February 2003 - Boost-users

boost::shared_ptr and stl algorithms
by Bill Eidson 15 Apr '05

15 Apr '05

std::vector< boost::shared_ptr<T> > TVec; I want to write a function object to run std::for_each. to do somework against members of T and accumulate results in functor. My attempts have failed in VC7 with memory errors in my operator() method. struct fo { bool operator()( boost::shared_ptr<T>& t ) { return true; } }; Thanks in advance bille

5 9

BGL and reversing / moving edges
by joshrose 17 Dec '03

17 Dec '03

G'day, I was hoping someone might be able to tell me if they've run into any situations using BGL where they needed to move (or reverse (i.e. flip the source and target)) an edge within a graph implemented as an adjacency_list. I'm working on an implementation of an algorithm for max. weight bipartite matching ("the assignment problem"), and one of the assumptions in the algorithm is that non-matched edges are directed from one subgraph, A, to the other subgraph, B, and matched edges are directed the other way. Over the course of growing the matching, a simple way to add and remove edges from the candidate set is just to reverse the edge, thus the need for some way to reverse the edge within BGL. I don't think I can just remove and then re-add the edge in the other direction, primarily because doing so will (I believe) invalidate any edge_descriptors referring to that edge. In particular, there are property maps keeping track of the weight of edges, the predecessors of certain nodes, etc. that use edge descriptors as keys. I've spent a fair bit of time reading the documentation and looking through the implementation of adjacency_lists (incl. edge_list_impl), and I'm not really sure of a way to either move an edge directly within a graph, or even to create a new edge from another s.t. they have the same edge_descriptor (in general a bad thing, probably). Anyone have any thoughts on this? Thanks, -Josh

4 10

bjam and windows command line problem.
by Fernando Cacciola 07 May '03

07 May '03

Hi, I'm having a hard time trying to use bjam to build boost 1.29.0. I have Win98 in Spanish, and I use Borland C++ Builder 5.0 (its bcc32.exe corresponds to is BCC5.5.1) I have boost 1.29.0 installed at: "C:\Libraries\boost\Repository\boost" I've downloaded bjam for Microsoft Windows: >bjam -v Boost.Jam Version 3.1.0. OS=NT. Copyright 1993-2002 Christopher Seiwald and Perforce Software, Inc. Copyright 2001 David Turner. Copyright 2001-2002 David Abrahams. I've temporarily modified the root 'jamfile' so that it builds only the Test library. and I've done: subst d: "c:\archivos de programa\borland\cbuilder5" set BCCROOT=d: (you'll see why in a moment) If I try: c:\>cd C:\Libraries\boost\Repository\boost" C:\Libraries\boost\Repository\boost> bjam -sTOOLS=borland I receive messages of the form: borland-C++-action libs\test\build\bin\libpem.lib\borland\debug\runtime-link-static\threading-s ingle\execution_monitor.obj "d:\bin\bcc32" -j5 -q -c -P -w -Ve -Vx -a8 -v -Od -vi- -tWC -WM- -w-8001 -I"libs\test\build" -I"C:\Libraries\boost\Repository\boost" -I"d:\includ e" -o"libs\test\build\bin\libpem.lib\borland\debug\runtime-link-static\thre ading-single\execution_moni Error E2266: No file names given tor.obj" "libs\test\build\../src\execution_monitor.cpp" C:\WINDOWS\TEMP\JAMTMP00.BAT [2] Unknown command "TOR.OBJ" "LIBS\TEST\BUILD\.." "d:\bin\bcc32" -j5 -q -c -P -w -Ve -Vx -a8 -v -Od -vi- -tWC -WM- -w-8001 -I"libs\test\build" -I"C:\Libraries\boost\Repository\boost" -I"d:\includ e" -o"libs\test\build\bin\libpem.lib\borland\debug\runtime-link-static\thre ading-single\execution_monitor.obj" "libs\test\build\../src\execution_monitor.cpp" ...failed borland-C++-action libs\test\build\bin\libpem.lib\borland\debug\runtime-link-static\threading-s ingle\execution_monitor.obj... If I'm interpreting this output correctly, the problem is that the command line feed to bcc32 exceeds 256 characters. What can be done about this? Is there a way for 'borland-tools.jam' to put the bcc32 parameters in a tmp response file and use: bcc32.exe @tmprsp.rsp? Any help will be appreciated. TIA, -- Fernando Cacciola

4 22

boost/type_traits/config.hpp
by H. Morishima 26 Mar '03

26 Mar '03

My compiler issues the following warning message (boost v1.29.0, G++ 3.2 with warning level high): boost_1_29_0/boost/type_traits/config.hpp:90:55: warning: "BOOST_MSVC" is not defined boost/type_traits/config.hpp: line 90 is: # if (........ || BOOST_MSVC > 1301 || ........) Shouldn't it be as follows? # if (........ || #defined(BOOST_MSVC) && BOOST_MSVC > 1301 || ........) H. Morishima

4 5

adjacency_matrix
by agiatojon 12 Mar '03

12 Mar '03

Hello all, I'm new to the list and to Boost in general and am trying to use an adjacency_matrix in order to plot coordinates (hopefully that is a good use for it). I just started trying to learn it and have entered the sample program from page 237 into my compiler (VS.NET) Below is the program I have entered. I am getting errors in that print_vertices, print_edges, and print_graph are undefined and are not members of boost. Any help apprechiated! Thanks. /** * adjaMatri.cpp * Nov 7th, 2002 * Boost Graph Library, pg. 237 */ #include <iostream> using std::cout; using std::cin; using std::endl; #include <boost/graph/adjacency_matrix.hpp> using namespace boost; int main() { enum { A, B, C, D, E, F, N }; const char *name = "ABCDEF"; typedef boost::adjacency_matrix<boost::undirectedS> UGraph; UGraph ug(N); add_edge(B, C, ug); add_edge(B, F, ug); add_edge(C, A, ug); add_edge(D, E, ug); add_edge(F, A, ug); cout << "vertex set: "; boost::print_vertices(ug, name); cout << endl; cout << "edge_set: "; boost::print_edges(ug, name); cout << endl; cout << "incident edges: "; boost::print_graph(ug, name); cout << endl; cout << "Press enter to exit..."; cin.get(); return EXIT_SUCCESS; } Jon Agiato JonAgiato(a)nyc.rr.com

2 1

Re: [Boost-Users] NDEBUG, affects floating point math?
by Stephen Crowley 05 Mar '03

05 Mar '03

On Fri, Feb 28, 2003 at 09:53:56AM +0100, Joerg Walter wrote: > > I wrote a small piece of C code to calculate a symmetric-matrix by > > row-matrix product by directly accessing the memory (which btw is about > > 1000% faster than ublas's generated code with gcc 3.2.3 and -O3, why is > > ublas so slow?) > > If you have *not* defined NDEBUG, ublas uses many internal consistency > checks and usually no expression templates. I was referring to the case when NDEBUG is defined and comparing symmetric_matrix by general_matrix products. uBLAS is quite slow with this operation. However, I compared my symmetric_matrix by general_matrix_transpose product and ublas was only 3-4 times slower than the optimized C, not too bad. > > And when I compile with -DNDEBUG the results differ > > by around +/- 1e-16, but when its not defined the results match almost > > perfectly. > > You've probably hit a bug, may be an old one (please consider to check > against CVS, too). Could you please send or post a small sample, if the > problem persists? I think I know the source of the problem now, and I'm not sure if it's one that boost should handle or not. I attached my test code at the end of this email, only 2.7k, I hope it isn't inappropriate for the list. 387 floating point registers are stored as 80 bits internally and rounded to 64 bits when assigned to memory. SSE2 floating-point registers are stored in the native 64-bit format. The following table lists the compiler flags used and the accumulated absolute difference between the two products over 100 randomly generated matrices. gcc version 3.2.3 20030221 (Debian prerelease) no compiler flags diff sum [5,3]((0,0,0),(0,0,0),(0,0,0),(0,0,0),(0,0,0)) -DNDEBUG diff sum [5,3]((0,0,0),(0,0,0),(0,0,0),(0,0,0),(0,0,0)) -O1 diff sum [5,3]((2.91694e-15,2.65348e-15,2.7027e-15),(4.4049e-15,3.29251e-15,3.44863e-15),(4.10436e-15,4.14816e-15,3.24393e-15),(3.57873e-15,3.75741e-15,2.64068e-15),(5.03091e-15,2.79637e-15,4.06099e-15)) -O1 -DNDEBUG diff sum [5,3]((0,0,0),(0,0,0),(0,0,0),(0,0,0),(0,0,0)) -O2 diff sum [5,3]((4.37584e-15,4.8659e-15,3.61516e-15),(4.61783e-15,3.77736e-15,4.04169e-15),(4.46691e-15,5.10876e-15,5.00576e-15),(3.64118e-15,4.91968e-15,5.17641e-15),(3.97273e-15,4.25528e-15,3.75394e-15)) -O2 -DNDEBUG diff sum [5,3]((0,0,0),(0,0,0),(0,0,0),(0,0,0),(0,0,0)) -O3 diff sum [5,3]((4.96131e-15,3.88968e-15,4.34245e-15),(5.07754e-15,5.91974e-15,4.19109e-15),(4.89539e-15,4.29279e-15,5.06344e-15),(4.79781e-15,4.66033e-15,3.77996e-15),(4.75227e-15,4.72539e-15,3.33414e-15)) -O3 -DNDEBUG diff sum [5,3]((4.60569e-15,2.67928e-15,2.7079e-15),(3.21054e-15,2.98112e-15,2.61249e-15),(3.91701e-15,3.60822e-15,2.93515e-15),(3.23006e-15,3.42434e-15,3.53363e-15),(3.18105e-15,2.77035e-15,3.67198e-15)) -O1 -mfpmath=sse,387 -march=pentium4 -msse2 diff sum [5,3]((0,0,0),(0,0,0),(0,0,0),(0,0,0),(0,0,0)) -O1 -mfpmath=sse,387 -march=pentium4 -msse2 -DNDEBUG diff sum [5,3]((0,0,0),(0,0,0),(0,0,0),(0,0,0),(0,0,0)) -O2 -mfpmath=sse,387 -march=pentium4 -msse2 diff sum [5,3]((0,0,0),(0,0,0),(0,0,0),(0,0,0),(0,0,0)) -O2 -mfpmath=sse,387 -march=pentium4 -msse2 -DNDEBUG diff sum [5,3]((0,0,0),(0,0,0),(0,0,0),(0,0,0),(0,0,0)) -O3 -mfpmath=sse,387 -march=pentium4 -msse2 diff sum [5,3]((0,0,0),(0,0,0),(0,0,0),(0,0,0),(0,0,0)) -O3 -mfpmath=sse,387 -march=pentium4 -msse2 -DNDEBUG diff sum [5,3]((3.62991e-15,3.74006e-15,4.62824e-15),(4.00548e-15,4.5311e-15,3.60129e-15),(3.48332e-15,5.11743e-15,3.08174e-15),(3.49373e-15,4.05752e-15,4.2865e-15),(4.81169e-15,4.51549e-15,4.18589e-15)) Is this perhaps a compiler bug? Part of the problem actually comes from whether the products are accumulated into a register by the compiler, or whether they are accumulated into memory. If accumulated into a 387 register it has higher-precision before it is rounded back down to 64-bits to be stored into memory. If the compiler generates code to accumulate the products directly into memory then the results of each individual product are rounded to 64-bits at every iteration and the result will be slightly less accurate. Now, when the -msse2 flag is used, the registers are 64-bit only and thus no accuracy is gained by using them, only speed. -- Stephen [Non-text portions of this message have been removed]

3 4

Re: [Boost-Users] stop a thread
by William E. Kempf 03 Mar '03

03 Mar '03

> From: "hlh771 <hlh771(a)yahoo.com.cn>" <hlh771(a)yahoo.com.cn> > > How to stop(or cancel) a thread? Suppose I have a main thread and a > worker thread, the worker thread does something looply,and when the > main thread get a message indicating the process should exit, I want > stop the worker thread,how to do that?I have read other thread > library,their thread class supply a method called stop or cancel > which stops the thread. Their Implementation is:the thread class has > a data member indicating whether the thread should stop,and the > thread function looply check the data member whether the thread > should stop.Generally the data member's type is a bool or Event. > Is it the best way to stop the thread? How the boost to deal with > this? > Thanks in advance. Today you should use a flag and a condition/mutex combination. Direct support for cancellation is being provided in a future release. William E. Kempf wekempf(a)cox.net

3 4

Exporting references into boost::python
by Nicholas Francis 02 Mar '03

02 Mar '03

A newbie question regarding boost::python. I need to make a templatized function that returns a boost::python::object containing a reference to an object I already have. Basically: template<class T> object convertToPython (T &obj) { return object (obj) } All classes I pass through T are already wrapped (they are also not copy-constructible). I tried with the code found at python.orgs wiki: template<class T> T& identity(T& x) { return x; } template<class T> object get_object_reference (T& x) { object f = make_function (&identity<T>, return_value_policy<reference_existing_object()); return f(x); } this code gave me a compile error (on CW 8.3, Mac OS X). Apparently CW is convinced that the &identity<T> maps to void. I then tried to hardcode a type just to see if it worked at all: InputManager &identity (InputManager &x) { return x; } object get_object_reference (InputManager &x) { object f = make_function ( &identity, return_value_policy<reference_existing_object()); return f(x); } This compiled, but gave me a typeerror at runtime: (No to_python (by-value) converter found for c++ type: InputManager) I've seccesfully wrapped functions like this: InputManager &GetInputManager(); BOOST_PYTHON_MODULE (Engine) { def ( "GetInputManager", &GetInputManager, return_value_policy<reference_existing_object() ); } So it should be able to work. Any help would be greatly appreciated.... TIA, Nicholas

2 1

VC7 says 'fatal C1067' compiling multi_array?
by Clift Norris 01 Mar '03

01 Mar '03

Greetings: I'm using VC7 and boost_1_29_0 to compile this code that is using multi_array.hpp. typedef boost::multi_array<double, 4> theta4_t; typedef theta4_t::index index; theta4_t Theta_Trun( boost::extents[10000][2][10][256] ); I'm receiving this error: "fatal error C1067: compiler limit: debug information module size exceeded" Of course, the code compiles fine using the "Release" configuration, but I need the "Debug" one also. Is there a workaround for this, other than the 'make your variables shorter' suggested in the VC7 documentation? Thanks, -Clift

2 1

are namespaces of boost libs required to be lowercase?
by Markus Werle 28 Feb '03

28 Feb '03

Hi! If someone wants to contribute code to boost: Is she required to provide the names of all namespaces lowercase? E.g. someone provides Elefant, the big library. Is she allowed to have namespace Elefant and perhaps namespace Elefant::Private, etc. I could not find any hint except this here, which did not really help me, since I do not know the standard library's conventions. Any link/ hint appreciated: -- citation from the boost docs -- "Naming conventions rationale The C++ standard committee's Library Working Group discussed this issue in detail, and over a long period of time. The discussion was repeated again in early boost postings. A short summary: Naming conventions are contentious, and although several are widely used, no one style predominates. Given the intent to propose portions of boost for the next revision of the C++ standard library, boost decided to follow the standard library's conventions." best regards, Markus

4 6