compute dimensionally anyway but there's rare cases when they do. I have gotten the engineers (those giving me specs for code I write) to decide on Kv instead of doing both Kv and Cv (it's a basic conversion). The options I see are: 1. Convert both Kv and Cv into m3/s and require engineers to give equations with that in mind. This will be very unnatural for them though. 1a. Do the conversion but don't tell them I'm doing it. This one seems the least preferable as we've tried to make sure eng. documents and code match and further...it might be hard for maintenance developers later. 2. Pretend...just use m3ps and pretend I'm using Kv (m3/h). Don't convert or anything...just lie. I have a hard time with that one since it sort of goes against the purpose of the library and again would seem to introduce confusion for maintenance. 3. Use different types either in the form of a different unit "system" or using BOOST_STRONG_TYPEDEF...or something. What are some suggestions here?

2 4

Boost-Pro Windows Installer for Boost 1.40
by Klaim 09 Sep '09

09 Sep '09

Hi! I'm not sure how and when the windows installer for Boost 1.40 is planned. Where can I find the informations about that? Will it be available soon? I'm waiting for this installer before updating my Boost version. Thanks for you time. Joël Lamotte.

4 5

[iostreams] regex_filter how-to
by Michał Nowotka 09 Sep '09

09 Sep '09

Hello I have following problem: I need to filter some records from one file and save it to another in my c++ application. For example from this file: /////////////////// in.txt //////////////////////////////// http://google.com http://yahoo.com ... http://google.com/analytics //////////////////////////////////////////////////////////// I want to only extract lines that match regex: ^(?:http://google.com).* to get: //////////////// out.txt /////////////////////////////// http://google.com ... http://google.com/analytics ///////////////////////////////////////////////////////// So I wrote something like this: class Writer { public: Writer() :matchesCount_(0){} virtual std::string operator() (const boost::match_results<const char*>& result) { matchesCount_ = result.size(); return aux_; } int getMatchesCount() const { return matchesCount_; } virtual ~Writer(){} private: std::string aux_; //this i completely useless but i must return something in operator() int matchesCount_; }; /////////////////////////////////////////////////////////////////////////////// class FileWriter : public Writer { public: FileWriter(std::ostream* of) :of_(of) {} std::string operator() (const boost::match_results<const char*>& result) { *of_ << *result.begin() << endl; return Writer::operator()(result); } private: std::ostream* of_; }; /////////////////////////////////////////////////////////////////////////////// int main(int argc, char *argv[]) { boost::regex match_lower("^(?:http://google.com).*"); std::ofstream out("out.txt"); string str; filtering_istream first(boost::iostreams::regex_filter(match_lower, FileWriter(&out))); first.push(file_source("in.txt", ios_base::in)); first.ignore();// my output is a side effect of filtering so I don't have to process this stream return 0; } ///////////////////////////////////////////////////////////////////////////// It works fine for short files (IMO for files which size is smaller then size of stream buffer). But I work with very large files (~4,7 GB) and then this is not a good solution. Do you have any idea how to solve it? -- Regards Michał Nowotka

3 4