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

boost lib naming conventions on OSX (darwin)
by Simmons, Aaron 10 Sep '04

10 Sep '04

I have built and installed boost on Mac OSX (using the 'darwin' toolset). I've noticed that the libboost libraries that are built don't follow the boost naming conventions. Specifically, they don't include the 'osx' toolset identifier as specified in the 'Getting Started' documentation at http://boost.org/more/getting_started.html#Results Is this a known issue? Thanks, Aaron Simmons

1 0

Building Boost on Mac OS X
by John Meinel 10 Sep '04

10 Sep '04

I've been trying to build boost-1_31_0 on Mac OSX (10.3.5). I first tried with the default compiler, and then I installed XCode 1.5 to get gcc build 1666, which seems to be what is listed on the main site. After all this, the problem I get is: c++ unrecognized option `-shared' As near as I can tell, it builds everything except for the shared libraries. I would really prefer to use the shared libraries, though if I can get my program to compile against the static it probably would be okay. Does anywone know of a workaround? I read something online that said "try using -G instead of -shared", but that just gave me -G as an unrecognized option. I'm trying to build CVS right now, but I don't know how stable/etc it is. I would much rather use a release, because if I download it later, I know it will still be the same code. Any pointers would be appreciated. John =:->

2 2

RE: [Boost-users] Help with Path Constructor
by John Wilkinson 09 Sep '04

09 Sep '04

Is this a Unicode build? I'm assuming not, since you don't cast the result of GetBuffer, but your code will not work properly with Unicode. In any case, you don't need to use GetBuffer to get a pointer (const TCHAR *) to the buffer; just cast to LPCTSTR (which works for either Unicode or multibyte). It also might help to know what version of Visual Studio you're using. John -----Original Message----- From: boost-users-bounces(a)lists.boost.org [mailto:boost-users-bounces@lists.boost.org] On Behalf Of RYAN ROSARIO Sent: Wednesday, September 08, 2004 11:30 PM To: boost-users(a)lists.boost.org Subject: [Boost-users] Help with Path Constructor My MFC has a text box that stores the result from a "choose directory" dialog box. I encounter problems when I try to create a path object from the value of this textbox. The debugger shows the value of scanPath (what I call my path object) to be "" which is then causing hangs and exceptions... Below is the code I am trying to use. I do not see why the constructor will not recognize the parameter I am passing to it: //m_scanPath is a control variable for the textbox containing the directory. CString temp2; //Since the stuff in the textbox is a CString. m_scanPath.GetWindowText(temp2); //Get the directory entered in the textbox. const char * tempString = temp2.GetBuffer(); //Convert it to a const char *. fs::path scanPath(tempString); //This does not work even though the param is a const char *) What am I doing wrong? I imagine it is some type of type casting problem? Thanks in Advance, Ryan Do you Yahoo!? Express yourself with Y! Messenger! Free. Download now.

3 4

autoconf m4 macro for boost
by Rodolfo Schulz de Lima 09 Sep '04

09 Sep '04

(answering the mail from Thomas Lemaire) I've attached to this mail the .m4 macros that I made to configure boost automatically in my autoconf/automake projects. It's not complete, nor optimal, but suit my needs. Put it in your /usr/share/aclocal directory and in your configure.ac put: RS_BOOST([1.32.0],[],[AC_MSG_ERROR([Package requires the Boost C++ libraries!])]) RS_BOOST_REGEX RS_BOOST_THREAD RS_BOOST_DATE_TIME RS_BOOST_PROGRAM_OPTIONS it'll create acsubst'ed BOOST_CPPFLAGS, BOOST_LIBS and BOOST_LIBS_R Since I use it to compile for ming32 (win32) and gcc (linux) targets, it's good to have AC_CANONICAL_TARGET right before AM_INIT_AUTOMAKE if you want the m4 to choose between mgw and gcc libraries according to the target you're building for. And if you define something like AC_ARG_ENABLE(debug,....), thus defining enable_debug variable, the m4 will select the debug/release libraries whether enable_debug is yes or no. Then you should put AM_CPPFLAGS=@BOOST_CPPFLAGS@, LIBS=@BOOST_LIBS@ (or @BOOST_LIBS_R@ if you want the recurrent (thread-safe) version of the libraries) somewhere in a Makefile.am That's it... I hope it'll be as useful to you as it is being to me. Regards, rod

1 0

RE: [Boost-users] Help with Path Constructor
by John Wilkinson 09 Sep '04

09 Sep '04

Look in Project/<project name> Properties. In the resulting dialog, pick General under Configuration Properties. In the listbox on the right, look at "Character Set", which will tell you whether or not you are using Unicode. I think that Visual Studio defaults to Unicode builds these days, so you should really check this to make sure. A binary dump of the contents of temp2 after the GetWindowText call might be helpful (binary dump so that we can tell if it's Unicode or not). Try this block of code in place of your original: //m_scanPath is a control variable for the textbox containing the directory. CString temp2; //Since the stuff in the textbox is a CString. m_scanPath.GetWindowText(temp2); //Get the directory entered in the textbox. fs::path scanPath((LPCTSTR) temp2); //This does not work even though the param is a const char *) Let me know if it compiles, and what happens when you run it. John -----Original Message----- From: boost-users-bounces(a)lists.boost.org [mailto:boost-users-bounces@lists.boost.org] On Behalf Of RYAN ROSARIO Sent: Thursday, September 09, 2004 12:36 PM To: boost-users(a)lists.boost.org Subject: RE: [Boost-users] Help with Path Constructor I am using Visual Studio 2003. I don't think it is a Unicode build, but I am not totally sure. John Wilkinson <jwilkinson(a)tsystem.com> wrote: Is this a Unicode build? I'm assuming not, since you don't cast the result of GetBuffer, but your code will not work properly with Unicode. In any case, you don't need to use GetBuffer to get a pointer (const TCHAR *) to the buffer; just cast to LPCTSTR (which works for either Unicode or multibyte). It also might help to know what version of Visual Studio you're using. John -----Original Message----- From: boost-users-bounces(a)lists.boost.org [mailto:boost-users-bounces@lists.boost.org] On Behalf Of RYAN ROSARIO Sent: Wednesday, September 08, 2004 11:30 PM To: boost-users(a)lists.boost.org Subject: [Boost-users] Help with Path Constructor My MFC has a text box that stores the result from a "choose directory" dialog box. I encounter problems when I try to create a path object from the value of this textbox. The debugger shows the value of ! scanPath (what I call my path object) to be "" which is then causing hangs and exceptions... Below is the code I am trying to use. I do not see why the constructor will not recognize the parameter I am passing to it: //m_scanPath is a control variable for the textbox containing the directory. CString temp2; //Since the stuff in the textbox is a CString. m_scanPath.GetWindowText(temp2); //Get the directory entered in the textbox. const char * tempString = temp2.GetBuffer(); //Convert it to a const char *. fs::path scanPath(tempString); //This does not work even though the param is a const char *) What am I doing wrong? I imagine it is some type of type casting problem? Thanks in Advance, Ryan Do you Yahoo!? Express yourself with Y! Messenger! Free. Download now. _______________________________________________ Boost-users mailing list Boost-users(a)lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users Do you Yahoo!? New and Improved Yahoo! Mail - Send 10MB messages!

1 0

Help with Path Constructor
by RYAN ROSARIO 09 Sep '04

09 Sep '04

My MFC has a text box that stores the result from a "choose directory" dialog box. I encounter problems when I try to create a path object from the value of this textbox. The debugger shows the value of scanPath (what I call my path object) to be "" which is then causing hangs and exceptions... Below is the code I am trying to use. I do not see why the constructor will not recognize the parameter I am passing to it: //m_scanPath is a control variable for the textbox containing the directory. CString temp2; //Since the stuff in the textbox is a CString. m_scanPath.GetWindowText(temp2); //Get the directory entered in the textbox. const char * tempString = temp2.GetBuffer(); //Convert it to a const char *. fs::path scanPath(tempString); //This does not work even though the param is a const char *) What am I doing wrong? I imagine it is some type of type casting problem? Thanks in Advance, Ryan --------------------------------- Do you Yahoo!? Express yourself with Y! Messenger! Free. Download now.

2 2

boost installation
by Sekhara Pillai, Ramya Sobhana 09 Sep '04

09 Sep '04

I just started using the boost. For my project I need to install boost -1.30.2 But I am getting some issues in the installation. I will tell you the steps I did over here 1. run the build.bat in boost-jam-3.1.10 2.build vcvrs32.bat in ../vc7/bin 3. Change dir to the dir ../boost.1.30.2 4. Then did C:\NonProject\technical\DOC_NET\Boost\boost-1.30.2>bjam "-sTOOLS=vc7" install --------------------------------------------------------------------- skipping Boost.Python library build due to missing or incorrect configuration couldn't find Python.h in "/usr/include" You can configure the location of your python installation by setting: PYTHON_ROOT - currently "/usr" PYTHON_VERSION - The 2-part python Major.Minor version number (e.g. "2.2", NOT "2.2.1") - currently "2.2" The following are automatically configured from PYTHON_ROOT if not otherwise set: PYTHON_LIB_PATH - path to Python library object; currently "/usr/libs" --------------------------------------------------------------------- don't know how to make install ...found 1 target... ...can't find 1 target... Please help me for this

4 3

boost signals dll does not export all functions
by Gast 128 09 Sep '04

09 Sep '04

Dear all, Probably already a known problem, but if one uses the signals library as a dll, the dll does not contain all exports (e.g. signal_base_impl::~signal_base_impl()). Tracking this down it looks like that the file 'signal_base.cpp' does not see that it has to create exported functions. This is due to the fact that 'BOOST_HAS_DECLSPEC' is not defined, which is defined a compiler specific header file ('visualc.hpp'). However another cpp file from signals, 'connection.cpp', does create exported function. Thus as a work around I added '#include <boost/signals/detail/signals_common.hpp>' as first include in 'signal_base.cpp' (and also to the other cpp files of signals) and bjamed again. Will this issue resolved in the next release? Wkr, me

2 1

Directory_Iterator -> hangs app
by RYAN ROSARIO 09 Sep '04

09 Sep '04

I am still stuck with this issue of a directory_iterator causing a hang when it is declared. This is one of the most bizarre things I have ever encountered. Some more clues have arisen since my last post: - Upon hanging, I now get a message box: Unhandled exception at 0x77e73887 in PICtool-w32.exe: Microsoft C++ exception: boost::filesystem::filesystem_error @ 0x0012f170. I seem to be getting *somewhere* although the message seems very vague. - Upon executing the app in the explorer (not MSVC), I get a Runtime Error: The app has requested runtime to terminate in an unusual way. - I ran Spy++ and see to get quite a few WM_GETDLGCODE messages. I have read that this can signify a problem. Has anybody else had a problem like this with the directory_iterator? Any help or suggestion is greatly appreciated as I have been stuck for quite a while. :-( FYI: MSVC .NET 2003, running as Multi-Threaded Debug DLL using the multi-threaded version of the Boost::Filesystem library. Thanks Again, Ryan --------------------------------- Do you Yahoo!? Win 1 of 4,000 free domain names from Yahoo! Enter now.

4 9

Unit Test Framework for non-boost projects?
by Steven T. Hatton 08 Sep '04

08 Sep '04

I believe I figured out how the examples in the 'subproject libs/test/example ;' are being compiled. Basically it goes like this: bjam reads the Jamfile which contains this rule: rule boost-test-example ( example-name : lib-name ) { exe $(example-name) : $(example-name).cpp <lib>../build/$(lib-name) : <sysinclude>$(BOOST_ROOT) <stlport-iostream>on <borland><*><cxxflags>-w-8080 ; } followed by a series of invocation of that rule such as: boost-test-example unit_test_example1 : boost_unit_test_framework ; The rule is evaluated by putting the first argument 'unit_test_example1' in the slot for the target name, and for the sources required to build. In addition to unit_test_example1.cpp, the expression <lib>../build/$(lib-name) is evaluated to add the library resulting from compiling other source files. These are named by reading $(BOOST_ROOT)/libs/test/build/Jamfile and expanding the target: lib boost_unit_test_framework : <template>boost_test_lib ../src/$(UTF_SOURCES).cpp ; The <template> adds some boiler-plate information and the ../src/$(UTF_SOURCES).cpp expands the following list: UTF_SOURCES = execution_monitor test_tools unit_test_parameters unit_test_log unit_test_monitor unit_test_result unit_test_suite unit_test_main supplied_log_formatters ; to create a list of file names of the form: ../src/execution_monitor.cpp If the library does not already exists, these are compiled, and the results are placed in the location specified by this directive: install test lib : [ unless $(NT) : <dll>boost_prg_exec_monitor <dll>boost_test_exec_monitor <dll>boost_unit_test_framework ] <lib>boost_prg_exec_monitor <lib>boost_test_exec_monitor <lib>boost_unit_test_framework ; Since they have already been compiled, bjam tells the compiler to link against them. After rebuilding all of boost and installing the headers and libraries to the default location, I am now able to compile one of the examples as follows: g++ -o unit_test_example5 unit_test_example5.cpp\ -lboost_unit_test_framework-gcc I don't fully understand why I need the -l switch, but adding it enabled the linker to succeed. I assume I could find a way to modify the above mentioned Jamfiles to do basically the same thing in my project directory, but the step of specifying the sources with which to build the libraries seems inappropriate for using the Unit Test Framework in my own project. How do I tell bjam to link against the correct library file? -- Regards, Steven

2 2

constant size arrays - why never a simulation of push pop?
by Frank Maddin 08 Sep '04

08 Sep '04

Maybe there is a good reason for this, but I've often wondered why there is no version of constant size arrays that support push_back() and pop_back(). In the past I've wanted to use a c_array in situations, but couldn't because at some point I wanted to use an algorithm or template function that used push/pop. So I end up using vector/reserve. I can't think of any problem off the top of my head. Seems like just some management of current size(), iterator manipulation etc. Maybe there is some reason I've never seen this or perhaps there is something I'd run into if I tried to roll my own based off of boost::array or c_array? Thanks Frank

2 1

ublas identity_matrix problem
by Heiko Bauke 08 Sep '04

08 Sep '04

Dear all, I discovered a problem in ublas identity_matrix class. I am doing linear algebra in finite fields. So my matrices contain no floats nor doubles. Instead I use some class that behaves like a number but does arithmetic in a finite field. Class instances are contructable by int, but _not_ default constructable by int. I don't want to have this. In matrix.hpp one finds the lines template<class T> typename identity_matrix<T>::value_type identity_matrix<T>::zero_=0; template<class T> typename identity_matrix<T>::value_type identity_matrix<T>::one_=1; These lines are problematic and the reason why my number class works not with the identity_matrix class. I think these lines should by changed into template<class T> typename identity_matrix<T>::value_type identity_matrix<T>::zero_=T(0); template<class T> typename identity_matrix<T>::value_type identity_matrix<T>::one_=T(1); to make the ublas code more general. with regards Heiko -- -- Gute Sitten haben für die Gesellschaft mehr Wert als alle -- Berechnungen Newtons. (Friedrich II., der Große, 1712-1786) -- Supercomputing in Magdeburg @ http://tina.nat.uni-magdeburg.de -- Heiko Bauke @ http://www.uni-magdeburg.de/bauke

1 0

How to remove custom property map from graph dynamically ?
by Raj 08 Sep '04

08 Sep '04

Hi, I have an adjacency list graph with three custom property maps attached. At some point during the run of the program, I would like to remove two of those maps from the graph. Is it possible ? How can I do it ? Thanks in advance Raja

4 4

RE: [Boost-users] Re: statically linking to Boost.Threads with msvc7
by Simmons, Aaron 08 Sep '04

08 Sep '04

> In Boost release 1.31, the Boost.Threads library doesn't support static > linking. This limitation will be removed in the next release (coming > soon) for: > > a) anyone using the pthreads implementation of Boost.Threads > b) anyone using Visual C++ on Win32 > > Hopefully it will be removed for other Win32 compilers in the future. How was the TSS problem solved for win32-threads? What's the ETA for the next version? Thanks, aaron

2 1

macro m4 for boost
by Thomas Lemaire 08 Sep '04

08 Sep '04

Hi boost-users, I am looking for a macro m4 (used by autoconf) that automatically detects boost installation directory. Has anybody got it ? thomas

1 0

dynamically created program options
by Rodolfo Schulz de Lima 08 Sep '04

08 Sep '04

Hi, I wonder if it's possible to handle dynamically created program options with boost::program_options. For instance: I have a config file that specifies configure options for dynamic loaded modules in my project. So the cfg file would be: [module.foo1] module_file = foo.so listen_port = 564 enabled = true [module.foo2] module_file = foo.so listen_port = 565 enabled = false [module.bar1] module_file = bar.so listen_port = 145 enabled = true As can be seen, foo1, foo2 and bar1 names are defined by the user in the configure file, and the program doesn't know about them before parsing the configuration. But if we parse that (somehow), an exception saying that the option "module.foo1.module_file" is invalid will be thrown. Can I manage the options_description "on the fly"? Another less complicated situation would be something like this: modules = foo1 foo2 bar1 [module.foo1] . . . [module.foo2] . . . [module.bar1] . . . So when parsing the "modules" option, we would know the new options name and it would be a matter of appending more option descriptions to the object. Regards, rod

2 1

Min-Cut o Max-Flow in a graph?
by Daniele Carlucci 08 Sep '04

08 Sep '04

Hi, How can I find the min-cut, or the max-flow in a graph with all the costs of edges equal one? Thanks Daniele

1 0

Comiling Boost with VC++.net
by Andreas Kempf 07 Sep '04

07 Sep '04

My IDE for C++ is VC++.net I heard that it is impossible to compile Boost with its compiler. I replaced the compiler with the VC++.net 2003 Toolkit Compiler, but it still doesn't work. Can anyone help me compiling with the VC++.net 2003 Toolkit?

2 1

RE: [Boost-users] signal slot flexibility
by Frank Maddin 07 Sep '04

07 Sep '04

Ahh, so now I'm able to do such things as this. Hello* h_ptr = new Hello(); World* w_ptr = new World(); typedef boost::signal< int (int, float) > IntFloatSignal; IntFloatSignal sig; sig.connect(boost::bind(&Hello::alternate_int_float, h_ptr, _1, _2)); sig.connect(boost::bind(&World::alternate_int_float, w_ptr, _1, _2)); float temp = 0.0f; sig(1, temp); > I don't see much benefit to extra (explicit) level of indirection here. > Most slots are short bind expressions that don't ever need to be > revisited, so it seems that it could be a bit challenging to determine > where some slots should reside: always in the class containing the > member function that is called (assuming one only uses member > functions!)? I guess I was thinking that since nothing is done with inheritance, it provides less opportunity to do things that might get you into trouble. If I've a hierarchy that looks something like this: BaseObject SimpleObject : BaseObject, signals::trackable ComplexObject : SimpleObject ReallyComplexObject : ComplexObject, signals::trackable where a SimpleObject can get a signal of one type and a ReallyComplexObject object can get another, is trackable inheritance ok with this or is there some other recommended method? Thanks again, Frank -----Original Message----- From: boost-users-bounces(a)lists.boost.org [mailto:boost-users-bounces@lists.boost.org] On Behalf Of Doug Gregor Sent: Tuesday, September 07, 2004 2:38 PM To: boost-users(a)lists.boost.org Subject: Re: [Boost-users] signal slot flexibility On Sep 7, 2004, at 1:36 PM, Frank Maddin wrote: > If I add a function to both Hello and World, from the "Hello World" > example, called alternate() and set up signals like so: > > > > boost::signal<void (Hello*) > sig_h; > > boost::signal< void (World*) > sig_w; > > > > sig_h.connect(boost::mem_fn(&Hello::alternate)); > > sig_w.connect(boost::mem_fn(&World::alternate)); > > > > I have to define two different signals since the object type is part > of the signal definition. Is there an easy way to get a single signal > to call the alternate() function of both Hello and World? Without Hello and World being related (e.g., sharing a base class declaring a virtual "alternate" function), I don't see a way. Still, when you call these signals you have to know whether you have a "Hello" in hand or a "World"; perhaps you wanted to store the "Hello" and "World" objects within the signal, and pass it no arguments? For instance, boost::signal<void()> sig; sig.connect(boost::bind(&Hello::alternate, &my_hello)); sig.connect(boost::bind(&World::alternate, &my_world)); > I recently found this signal slot solution. > http://www.codeproject.com/cpp/ElmueSignalsandSlots.asp > > There is an extra level of indirection here, in that you define a slot > that points to the function. I'd be interested to see what people on > this forum think about this solution. If I understand things > correctly, this solution doesn't seem to fit into the categories > listed in the "Choice of Slot Definitions" docs. I don't see much benefit to extra (explicit) level of indirection here. Most slots are short bind expressions that don't ever need to be revisited, so it seems that it could be a bit challenging to determine where some slots should reside: always in the class containing the member function that is called (assuming one only uses member functions!)? > Also a general signal slot question. What are the main reasons why > developers don't use an inherited pure interface for every callback > type? A signal would be a template class with a list of pointers to > the interface and some methods to operate on the list. Not flexible > enough? You often want only one event out of five, so it's a pain to drop in all of the stubs needed to ignore the other events. Also, you don't get a choice of names to use: if the writer of the interface says it's "on_click", you must write a function named "on_click" even though you'd rather write something like "submit_current_query" that describes the action taken when the event fires, not the event itself (which is essentially irrelevant). > Doesn't scale well to larger uses? Time spent calling virtual > functions? I saw this in the boost signal docs that might apply - > "the use of a large number of small adaptor classes containing virtual > functions has been found to cause an unacceptable increase in the size > of executables (polymorphic class types require more code than > non-polymorphic types)". This is more of an implementation detail; using function pointers instead of virtual functions tends to reduce the amount of code generated for libraries such as Boost.Signals or Boost.Function. Doug _______________________________________________ Boost-users mailing list Boost-users(a)lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users

1 0

signal slot flexibility
by Frank Maddin 07 Sep '04

07 Sep '04

Hi, I have some questions about signals. If I add a function to both Hello and World, from the "Hello World" example, called alternate() and set up signals like so: boost::signal<void (Hello*) > sig_h; boost::signal< void (World*) > sig_w; sig_h.connect(boost::mem_fn(&Hello::alternate)); sig_w.connect(boost::mem_fn(&World::alternate)); I have to define two different signals since the object type is part of the signal definition. Is there an easy way to get a single signal to call the alternate() function of both Hello and World? I recently found this signal slot solution. http://www.codeproject.com/cpp/ElmueSignalsandSlots.asp There is an extra level of indirection here, in that you define a slot that points to the function. I'd be interested to see what people on this forum think about this solution. If I understand things correctly, this solution doesn't seem to fit into the categories listed in the "Choice of Slot Definitions" docs. Also a general signal slot question. What are the main reasons why developers don't use an inherited pure interface for every callback type? A signal would be a template class with a list of pointers to the interface and some methods to operate on the list. Not flexible enough? Doesn't scale well to larger uses? Time spent calling virtual functions? I saw this in the boost signal docs that might apply - "the use of a large number of small adaptor classes containing virtual functions has been found to cause an unacceptable increase in the size of executables (polymorphic class types require more code than non-polymorphic types)". Thanks Frank

2 1

Re: [Boost-users] bjam with mingw
by Brewer, Christopher 07 Sep '04

07 Sep '04

Angus- I had similar problems with bjam under MinGW. By experimentation, I discovered that I could make bjam work under MinGW by moving it from it's previous location at 'c:/msys/1.0/bin' to 'c:/msys/1.0/mingw/bin'. FYI, you can find this thread starting at Boost-users Digest, Vol 250, Issue 1, and ending (successully :) at Boost-users Digest, Vol 253, Issue 1. HTH, Chris Brewer

3 2

bjam with mingw
by Angus Leeming 07 Sep '04

07 Sep '04

Please forgive the newbie questions, but I am trying and failing to build boost with MinGW. I've grabbed the CVS sources of boost from the sourceforge anonymous mirror. I've installed MinGW and MinSYS. The MinGW tools are in my path. From the MinSYS console: $ cd ~/boostcvs $ which which c++ /mingw/bin/c++ $ c++ --version c++.exe (GCC) 3.2.3 (mingw special 20030504-1) Just to check that the tools are working: $ cd libs/spirit/phoenix/test $ c++ -I$HOME/boostcvs -o new_test new_test.cpp $ ./new_test /////////////////////////////////////////////////////////////////////////////// Tests concluded SUCCESS!!! /////////////////////////////////////////////////////////////////////////////// Now trying to do something similar using bjam: $ cd ~/boostcvs/libs/test/build $ bjam '-sTOOLS=mgw' CALL "C:\Program Files\Microsoft Visual Studio .NET 2003\VC7\bin\VCVARS32.BAT" >nul "C:\Program Files\Microsoft Visual Studio .NET 2003\VC7\bin\cl" /Zm800 -nologo /EHsc -c -DBOOST_TEST_NO_AUTO_LINK=1 /Z7 /Od /Ob0 /EHsc /GR /MTd /Op /Zc:forScope /Zc:wchar_t -I"..\..\..\bin\boost\libs\test\build" -I"d:\MinSYS\1.0\home\Angus\boost_cvs" -Fo"..\..\..\bin\boost\libs\test\build\libboost_unit_test_framework.lib\vc7.1\debug\runtime-link-static\threading-multi\unit_test_main.obj" -Tp"d:\MinSYS\1.0\home\Angus\boost_cvs\libs\test\build\../src/unit_test_main.cpp" ...failed vc-C++ ..\..\..\bin\boost\libs\test\build\libboost_unit_test_framework.lib\vc7.1\debug\runtime-link-static\threading-multi\unit_test_main.obj... vc-C++ ..\..\..\bin\boost\libs\test\build\libboost_unit_test_framework.lib\vc7.1\debug\runtime-link-static\threading-multi\supplied_log_formatters.obj The system cannot find the path specified. The system cannot find the path specified. [ snip all other, similar messages ] For the life of me I can't see why bjam is trying to build the VC++ target. I can't even run $ bjam --help because it just does the same thing, trying and failing to build stuff with VC++. I have also tried running from the top level boost directory and with the '-sMINGW_ROOT_DIRECTORY=/d/mingw' command line option to bjam, all to no avail. (From my understanding of the docs at http://www.boost.org/tools/build/v1/mingw-tools.html, that shouldn't be necessary.) I have changed the TOOLS name to 'mingw' rather than 'mgw'. Again , no dice. Finally, in desperation, opening up a DOS prompt and repeating the exercise. Nope. Same result. Can anyone help me out? Regards, Angus

2 3

RE: [Boost-users] output stream cloning?
by Ben Hutchings 07 Sep '04

07 Sep '04

Jon Biggar <jon(a)levanta.com> wrote: > Is there any package in Boost (or upcoming in the Boost review queue) > that would let me create an output stream that clones its output to > one or more underlying output streams? Jonathan Turkanis's Iostream library, which is currently under review and has been mostly positively received, seems to make it pretty trivial to write such a thing - see <http://lists.boost.org/MailArchives/boost/msg69799.php> and <http://lists.boost.org/MailArchives/boost/msg69827.php>.

1 0

problem with CVS Boost.Python with MSVC
by Stephen D Evans 07 Sep '04

07 Sep '04

The CVS change to boost/boost/python/detail/prefix.hpp from revision 1.3 to 1.4 has obviated the effect of boost/boost/python/detail/wrap_python.hpp in _DEBUG when compiling with MSVC. Python now forces the debug library as if BOOST_DEBUG_PYTHON is permanently in effect. Revision 1.3 does not exhibit this problem. The added #include <pyconfig.h> is the cause. Stephen D Evans

1 1

dangerous use of boost::bind
by Gast 128 07 Sep '04

07 Sep '04

Dear all, In my work I use the boost libraries quite frequently, especially boost::bind to create function objects. However its use is not as transparent as I hoped for. Especially using functions in boost bind with signatures returning const& can give unpredictable results. For example: struct KX { KX (int i); //etc. }; typedef std::pair<int, KX> pair_t; pair_t do_kx_pair(int i) { return std::make_pair(i, KX(i)); } //following will go wrong: KX x = boost::bind(&pair_t::second, boost::bind(&do_kx_pair, i))(); The last line compiles on Visual .NET 2003 without any warning (even on level 4). However after spending two days in boost bind source code, I could see that it actually returns const& to a temporary created by boost bind. The reason for this is that arguments are copied by value, and thus functions with const & (or in this case pointer to member) can return then const & from temporaries. Is there no way that at least boost::bind could issue a compiler warning? Because now I have to think deep before I create a nested boost bind construction. For example the following construction is not wrong; the const & is again taken from a temporary which boost::bind holds, but this is destroyed after the expression: KX x = boost::bind(&pair_t::second, do_kx_pair(i))(); If this is not solvable, perhaps it's a good idea to include this case in the documentation. Wkr, me (btw boost::bind is a great improvement over the std::mem_fun and std::bind1st etc.)

2 1