Nabble - Boost - Users

Re: [test] Linking against boost_unit_test_framework causes double-deletion in std::ostringstream

2009-11-11T03:08:30Z

On 2009-11-10 23:48:49 +1100, Roman Perepelitsa
<roman.perepelitsa@...> said:

> As far as I remember, boost_unit_test_framework defines main function.
> Having two main functions in the same application might be what's causing
> your problems. Another possibility is that the library and the program are
> compiled with different standard libraries.

Good thoughts, but:

1. Boost Test does not define main unless you tell it to, and that's
only in a header. The library itself doesn't contain main:

% nm /opt/local/lib/libboost_unit_test_framework-mt-d.dylib |c++filt|grep main
00000000000759e6 T boost::runtime::cla::argv_traverser::remainder(int&, char**)
000000000002e26a T boost::unit_test::unit_test_main(bool (*)(), int, char**)
0000000000136670 S typeinfo for std::domain_error
00000000000b2a0a S typeinfo name for std::domain_error
%

2. I compiled the library myself, or at least MacPorts did, using the
same standard library.

I'm pretty sure it's caused by the unit test library doing something
dodgy at static initialisation time. Though I don't know exactly what...

Mainly I'm interested in whether anyone else can replicate this problem.

_______________________________________________
Boost-users mailing list
Boost-users@...
http://lists.boost.org/mailman/listinfo.cgi/boost-users

doubt on lambda expression

2009-11-11T03:04:49Z

Hi All

I started playing with boost::lambda and I need some help or some clarification about it

or some hint to better understanding how does it work.

I have this piece of code

using namespace boost::lambda;

using boost::lambda::var;

int main(int argc, char** argv) {

std::vector< int > myList;

myList.push_back(100);

myList.push_back(200);

int interval = 20;

std::for_each( myList.begin(), myList.end(),

(

std::cout << "value: " << _1 << "\n",

std::cout << "interval: " << var( interval ) << "\n",

var( interval )++

));

}

I understood that writing a lambda expression would be a substitute

for writing a functor.

I expected the following result

value: 100

interval: 20

value: 200

interval: 21

Instead I had the following

intervalfirst value: 100

200

Why the strings "interval" and "first" are not written for each value and how can I achieve this?

Kind Regards

Screensaver, sfondi e scene: scarica i gadget di Messenger
_______________________________________________
Boost-users mailing list
Boost-users@...
http://lists.boost.org/mailman/listinfo.cgi/boost-users

Re: concatenate strings?

2009-11-11T03:04:46Z

> hi, i'm wondering if there is some function into boost to concatenate
> several strings to have just a long one?

Why can't you use standard operator+ or operator+= ?
_______________________________________________
Boost-users mailing list
Boost-users@...
http://lists.boost.org/mailman/listinfo.cgi/boost-users

Re: Constructing a shared pointer from a const pointer

2009-11-11T03:02:17Z

I got out of the problem.

The signature of bar is bar( int xpointer );

and foo calls bar( (int) &x).

All I have to pass to bar is an identity of x.

Johan Nilsson-4 wrote:

veerus wrote:
> Hi All,
>
> I have been in trouble for a day solving this seemingly simple
> problem - constructing shared_ptr<X> from const X& without copying
> the X object.
>
> The programming logic does not allow me to construct the shared_ptr
> like this: SX x = SX( new X( x ) );
>
> Consider the following piece of code:
>
> void foo( const X& x )
> {
> typedef shared_ptr<X> SX;
> SX px = SX( &x );
>
> bar( px );
> }
>
> The compiler gives the following error:
> In shared_ptr.hpp at line 184: 'initializing' cannot convert from
> 'const X *' to 'X*'.

To make it compile, you could simply change the typedef to "typedef
shared_ptr<X const> SX;".

However, the code is suspect - depending on how the caller of foo declares
and manages the variable used as "x", you're likely to either encounter a
delete of a (invalid reference to a) stack object sooner or later, or
perhaps a double delete. If I saw the above during a code review I would say
foo needs to be redesigned.

If you're relying on identity here, or if X is non-copyable, and want to
transfer ownership of "x" into "bar" you could define foo as:

void foo(std::auto_ptr<X> px)
{
bar(shared_ptr<X>(px));
}

[It's hard to tell without knowing the full context]

/ Johan

_______________________________________________
Boost-users mailing list
Boost-users@lists.boost.org
http://lists.boost.org/mailman/listinfo.cgi/boost-users

concatenate strings?

2009-11-11T03:01:32Z

hi, i'm wondering if there is some function into boost to concatenate several strings to have just a long one?
thanks

--
Ing Pedro Alonso Chaparro Valero
Ciudad Politecnica de la Innovación
iTEAM - Mobile Communications Group
Polytechnic University of Valencia
C\ Camino de Vera S/N, Edificio 8G
46022 Valencia, Spain

_______________________________________________
Boost-users mailing list
Boost-users@...
http://lists.boost.org/mailman/listinfo.cgi/boost-users

Re: BGL - parallel/distributed

2009-11-11T02:14:52Z

Thank you very much. I thought that the constructor of graph would also
take care about initializing the mpi (if it has distributedS selector).
Some working example in the documentation of graph_parallel would help a
lot for sure. Now it seems that everything works for me. Thank you once
more.

Nick Edmonds wrote:

>> To be even more precise, it exits the code in function MPI_Comm_dup
>> of mpi library. On linux I guess you have to have the mpi running,
>> but I don't know how it is on windows with mpich2.
>>
>
> You never initialized the Boost.MPI environment. See the tests, but
> you're looking for something like:
>
> mpi::environment env(argc, argv);
>
> This basically functions like MPI_Init() plus some Boost.MPI
> initialization. Comm_dup fails because the communicator you're trying
> to dup doesn't exist.
>
> -Nick
>
> _______________________________________________
> Boost-users mailing list
> Boost-users@...
> http://lists.boost.org/mailman/listinfo.cgi/boost-users
>
>

--
Dipl.-Ing. Ondrej Sluciak
Room CG-04-06
Vienna University of Technology, Austria
Institute of Communications and Radio-Frequency Engineering
Gusshausstrasse 25-29/389
http://www.nt.tuwien.ac.at

_______________________________________________
Boost-users mailing list
Boost-users@...
http://lists.boost.org/mailman/listinfo.cgi/boost-users

Re: [Asio] Is shared_ptr with async_send_to ok?

2009-11-11T01:04:21Z

> Socket.async_send_to(boost::asio::buffer(*SharedBuffer, BufferSize), DestinationEndpoint, boost::bind(&MyClass::OnPacketSent, this, boost::asio::placeholders::error, boost::asio::placeholders::bytes_transferred));
> // Shared buffer goes out of scope here, immediately!

You do not copy SharedBuffer, so it's destroyed when goes out of
scope. If you want SharedBuffer to outlive the handler, you can store
the shared_ptr within the handler, like this:

// UNTESTED!
using boost::bind;
using boost::protect;
Socket.async_send_to(buffer(*SharedBuffer, BufferSize), DestinationEndpoint,
bind(protect(bind(&MyClass::OnPacketSent, this, _1, _2)), _1, _2,
SharedBuffer)); // you can revert back to the asio placeholders
_______________________________________________
Boost-users mailing list
Boost-users@...
http://lists.boost.org/mailman/listinfo.cgi/boost-users

Re: Constructing a shared pointer from a const pointer

2009-11-10T23:37:51Z

veerus wrote:

> Hi All,
>
> I have been in trouble for a day solving this seemingly simple
> problem - constructing shared_ptr<X> from const X& without copying
> the X object.
>
> The programming logic does not allow me to construct the shared_ptr
> like this: SX x = SX( new X( x ) );
>
> Consider the following piece of code:
>
> void foo( const X& x )
> {
> typedef shared_ptr<X> SX;
> SX px = SX( &x );
>
> bar( px );
> }
>
> The compiler gives the following error:
> In shared_ptr.hpp at line 184: 'initializing' cannot convert from
> 'const X *' to 'X*'.

To make it compile, you could simply change the typedef to "typedef
shared_ptr<X const> SX;".

However, the code is suspect - depending on how the caller of foo declares
and manages the variable used as "x", you're likely to either encounter a
delete of a (invalid reference to a) stack object sooner or later, or
perhaps a double delete. If I saw the above during a code review I would say
foo needs to be redesigned.

If you're relying on identity here, or if X is non-copyable, and want to
transfer ownership of "x" into "bar" you could define foo as:

void foo(std::auto_ptr<X> px)
{
bar(shared_ptr<X>(px));
}

[It's hard to tell without knowing the full context]

/ Johan

_______________________________________________
Boost-users mailing list
Boost-users@...
http://lists.boost.org/mailman/listinfo.cgi/boost-users

Re: Boost.Asio: "End of file" when receiving data

2009-11-10T23:20:36Z

On Tue, Nov 10, 2009 at 7:16 PM, tiredashell <tiredashell@...> wrote:
> In addition to the "parameter is incorrect" error which I inquired about
> earlier (and remains unsolved),

This may indicate that you're trying to set a socket option with an
invalid value for your system.

> I am getting "End of file" sometimes when I
> attempt to receive data such as in the code below.

This usually indicates that the other party closed the socket.

> Please offer some
> guidance as to its cause, because there is absolutely no documentation on
> what either of these errors mean.

These are well-known system errors. You should familiarize yourself
with the particulars of the socket library implementation for the
system that you are working on. e.g. Learn what particular errors
your system can raise, and when.

Jon
_______________________________________________
Boost-users mailing list
Boost-users@...
http://lists.boost.org/mailman/listinfo.cgi/boost-users

Re: Constructing a shared pointer from a const pointer

2009-11-10T22:52:02Z

To add more to the above query,
assume that 'bar' is defined as follows:

map<SX, SX> m;

bar( SX px )
{

m.insert( pair<SX, SX>(px, someFunction( px ) ));

}

Can we do with a weak_ptr<X> instead of shared_ptr<X> ?

Thanks

Constructing a shared pointer from a const pointer

2009-11-10T22:41:17Z

Hi All,

I have been in trouble for a day solving this seemingly simple problem - constructing shared_ptr<X> from const X& without copying the X object.

The programming logic does not allow me to construct the shared_ptr like this: SX x = SX( new X( x ) );

Consider the following piece of code:

void foo( const X& x )
{
typedef shared_ptr<X> SX;
SX px = SX( &x );

bar( px );
}

The compiler gives the following error:
In shared_ptr.hpp at line 184: 'initializing' cannot convert from 'const X *' to 'X*'.

Please suggest any procedure to solve this problem.

Boost.Asio: "End of file" when receiving data

2009-11-10T18:16:18Z

In addition to the "parameter is incorrect" error which I inquired about earlier (and remains unsolved), I am getting "End of file" sometimes when I attempt to receive data such as in the code below. Please offer some guidance as to its cause, because there is absolutely no documentation on what either of these errors mean.

char reply[1024];

boost::system::error_code error;

size_t reply_length = s.receive(boost::asio::buffer(reply, 1024), 0, error);

if (error) cout << error.message() << endl;

_______________________________________________
Boost-users mailing list
Boost-users@...
http://lists.boost.org/mailman/listinfo.cgi/boost-users

Re: running a server "in background mode" ??

2009-11-10T16:46:42Z

> io_service.run();
> printf("Writting a line 1 \n");

The io_service can't exit the run command until all the work is
complete. You need to put the io_service.run() into its own thread if
you want to still execute things in main.

boost::thread runningService(boost::bind(&boost::asio::io_service::run,
io_service));
printf("Writting a line 1 \n");
sleep(5);
printf("Writting a line 2 \n");
runningService.join();

This will allow you to continue and execute code in your main function.
The problem you'll run into is trying to close the thread properly. I
used the HTTP Server 2 example to create a class that ran my process but
responded to an outside event to signal the class to stop processing.

http://www.boost.org/doc/libs/1_40_0/doc/html/boost_asio/example/http/server2/win_main.cpp

Ryan
_______________________________________________
Boost-users mailing list
Boost-users@...
http://lists.boost.org/mailman/listinfo.cgi/boost-users

Re: [Fusion] Views of Associative Sequences

2009-11-10T15:33:43Z

er schrieb:
> Sorry for troubling you with more trivialities, but by HEAD do you mean
> the latest update?
Yeah, the most up-to-date revision of the fusion trunk.
>
> Could you please confirm if this is the link from which I should copy
> fusion to my current release so I can start using the new facilities you
> have provided:
>
> http://svn.boost.org/svn/boost/trunk/boost/fusion/
That's correct.

svn co http://svn.boost.org/svn/boost/trunk/boost/fusion fusion

-Christopher

_______________________________________________
Boost-users mailing list
Boost-users@...
http://lists.boost.org/mailman/listinfo.cgi/boost-users

[inteprocess] Use of duration for the message_queue timed functions

2009-11-10T13:46:13Z

A suggestion for the message_queue class is to overload timed_send and timed_receive so that they can accept a duration (say in milliseconds). I believe currently they work on an absolute time being reached.

Thanks,

- Achal.

_______________________________________________
Boost-users mailing list
Boost-users@...
http://lists.boost.org/mailman/listinfo.cgi/boost-users

[Asio] Is shared_ptr with async_send_to ok?

2009-11-10T13:39:28Z

I'm using a UDP socket to send data asynchronously. I know I need to preserve the data in a buffer until the async write is complete. Can I use a shared_ptr to accomplish this? Basically, my question is, will async_send_to keep a copy of my shared_ptr so that the data doesn't get deleted? Do I need to use a particular overload of the boost::asio::buffer function to ensure that a copy of the shared_ptr is stored?

Here is an example:

boost::asio::streambuf Buffer;
// ... buffer is filled with data ...
int BufferSize = Buffer.size();
boost::shared_ptr<unsigned char> SharedBuffer(new unsigned char[BufferSize]);
Buffer.sgetn(*SharedBuffer, BufferSize);
Buffer.consume(BufferSize);
Socket.async_send_to(boost::asio::buffer(*SharedBuffer, BufferSize), DestinationEndpoint, boost::bind(&MyClass::OnPacketSent, this, boost::asio::placeholders::error, boost::asio::placeholders::bytes_transferred));
// Shared buffer goes out of scope here, immediately!

_______________________________________________
Boost-users mailing list
Boost-users@...
http://lists.boost.org/mailman/listinfo.cgi/boost-users

Re: [Fusion] Views of Associative Sequences

2009-11-10T13:17:43Z

Christopher Schmidt wrote:
>
> Unfortunately the changes in <boost/fusion/view/joint_view/*> are not
> stand-alone, and I was too late to merge them into the 1.41 release branch.

OK.

> I suggest you to work with a copy of the HEAD of the trunk.
>

Sorry for troubling you with more trivialities, but by HEAD do you mean
the latest update?

Could you please confirm if this is the link from which I should copy
fusion to my current release so I can start using the new facilities you
have provided:

http://svn.boost.org/svn/boost/trunk/boost/fusion/

Thanks.

_______________________________________________
Boost-users mailing list
Boost-users@...
http://lists.boost.org/mailman/listinfo.cgi/boost-users

Re: Two questions on boost::function

2009-11-10T12:32:43Z

AMDG

CHRISTOPHE KOUDELLA, CITADEL INVESTMENT G wrote:
> Hi, I am hopng to get some help from boost::function users on two questions I have.
>
> First, is it possible to extract the underlying functor out of a function? The reason for this is that it is useful to have the boost::function abstraction to move around functors in a generic fashion, but the interface is obviously limited to operator(), so being able to extract the underlying class out of a boost::function would be useful. I am thinking mostly about applied math applications in which functions are defined by more than an operator, whence the interface of boost::function is sometimes insufficient.
>

If you have to extract the underlying type, boost::function is probably
not appropriate.

> Second, I was looking to implement a "function algebra" for boost::function (of a given arity, for a start). This algebra would implement sum, difference, product, composition, etc... operations on a pair of like boost::functions. In a template implementation in order to implement the operator() of the result boost::function, I need to access the result type and the argument types of the boost::functions being operated on. So, suppose I want to implement a product operation on a pair of template functors, then in order to define the operator() of the product of the functors, I need the result and argument types. Does boost::function make these types available as some typedefs? Does what I say make at all sense or would you proceed entirely differently in order to define such operations on boost::function?
>

Something like this can be done using Boost.Lambda.

boost::function<int(int)> f1, f2, f3;
//...
using namespace boost::lambda;
// sum
f3 = bind(f1, _1) + bind(f2, _1);
// composition
f3 = bind(f1, bind(f2, _1));

In Christ,
Steven Watanabe

_______________________________________________
Boost-users mailing list
Boost-users@...
http://lists.boost.org/mailman/listinfo.cgi/boost-users

Two questions on boost::function

2009-11-10T11:58:11Z

Hi, I am hopng to get some help from boost::function users on two questions I have.

First, is it possible to extract the underlying functor out of a function? The reason for this is that it is useful to have the boost::function abstraction to move around functors in a generic fashion, but the interface is obviously limited to operator(), so being able to extract the underlying class out of a boost::function would be useful. I am thinking mostly about applied math applications in which functions are defined by more than an operator, whence the interface of boost::function is sometimes insufficient.

Second, I was looking to implement a "function algebra" for boost::function (of a given arity, for a start). This algebra would implement sum, difference, product, composition, etc... operations on a pair of like boost::functions. In a template implementation in order to implement the operator() of the result boost::function, I need to access the result type and the argument types of the boost::functions being operated on. So, suppose I want to implement a product operation on a pair of template functors, then in order to define the operator() of the product of the functors, I need the result and argument types. Does boost::function make these types available as some typedefs? Does what I say make at all sense or would you proceed entirely differently in order to define such operations on boost::function?

Thanks for any input,
CK
_______________________________________________
Boost-users mailing list
Boost-users@...
http://lists.boost.org/mailman/listinfo.cgi/boost-users

Re: BGL - parallel/distributed

2009-11-10T11:56:06Z

> To be even more precise, it exits the code in function MPI_Comm_dup
> of mpi library. On linux I guess you have to have the mpi running,
> but I don't know how it is on windows with mpich2.
>

You never initialized the Boost.MPI environment. See the tests, but
you're looking for something like:

mpi::environment env(argc, argv);

This basically functions like MPI_Init() plus some Boost.MPI
initialization. Comm_dup fails because the communicator you're trying
to dup doesn't exist.

-Nick

_______________________________________________
Boost-users mailing list
Boost-users@...
http://lists.boost.org/mailman/listinfo.cgi/boost-users

boost::graph - Newbie question about vertex properties

2009-11-10T08:13:04Z

Hello dear mailing list members,

i'm new to this mailing list and hope i'm right here and that my english is not
too annoying.

I have to to some distance calculation and decided to use the
boost::graph-framework to do this as i have to develop a C++ application using
Visual C++ Express Edition 2008. I have to include some background details
information in the vertices and edges of my graph and chose to append such meta
info objects to the vertices and edges with the help of properties. I defined
the following types in the Header-Class:

[code]
struct vtx_ref_t {
typedef boost::vertex_property_tag kind;
};
struct edge_ref_t {
typedef boost::edge_property_tag kind;
};

// VtxMetaClass and EdgeMetaClass are some Classes containing some
// infos needed for calculation
typedef boost::property<vtx_ref_t, VtxMetaClass*> VtxProperty;
typedef boost::property<edge_ref_t, EdgeMetaClass*> EdgeProperty;
typedef boost::adjacency_list<boost::vecS, boost::vecS,
boost::directedS, VtxProperty, EdgeProperty> MyGraphType;

typedef boost::graph_traits<MyGraphType>::vertex_descriptor vertex_t;
typedef boost::graph_traits<MyGraphType>::edge_descriptor edge_t;
[/code]

In the Main-Method i tried to use the function boost::put to append such a meta
class object infoobj to a vertex (whose vertex descriptor vtx i only have) i
have created somewhere else where i cannot append the infoobject cause i don't
know at that time which one to append.

[code]
MyGraphType graph(100);
vertex_t vtx = boost::add_vertex(graph);
VtxMetaClass* infoobj;
...
boost::property_map<MyGraphType, vtx_ref_t>::type vtxref =
get(vtx_ref_t(), graph);
boost::put(vtxref, graph, vtx, infoobj);
[/code]

But i only get a compilation error telling me that:
1>e:\development\VC++\project\include\boost/pending/property.hpp(46) : error
C2039: 'kind': is no element of
'boost::vec_adj_list_vertex_property_map<Graph,GraphPtr,ValueType,Reference,Tag>'

What is wrong with that, i did what was suggested in the online manual ?

I know there's also the possibility to use boost::put with the index of the
vertex that has to be manipulated, but i don't know how to query the index from
the vertex_descriptor i have, perhaps you came across a way to do this yet ...

Perhaps you could help me with this problem ?
Thank you very much in anticipation,
Stephen

_______________________________________________
Boost-users mailing list
Boost-users@...
http://lists.boost.org/mailman/listinfo.cgi/boost-users

Re: [test] Linking against boost_unit_test_framework causes double-deletion in std::ostringstream

2009-11-10T07:33:35Z

Roman Perepelitsa wrote:

> 2009/11/10 Alastair Rankine <arsptr@...>
>>Just curious if anyone else has seen this or anything like it? Platform
>>is Mac OS X 10.6.1, gcc 4.2.1, boost 1.40.0 installed from MacPorts,
>>compiling 64-bit (default).
I have crashes and "double deletions" in my unit tests, too.
I am running Linux, gcc 4.2.1, Boost 1.41.0-beta1;
(I have no crashes or double deletions with Boost 1.40.0.)

> As far as I remember, boost_unit_test_framework defines main function.
> Having two main functions in the same application might be what's causing
> your problems.
Definitely no two mains; 'my' main comes from a macro provided by
Boost.Test.

> Another possibility is that the library and the program are
> compiled with different standard libraries.
I have built both boost and my programs/tests myself, so I can rule the
"different standard library" thing out.

Best regards

_______________________________________________
Boost-users mailing list
Boost-users@...
http://lists.boost.org/mailman/listinfo.cgi/boost-users

Re: [BGL] Schultes and Sanders Contraction Hierarchies Routing

2009-11-10T05:40:05Z

Has anyone looked at supporting Drs Schultes and Sanders Contraction Hierarchies Routing using Boost Graph?

No. For some reason, good results in graph algorithms don't filter down to us too quickly :)

If Andrew, Jeremiah or others have time to look at some of the papers, I would be interested in your thoughts on supporting this or even knowing that BGL might be the wrong tool to implement this with.

I only have time to give the paper a very brief look this morning, but it seems like it might be feasible to reproduce this with the BGL. However, there are so many components to the technique, that it would require careful planning. For example, contraction may be difficult to implement. The vast majority of our use cases target graph construction and query. Fully dynamic graphs... well, they're supported, but maybe not as well as we'd like.

I would certainly be interested to see somebody try to put this together for the BGL. I think it might drive a lot of interesting requirements for the library, and identify areas that we could improve on.

Andrew Sutton
andrew.n.sutton@...

_______________________________________________
Boost-users mailing list
Boost-users@...
http://lists.boost.org/mailman/listinfo.cgi/boost-users

Re: [test] Linking against boost_unit_test_framework causes double-deletion in std::ostringstream

2009-11-10T04:48:49Z

2009/11/10 Alastair Rankine <arsptr@...>

#include <sstream>

int main()
{
std::ostringstream str;
str << "Help Me!";

return 0;
}

Here's what I'm doing:

% g++ -o streamtest_bad -ggdb -L/opt/local/lib -lboost_unit_test_framework-mt-d streamtest.cpp

As far as I remember, boost_unit_test_framework defines main function. Having two main functions in the same application might be what's causing your problems. Another possibility is that the library and the program are compiled with different standard libraries.

Roman Perepelitsa.

_______________________________________________
Boost-users mailing list
Boost-users@...
http://lists.boost.org/mailman/listinfo.cgi/boost-users

Re: running a server "in background mode" ??

2009-11-10T04:29:57Z

>     boost::asio::io_service io_service;
>     tcp_server server(io_service);
>     io_service.run();

io_service::run() does not return until there's no work for io_service.
But you can run io_service in another thread:

new boost::thread(&asio::io_service::run, &io_service); // note that
io_service object can't be local here, as it must outlive the thread!
_______________________________________________
Boost-users mailing list
Boost-users@...
http://lists.boost.org/mailman/listinfo.cgi/boost-users

Re: BGL - parallel/distributed

2009-11-10T04:24:02Z

To be even more precise, it exits the code in function MPI_Comm_dup of
mpi library. On linux I guess you have to have the mpi running, but I
don't know how it is on windows with mpich2.

Jeremiah Willcock wrote:

> On Mon, 9 Nov 2009, Ondrej Sluciak wrote:
>
>> Ok, I tried the simplest possible example, but just didn't figure out
>> where the problem is. I am on windows, mpich2 installed, all boost
>> libs correctly compiled with mpi, but if I do something like
>>
>> #include <boost/graph/use_mpi.hpp>
>> #include <boost/graph/distributed/selector.hpp>
>> #include <boost/graph/distributed/adjacency_list.hpp>
>> #include <boost/graph/distributed/mpi_process_group.hpp>
>>
>> void main(){
>> using namespace boost;
>> typedef
>> adjacency_list<vecS,distributedS<graph::distributed::mpi_process_group,
>> vecS>, directedS> Graph;
>>
>> Graph g(4);
>>
>> };
>>
>> I always get linking problems like:
>> Error 7 error LNK2019: unresolved external symbol "void __cdecl
>> boost::graph::distributed::synchronize(class
>> boost::graph::distributed::mpi_process_group const &)"
>> (?synchronize@distributed@graph@boost@@YAXABVmpi_process_group@123@@Z)
>> referenced in function "public: __thiscall
>> boost::adjacency_list<struct boost::vecS,struct
>> boost::distributedS<class
>> boost::graph::distributed::mpi_process_group,struct
>> boost::vecS,struct boost::defaultS>,struct boost::directedS,struct
>> boost::no_property,struct boost::no_property,struct
>> boost::no_property,struct boost::listS>::~adjacency_list<struct
>> boost::vecS,struct boost::distributedS<class
>> boost::graph::distributed::mpi_process_group,struct
>> boost::vecS,struct boost::defaultS>,struct boost::directedS,struct
>> boost::no_property,struct boost::no_property,struct
>> boost::no_property,struct boost::listS>(void)"
>> (??1?$adjacency_list@UvecS@boost@@U?$distributedS@Vmpi_process_group@distributed@graph@boost@@UvecS@4@UdefaultS@4@@2@UdirectedS@2@Uno_property@2@U52@U52@UlistS@2@@boost@@QAE@XZ)
>> boostTest2.obj
>>
>> I think I have included all the necessary libs (including mpi.lib
>> from mpich directory), so I still don't understand where is the
>> problem. Does anyone tried distributed graph class?
>
> Are you also including the graph_parallel library?
> Boost.Graph.Parallel is not header-only, unlike almost all of the
> sequential Boost.Graph library.
>
> -- Jeremiah Willcock
> _______________________________________________
> Boost-users mailing list
> Boost-users@...
> http://lists.boost.org/mailman/listinfo.cgi/boost-users
>
>

Re: running a server "in background mode" ??

2009-11-10T03:29:57Z

thanks very much for your support!!! i've understood that code, but however when i run it , i don't really see some difference, the server keeps in listening mode, blocking the execution of the rest of the code : i've added two simple line to print something on the console, and what i want is execute the server,that it keeps open in background but inmediatly execute too the printf code lines. main fuction would be :
.
.
.
.

int main()
{
try
{
    boost::asio::io_service io_service;
    tcp_server server(io_service);
    io_service.run();
    printf("Writting a line 1 \n");
    sleep(5);
    printf("Writting a line 2 \n");

}
catch (std::exception& e)
{
    std::cerr << e.what() << std::endl;
}

return 0;
}

2009/11/9 Maxime van Noppen <maxime@...>

pedro chaparro wrote:
> Hi, i need to use a server inside of other code application, i take the
> example from
> http://www.boost.org/doc/libs/1_40_0/doc/html/boost_asio/tutorial/tutdaytime2/src.html
> and it works but the server stay in listening mode and there is other
> code after that i'd like continue executing, is possible tu run the
> server let's say like in a background mode?

Hi,

You should look to how asynchronous operations work :

http://www.boost.org/doc/libs/1_40_0/doc/html/boost_asio/tutorial/tutdaytime3.html

--
Maxime
_______________________________________________
Boost-users mailing list
Boost-users@...
http://lists.boost.org/mailman/listinfo.cgi/boost-users

--
Ing Pedro Alonso Chaparro Valero
Ciudad Politecnica de la Innovación
iTEAM - Mobile Communications Group
Polytechnic University of Valencia
C\ Camino de Vera S/N, Edificio 8G
46022 Valencia, Spain

_______________________________________________
Boost-users mailing list
Boost-users@...
http://lists.boost.org/mailman/listinfo.cgi/boost-users

Re: [Fusion] Views of Associative Sequences

2009-11-10T03:20:29Z

er schrieb:
> I wonder if I should wait for v1.41 (when : is there an estimate?) or if
> I can already remove joint_view.hpp from my current copy of boost and
> replace it with the new one : do you remember if the changes in
> joint_view standalone?

Unfortunately the changes in <boost/fusion/view/joint_view/*> are not
stand-alone, and I was too late to merge them into the 1.41 release branch.
I suggest you to work with a copy of the HEAD of the trunk.

-Christopher

_______________________________________________
Boost-users mailing list
Boost-users@...
http://lists.boost.org/mailman/listinfo.cgi/boost-users

[test] Linking against boost_unit_test_framework causes double-deletion in std::ostringstream

2009-11-10T02:50:20Z

Just curious if anyone else has seen this or anything like it? Platform
is Mac OS X 10.6.1, gcc 4.2.1, boost 1.40.0 installed from MacPorts,
compiling 64-bit (default).

Here's my test code:

#include <sstream>

int main()
{
std::ostringstream str;
str << "Help Me!";

return 0;
}

Here's what I'm doing:

% g++ -o streamtest_bad -ggdb -L/opt/local/lib
-lboost_unit_test_framework-mt-d streamtest.cpp
% ./streamtest_bad
streamtest_bad(63554) malloc: *** error for object 0x7fff7008a500:
pointer being freed was not allocated
*** set a breakpoint in malloc_error_break to debug
zsh: abort ./streamtest_bad
%

Works fine against other boost libraries, eg system:

% g++ -o streamtest_good -ggdb -L/opt/local/lib -lboost_system-mt-d
streamtest.cpp
% ./streamtest_good
%

Note that this seems to be a separate problem to that reported in
ticket 3432 (https://svn.boost.org/trac/boost/ticket/3432). At least, I
have verified that the framework::clear() function isn't being called
in the above code, which does tend to point the finger at a different
part of the library.

The backtrace is not particularly enlightening, but posted here for the
sake of completeness:

(gdb) bt
#0 0x00007fff87f92ff6 in __kill ()
#1 0x00007fff88034072 in abort ()
#2 0x00007fff87f4b095 in free ()
#3 0x0000000100005533 in __gnu_cxx::new_allocator<char>::deallocate
(this=0x7fff5fbff5df, __p=0x7fff7008a500 "") at new_allocator.h:97
#4 0x0000000100005578 in std::string::_Rep::_M_destroy
(this=0x7fff7008a500, __a=@0x7fff5fbff63f) at basic_string.tcc:431
#5 0x0000000100005939 in std::string::_Rep::_M_dispose
(this=0x7fff7008a500, __a=@0x7fff5fbff63f) at basic_string.h:238
#6 0x000000010000596a in std::basic_string<char,
std::char_traits<char>, std::allocator<char> >::~basic_string
(this=0x7fff5fbff670) at basic_string.h:493
#7 0x000000010002ad55 in std::basic_stringbuf<char,
std::char_traits<char>, std::allocator<char> >::overflow
(this=0x7fff5fbff7f8, __c=72) at sstream.tcc:125
#8 0x0000000100029db0 in std::basic_streambuf<char,
std::char_traits<char> >::xsputn (this=0x7fff5fbff7f8, __s=0x100000e7e
"Help Me!", __n=8) at streambuf.tcc:102
#9 0x00007fff806e0ae4 in std::__ostream_write<char,
std::char_traits<char> > ()
#10 0x00007fff806deb98 in std::__ostream_insert<char,
std::char_traits<char> > ()
#11 0x00007fff806dec58 in std::operator<< <std::char_traits<char> > ()
#12 0x0000000100000d6b in main () at streamtest.cpp:6

_______________________________________________
Boost-users mailing list
Boost-users@...
http://lists.boost.org/mailman/listinfo.cgi/boost-users

Re: BGL - parallel/distributed

2009-11-10T02:17:51Z

You are right, that was what I was missing. Thank you. Anyway, no I can
compile it and run it, but the program always exits at the line Graph
g(4) with exit code 0x1. It quits in the inizialization of
mpi_process_group::impl::impl(std::size_t num_headers, std::size_t
buffer_sz, communicator_type parent_comm)
: comm(parent_comm, boost::mpi::comm_duplicate),
oob_reply_comm(parent_comm, boost::mpi::comm_duplicate),
allocated_tags(boost::mpi::environment::max_tag())

exactly in libs/mpi/src/communicator.cpp
communicator::communicator(const ...) on the line
BOOST_MPI_CHECK_RESULT(MPI_Comm_dup, (comm, &newcomm));

It doesn't crash or anything, just exits the code "normally" with exit
code 1.
Do you have any idea what is wrong?

Jeremiah Willcock wrote:

Re: [Fusion] Views of Associative Sequences

2009-11-09T17:45:53Z

>
> There is already a feature request for associative views (ticket 3473).
> I will get down to it once I am granted write access to the trunk.

I see that the ticket is closed and the changes implemented in the trunk
: many thanks!

I wonder if I should wait for v1.41 (when : is there an estimate?) or if
I can already remove joint_view.hpp from my current copy of boost and
replace it with the new one : do you remember if the changes in
joint_view standalone?

_______________________________________________
Boost-users mailing list
Boost-users@...
http://lists.boost.org/mailman/listinfo.cgi/boost-users

Re: Boost.Asio:

2009-11-09T16:52:49Z

This is indeed on Windows, and interestingly enough it does appear to work on other platforms (OSX in particular does not give the error). I'm not familiar with "completion ports" but I looked it up and it appears they are created with "CreateIoCompletionPort" which doesn't exist anywhere in my code. The socket in question is actually created in a fairly normal way:

SOCKET new_sock = INVALID_SOCKET;

new_sock = accept(svr_sock, (struct sockaddr*)&sin, &size);

After that, a few other things are done to the socket -- namely, setting it as non-blocking using ioctlsocket, and using setsockopt for SO_LINGER and SO_OOBINLINE. Any help is greatly appreciated.

On Mon, Nov 9, 2009 at 3:37 AM, Tore Halvorsen <tore.halvorsen@...> wrote:

tiredashell <tiredashell <at> gmail.com> writes:

> The error in the subject line occurs when attempting to assign a native
> socket to a Boost.Asio socket. I realize it's not very helpful since I'm
> not explaining where I am getting nativeSocket, but I was hoping someone
> let me know in general the cause of such an error, since there doesn't
> appear to be any documentation on it. I am almost positive that it is a
> valid socket, and its integer value is normally around 740.

Is this on windows with native socket already mapped to a completion port?

_______________________________________________
Boost-users mailing list
Boost-users@...
http://lists.boost.org/mailman/listinfo.cgi/boost-users

_______________________________________________
Boost-users mailing list
Boost-users@...
http://lists.boost.org/mailman/listinfo.cgi/boost-users

Re: at_thread_exit() never getting called

2009-11-09T15:50:33Z

cowwoc wrote:
> PS: Is there a more reliable mechanism than atexit() that is
> guaranteed to support more than 32 hooks?

You could wrap the whole thing in a class, and use a static variable of that
class so that its destructor will be called on exit. Something like (not
compiled, not tested):

class my_thread
{
private:

boost::scoped_ptr<boost::thread> pt_;
bool shutdown_request_;
// more per-thread state, if needed

void threadproc(); // as before

public:

my_thread(): shutdown_request_( false ) {}

void start()
{
pt_.reset( new boost::thread( &my_thread::threadproc, this ) );
}

~my_thread()
{
if( pt_ )
{
shutdown_request_ = true;
pt_->join();
}
}
};

// ...

static my_thread s_th;

int main()
{

s_th.start();
// ...

} // the destructor of s_th should be called here

_______________________________________________
Boost-users mailing list
Boost-users@...
http://lists.boost.org/mailman/listinfo.cgi/boost-users

getting shape of numpy array

2009-11-09T15:36:16Z

Hi,

I am trying to interaface a numpy array using boost::python::numeric::array.

void function(numeric::array& a)
{
tuple shape=extract<tuple>(a.getshape());
double x;
int i,j;
for(i=0;i<shape[0];++i)
for(j=0;j<shape[1];++j)
x=extract<double>(a[make_tuple(i,j)]);
}

Now I had to add
numeric::array::set_module_and_type("numpy", "ndarray");

to the module, hope that was correct.

When I call this function in python, I get:
AttributeError: 'numpy.ndarray' object has no attribute 'getshape'

What is the correct way to do it?

Thanks!
Nathan

Re: BGL - parallel/distributed

2009-11-09T13:18:58Z

On Mon, 9 Nov 2009, Ondrej Sluciak wrote:

> Ok, I tried the simplest possible example, but just didn't figure out where
> the problem is. I am on windows, mpich2 installed, all boost libs correctly
> compiled with mpi, but if I do something like
>
> #include <boost/graph/use_mpi.hpp>
> #include <boost/graph/distributed/selector.hpp>
> #include <boost/graph/distributed/adjacency_list.hpp>
> #include <boost/graph/distributed/mpi_process_group.hpp>
>
> void main(){
> using namespace boost;
> typedef
> adjacency_list<vecS,distributedS<graph::distributed::mpi_process_group,
> vecS>, directedS> Graph;
>
> Graph g(4);
>
> };
>
> I always get linking problems like:
> Error 7 error LNK2019: unresolved external symbol "void __cdecl
> boost::graph::distributed::synchronize(class
> boost::graph::distributed::mpi_process_group const &)"
> (?synchronize@distributed@graph@boost@@YAXABVmpi_process_group@123@@Z)
> referenced in function "public: __thiscall boost::adjacency_list<struct
> boost::vecS,struct boost::distributedS<class
> boost::graph::distributed::mpi_process_group,struct boost::vecS,struct
> boost::defaultS>,struct boost::directedS,struct boost::no_property,struct
> boost::no_property,struct boost::no_property,struct
> boost::listS>::~adjacency_list<struct boost::vecS,struct
> boost::distributedS<class boost::graph::distributed::mpi_process_group,struct
> boost::vecS,struct boost::defaultS>,struct boost::directedS,struct
> boost::no_property,struct boost::no_property,struct boost::no_property,struct
> boost::listS>(void)"
> (??1?$adjacency_list@UvecS@boost@@U?$distributedS@Vmpi_process_group@distributed@graph@boost@@UvecS@4@UdefaultS@4@@2@UdirectedS@2@Uno_property@2@U52@U52@UlistS@2@@boost@@QAE@XZ)
> boostTest2.obj
>
> I think I have included all the necessary libs (including mpi.lib from mpich
> directory), so I still don't understand where is the problem. Does anyone
> tried distributed graph class?

Are you also including the graph_parallel library? Boost.Graph.Parallel
is not header-only, unlike almost all of the sequential Boost.Graph
library.

-- Jeremiah Willcock
_______________________________________________
Boost-users mailing list
Boost-users@...
http://lists.boost.org/mailman/listinfo.cgi/boost-users