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

Spirit
by Foster, Gareth 14 Oct '04

14 Oct '04

Hello all, I am trying to find some source code that uses spirit to parse C++ and convert it to HTML, i found some on google, but i misplaced it and i can't find it again. It looked like it could have been an example provided with the library, but i can't see it listed in the documentation. On a side note, anyone know the best way to use boost with cygwin? From source? Is there even another way? Thanks for any help! Gaz

3 2

boost::bind, MSVC7.1 and volatile-qualification
by Schalk_Cronje＠McAfee.com 14 Oct '04

14 Oct '04

Is there a known problem with boost::bind, volatile-qualified member functions and MSVC7.1? The following snippet of code happily compiles with various incantations of gcc, but fails with MSVC7.1. The same applies for 'const volatile'. #include <boost/function.hpp> #include <boost/bind.hpp> struct TEST { void gv0() volatile; void gvc0() volatile const; int gv2(int,int) volatile; int gvc2(int,int) volatile const; }; void instantiate() { TEST T; boost::function< void() > gv0 =boost::bind(&TEST::gv0,&T); // boost::function< int(int,int) > gv2 =boost::bind(&TEST::gv2,&T,_1,_2); // boost::function< void() > gcv0=boost::bind(&TEST::gvc0,&T); // boost::function< int(int,int) > gcv2=boost::bind(&TEST::gvc2,&T,_1,_2); } cl -c -TP -GS -GF -GR -vmv -vmg -EHsc -Zc:forScope -Zc:wchar_t -Op -Gm- -D_DEBUG -Od -Zi -Yd -MDd msvc71bind.cpp (Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 13.10.3077 for 80x86) ******\boost\mem_fn.hpp(256) : warning C4180: qualifier applied to function type has no meaning; ignored ******\boost\bind\bind_template.hpp(156) : see reference to class template instantiation 'boost::_mfi::dm<R,T>' being compiled with [ R=void (void) volatile, T=TEST ] msvc71bind.cpp(21) : see reference to class template instantiation 'boost::_bi::bind_t<R,F,L>' being compiled with [ R=void (__thiscall &)(void) volatile, F=boost::_mfi::dm<void (void) volatile,TEST>, L=boost::_bi::list_av_1<TEST *>::type ] ******\boost\mem_fn.hpp(264) : warning C4180: qualifier applied to function type has no meaning; ignored ******\boost\mem_fn.hpp(274) : warning C4180: qualifier applied to function type has no meaning; ignored ******\boost\mem_fn.hpp(288) : warning C4180: qualifier applied to function type has no meaning; ignored ******\boost\mem_fn.hpp(293) : warning C4180: qualifier applied to function type has no meaning; ignored ******\boost\mem_fn.hpp(307) : warning C4180: qualifier applied to function type has no meaning; ignored ******\boost\bind.hpp(1572) : warning C4180: qualifier applied to function type has no meaning; ignored msvc71bind.cpp(21) : see reference to function template instantiation 'boost::_bi::bind_t<R,F,L> boost::bind<void(void) volatile,TEST,TEST*>(void (__thiscall TEST::* ),A1)' being compiled with [ R=void (__thiscall &)(void) volatile, F=boost::_mfi::dm<void (void) volatile,TEST>, L=boost::_bi::list_av_1<TEST *>::type, A1=TEST * ] ******\boost\mem_fn.hpp(286) : error C4716: 'boost::_mfi::dm<void __thiscall TEST::(void)volatile ,TEST>::operator()' : must return a value

2 1

Re: Extending the boost::thread class
by arno.schaefer＠sqs.de 14 Oct '04

14 Oct '04

> For what it's worth, William Kempf began an implementation of thread > priorities and an exception-based thread cancellation scheme on the > thread_dev branch before he stopped working on Boost.Threads. If it's > completed enough (I haven't examined it closely enough yet to tell), > I hope to move it into the a Boost release after the upcoming 1.32. > That won't help you right now, however. I feel sorry about that, but perhaps it is possible to get a look to the actual state? Arno

2 1

Boost & VC2005 (Whidbey)
by S.I.M. van Keulen 14 Oct '04

14 Oct '04

Hi all, recently I started 'porting' some projects to 64 bit Windows (AMD) using Visual Studio 2005. I noticed that the Compiler had trouble with Boost. Searching the net didn't provide much helpfull information. I wondered if any of you had tried anything with Boost and VC2005/Whidbey compilers. Or if any of you could direct me to some information source on the net. Thank you! Serge van Keulen

3 3

Re: "smart list" container?
by David Jones 13 Oct '04

13 Oct '04

"Caleb Epstein" <caleb.epstein(a)gmail.com> wrote in message news:<989aceac041012063240b6aa3(a)mail.gmail.com>... > On Mon, 11 Oct 2004 10:19:54 -0600, David M. Jones <djones(a)keymark.com> wrote: > > - The container itself would act more or less identically to std::list; > > - The iterators on the container would act like boost::weak_ptr in that they > > would know that they are invalid when the list element that they point to is > > removed. > > Can't you just do this with a std::list<boost::shared_ptr>? You can > convert the *iterators to weak_ptrs as needed. Or am I missing > something? There are a couple of reasons that std::list<boost::shared_ptr> is less than ideal. First, using the algorithms becomes cumbersome/annoying. Instead of being able to write std::list<boost::shared_ptr<int> > lst; // ... Populate the list int i; // ... Set the value of i std::list<boost::shared_ptr<int> >::iterator = std::find(lst.begin(), lst.end(), i); I instead need to create a function object to represent a boost::shared_ptr<int> having the desired value and then use find_if(). Second, I would lose the ability to store iterators and then check them later for validity. Suppose that I have a std::list<boost::shared_ptr<int> > named lst and I have two std::list<boost::shared_ptr<int> >::iterator objects it1 and it2 that happen to reference the same node in the list. If I call lst.erase(it1) then it2 also becomes invalid. If I call boost::weak_ptr<int>(*it2) I get an error because I cannot dereference the iterator because it is already invalid. This is how std::list objects are supposed to work but what I want to be able to do is now call something like it2.expired() to determine that the node being refered to by it2 is not invalid/expired/removed. One way to get around this is to store boost::weak_ptr objects derived from std::list<boost::shared_ptr>::iterator objects instead of storing it1 and it2 directly. But then I lose the advantage of having stored iterators: I cannot (after checking for expiration and being at the end of the list) call operator++() to move the iterator forward. Fundamentally, I don't want to be storing a "list of nodes of shared pointers of objects" but a "list of nodes of objects whose pointers to each other are shared pointers". And I want the iterators to be weak pointers to the nodes. Potentially a way to do this would be to write a custom allocator my_alloc that declares typedef boost::shared_ptr pointer; and then create a std::list<T, my_alloc<T> > This works because std::list uses the allocator rebinding mechanism to call my_alloc<node<T> > rather than my_alloc<T>. (In fact, my_alloc<T> is never used to allocate anything in std::list<T>!) However, there is a problem because the interface (concept a la Austern) of boost::shared_ptr is different than that of a regular pointer. So when the std::list code tries to call _Myhead = 0; I get a compiler error. (I am using MSVC 7.1: although this particular error may be STL implementation dependent, I do believe that the problem I am describing will occur with any STL implementation.) But I digress... The question of why the boost::shared_ptr interface/concept does not conform to that of a normal pointer is probably a topic for a separate posting.

2 1

RE: [Boost-users] Extending the boost::thread class
by Foster, Gareth 13 Oct '04

13 Oct '04

Could you make it something like "enum { eLow, eNormal, eHight }; myThread.priority(eHigh);"? Maybe something non specific would be do-able, and it might still be useful to those who aren't after lots of control? Not that I am well clued up, just a thought. Cheers! Gaz -----Original Message----- From: boost-users-bounces(a)lists.boost.org [mailto:boost-users-bounces@lists.boost.org] On Behalf Of Ben Hutchings Sent: 13 October 2004 17:45 To: boost-users(a)lists.boost.org Subject: Re: [Boost-users] Extending the boost::thread class arno.schaefer(a)sqs.de wrote: > Hi all, > I want to use the boost thread library as a base for my multiplatform > thread class. After I have studied the documentation I recognize that > there are not all things available at the moment what I need, so I try > to find out what is the best way to start my implementation. What I > need, is to implement the following additional methods: suspend, resume, > stop (or terminate), set_priority, get_priority, get_state and I want to > have the possibility to create threads in a suspend mode. <snip> Why do you think you want suspend, resume and terminate? There are very few safe uses for these operations, and programs that use them tend to suffer from data corruption and/or deadlocks. It would be fairly easy to add support for priority settings to Boost.Thread, but since priority values are only meaningful in the context of a scheduling policy and policies vary between operating systems it would be hard to provide a genuinely portable priority system. _______________________________________________ Boost-users mailing list Boost-users(a)lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users

1 0

Extending the boost::thread class
by arno.schaefer＠sqs.de 13 Oct '04

13 Oct '04

Hi all, I want to use the boost thread library as a base for my multiplatform thread class. After I have studied the documentation I recognize that there are not all things available at the moment what I need, so I try to find out what is the best way to start my implementation. What I need, is to implement the following additional methods: suspend, resume, stop (or terminate), set_priority, get_priority, get_state and I want to have the possibility to create threads in a suspend mode. I want to use the boost thread as is and extend the needed features in my own class, but it seems so that their is at the moment no legal way to get the internal thread resource from a created thread, so that I can implement the missing and perhaps not portable features by my own? So what is the prefered way to extend the thread class? Is their a reasonable way to do it or should I build my own class from scratch? many thanks for any input and guess, regards Arno

3 2

about boost graph library -- subgraph
by qyang＠cs.ucr.edu 12 Oct '04

12 Oct '04

Hi, I am new here. I recently downloaded the boost library from www.boost.org. I need to use the "subgraph" function in the library. I tried to run the example file "subgraph.cpp", but the running result was different from what was given in the file. Below is my running result and sample output: //my result: G0: 0 --> 1 1 --> 2 3 2 --> 5 3 --> 4 --> 1 5 5 --> 3 0(0,1) 1(1,2) 2(1,3) 6(2,5) 3(4,1) 4(4,5) 5(5,3) G1: 0 --> 2 1 --> 2 2 --> 6(0,2) 4(1,2) G2: 0 --> 1 1 --> 0(0,1) //sample output G0: 0 --> 1 1 --> 2 3 2 --> 5 3 --> 4 --> 1 5 5 --> 3 0(0,1) 1(1,2) 2(1,3) 6(2,5) 3(4,1) 4(4,5) 5(5,3) G1: 2 --> 5 4 --> 5 5 --> 6(2,5) 4(4,5) G2: 0 --> 1 1 --> 0(0,1) //====================================================== Basically what happened was that the vertex indices of the subgraph G1 in my output are child's local indices, however, the vertex indices of the subgraph G1 in the sample output is parent's global indices. Can somebody explain this? Is it a bug in the boost code? Thanks very much, Qiaofeng

2 1

Thread library Win32 static linking for MSVC 2003 release?
by Emilio Schapira 12 Oct '04

12 Oct '04

Does anybody know when will be the next release of Boost that has the threads library fixed such that it is possible to link statically in win32 platforms compiling with MSVC++ 2003? I have read that it will be fixed for the next release. Also, how stable is the current version of the library in the CVS repository? I was browsing through it and it seems to have been fixed. Anybody has experiences with it? Thanks, Emilio.

2 2

"Compiler threading support not turned on" error
by Glen Simmons 12 Oct '04

12 Oct '04

I'm attempting to use Boost.threads on Mac OS X 10.3 and am getting the error "Compiler threading support is not turned on" when trying to compile. I'm using Xcode, which is Apple's IDE, so I'm not sure what to do when the error text helpfully suggests I set the correct command line option. As far as I know, there's nothing extra that you have to do to enable threading support in the compiler. I've written multi-threaded apps using pthreads and never had to deal with command line options. Any suggestions? Thanks, Glen Simmons

4 9

"smart list" container?
by David M. Jones 12 Oct '04

12 Oct '04

Hello, I have run into a situation where the following std::list-like container would be useful. - The container itself would act more or less identically to std::list; - The iterators on the container would act like boost::weak_ptr in that they would know that they are invalid when the list element that they point to is removed. Does anyone know of such a container in any of the boost libraries, in another library or a way to accomplish the task using other means (such as iterator adaptors)? I looked at the newly proposed "smart containers" library, but it seems to be solving a different problem despite the similarity with the name I am using to describe what I want. Thanks in advance, David M. Jones

2 1

two boost.threads taking much longer than one
by Andrew Chapman 11 Oct '04

11 Oct '04

Hi everyone, I've got an operation which takes 14 seconds on a single thread, but is taking up to 45 seconds when running via two boost.threads on a dual proc machine (2x 3ghz Xeons, Fedora Core 1). The thread functor objects are being passed a lot of references (about 20) to some very large std::vectors. One thread always reads/writes from the start of the vectors to their midpoint, and the second thread from (midpoint+1) to the end of the vectors. The only thing I can think of is that the threads are taking a long time to fork and get running, perhaps moving a lot of stack data around or something? When I look at the output of 'top' with threads displayed, I see the two threads appear, but they only use about 50-60% CPU utilization each. The system monitor shows mostly red (Kernel) as opposed to blue (User). Any ideas? -- Andrew Chapman Senior Technical Director - Framestore CFC

3 3

Re: Re: [Boost-users] function_traits inheritance
by Victor Morilla 11 Oct '04

11 Oct '04

Peter Dimov wrote: >Victor Morilla wrote: >> >> template <typename FnType, FnType* fn> >> struct Translator<FnType, fn, 0> : >> function_traits<FnType> >> { >> typedef result_type type; > typedef typename function_traits<FnType>::result_type type; >> }; The question is, why result_type type definition is not inherited from function_traits ? ______________________________________________ Renovamos el Correo Yahoo!: ¡100 MB GRATIS! Nuevos servicios, más seguridad http://correo.yahoo.es

2 1

function_traits inheritance
by Victor Morilla 11 Oct '04

11 Oct '04

Guess someone why this code fails ? (g++3.3) ---- begin ----------------------- #include <boost/type_traits.hpp> #include <iostream> using namespace std; using namespace boost; int foo() { return 10; } template <typename FnType, FnType* fn, int Arity = function_traits<FnType>::arity> struct Translator; template <typename FnType, FnType* fn> struct Translator<FnType, fn, 0> : function_traits<FnType> { typedef result_type type; }; int main(int argc, char** argv) { Translator<int (), &foo>::type x; } ------------------- end code ------------------ error: ISO C++ forbids declaration of `result_type' with no type --------------------- end error ------------------- And another question: is there a portable equivalent to: Translator<typeof(foo), &foo> Thanks ______________________________________________ Renovamos el Correo Yahoo!: ¡100 MB GRATIS! Nuevos servicios, más seguridad http://correo.yahoo.es

3 2

Problems in compilation with Boost
by Debayan Chaudhuri 11 Oct '04

11 Oct '04

Hi, I am evaluating Boost for use in our project. The platform OS is FreeBSD 4.9; which comes with gcc-2.95.3. I downloaded boost_1.30.0 and compiled the thread library on the FreeBSD box. When I try to compile a small test program using Boost threads; I encounter hundreds of compiler errors; mostly in the Boost header files. Is there a workaround to this ? Which version of Boost works with gcc-2.95.3 and what are the compiler options that will make it compile properly? regards, Debayan.

2 2

RE: [Boost-users] Problems with compiling Filesystem library on Solaris
by Barak Simon 11 Oct '04

11 Oct '04

By now I had time to check your suggestion, but I'm afraid I still get the same error when building boost libs: "/db/pub/eq/build/src/boost_1_31_0/boost/iterator/iterator_facade.hpp", line 622: Error: Could not find a match for boost::iterator_core_access::equal<boost::iterator_core_access::Facade1, boost::iterator_core_access::Facade2>(const boost::filesystem::directory_iterator, const boost::filesystem::directory_iterator, boost::is_convertible<boost::filesystem::directory_iterator, boost::filesystem::directory_iterator>). "/db/pub/eq/build/src/boost_1_31_0/libs/filesystem/build/../src/operations_posix_windows.cpp", line 176: Where: While instantiating "boost::operator==<boost::filesystem::directory_iterator, boost::filesystem::path, boost::single_pass_traversal_tag, boost::filesystem::path&, int, boost::filesystem::directory_iterator, boost::filesystem::path, boost::single_pass_traversal_tag, boost::filesystem::path&, int>(const boost::iterator_facade<boost::filesystem::directory_iterator, boost::filesystem::path, boost::single_pass_traversal_tag, boost::filesystem::path&, int>&, const boost::iterator_facade<boost::filesystem::directory_iterator, boost::filesystem::path, boost::single_pass_traversal_tag, boost::filesystem::path&, int>&)". "/db/pub/eq/build/src/boost_1_31_0/libs/filesystem/build/../src/operations_posix_windows.cpp", line 176: Where: Instantiated from non-template code. Are you sure that is the only line I need to comment out? Thanks, Barak ________________________ Barak Simon GED IT Core Platform 190 George St. Sydney, 2000 Phone +61 2 925 85070 Fax +61 2 925 95050 barak.simon+external@db. com To: boost-users(a)lists.boost.org Sent by: cc: boost-users-bounces@list Subject: RE: [Boost-users] Problems with compiling Filesystem library on Solaris s.boost.org 11/10/2004 11:45 Please respond to boost-users Thanks for the swift reply! In the interim I commented out the 2-3 problematic methods (which I am not using anyway) in the filesystem lib, but I like your solution better. A question on that though: On your solution, what exactly am I commenting out, that is, can the workaround break some other boost library or a significant assumption? Thanks, Barak Tomas.Puverle@morganstan ley.com To: boost-users(a)lists.boost.org Sent by: cc: boost-users-bounces@list Subject: RE: [Boost-users] Problems with compiling Filesystem library on Solaris s.boost.org 08/10/2004 18:47 Please respond to boost-users > Hi, > I get the following error from the Filesystem library when > building boost 1_31 on Solaris with sunpro compiler: > > "/db/pub/eq/build/src/boost_1_31_0/boost/iterator/iterator_fac > ade.hpp", line 622: Error: Could not find a match for > boost::iterator_core_access::equal<boost::iterator_core_access > ::Facade1, boost::iterator_core_access::Facade2>(const > boost::filesystem::directory_iterator, const > boost::filesystem::directory_iterator, > boost::is_convertible<boost::filesystem::directory_iterator, > boost::filesystem::directory_iterator>). > > "/db/pub/eq/build/src/boost_1_31_0/libs/filesystem/build/../sr c/operations_posix_windows.cpp", line 176: Where: While instantiating > "boost::operator==<boost::filesystem::directory_iterator, > boost::filesystem::path, boost::single_pass_traversal_tag, > boost::filesystem::path&, int, > boost::filesystem::directory_iterator, > boost::filesystem::path, boost::single_pass_traversal_tag, > boost::filesystem::path&, int>(const > boost::iterator_facade<boost::filesystem::directory_iterator, > boost::filesystem::path, boost::single_pass_traversal_tag, > boost::filesystem::path&, int>&, const > boost::iterator_facade<boost::filesystem::directory_iterator, > boost::filesystem::path, boost::single_pass_traversal_tag, > boost::filesystem::path&, int>&)". > > "/db/pub/eq/build/src/boost_1_31_0/libs/filesystem/build/../sr c/operations_posix_windows.cpp", line 176: Where: Instantiated from non-template code. > > > > It looks suspiciously similar to bug "[ 976241 ] Compiler > error for tokenizer on Solaris" from 2004-06-20, with the > difference that I am using version 5.3 of sunpro rather than 5.5. I have submitted a similar problem and the proposed fix recently. I am just waiting on our developers to confirm that it works before submitting it to boost, but it seems to fix the tokenizer problem, too. For a quick workaround (before the patch is applied) do this: At the top of boost/type_traits/detail/bool_trait_def.hpp you will see something like this: #if defined(__SUNPRO_CC) # define BOOST_TT_AUX_BOOL_TRAIT_VALUE_DECL(C) \ typedef mpl::bool_< C > type; \ enum { value = type::value }; \ /**/ # define BOOST_TT_AUX_BOOL_C_BASE(C) //1 Comment out the line marked with //1. That seems to have fixed my problems with both 5.3 and 5.5. HTH, Tom -------------------------------------------------------- NOTICE: If received in error, please destroy and notify sender. Sender does not waive confidentiality or privilege, and use is prohibited. _______________________________________________ Boost-users mailing list Boost-users(a)lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users -- This e-mail may contain confidential and/or privileged information. If you are not the intended recipient (or have received this e-mail in error) please notify the sender immediately and destroy this e-mail. Any unauthorized copying, disclosure or distribution of the material in this e-mail is strictly forbidden. _______________________________________________ Boost-users mailing list Boost-users(a)lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users -- This e-mail may contain confidential and/or privileged information. If you are not the intended recipient (or have received this e-mail in error) please notify the sender immediately and destroy this e-mail. Any unauthorized copying, disclosure or distribution of the material in this e-mail is strictly forbidden.

1 0

Function type transformation and manipulation traits?
by Aaron W. LaFramboise 11 Oct '04

11 Oct '04

Are there any Boost support for manipulating function types? Operations I have in mind include: setting return type, setting argument type, changing number of arguments, et cetera. If not, is there some generally accepted way to accomplish this, and is any work being done to add this feature to Boost? Aaron W. LaFramboise

2 3

RE: [Boost-users] Problems with compiling Filesystem library on Solaris
by Barak Simon 11 Oct '04

11 Oct '04

Thanks for the swift reply! In the interim I commented out the 2-3 problematic methods (which I am not using anyway) in the filesystem lib, but I like your solution better. A question on that though: On your solution, what exactly am I commenting out, that is, can the workaround break some other boost library or a significant assumption? Thanks, Barak Tomas.Puverle@morganstan ley.com To: boost-users(a)lists.boost.org Sent by: cc: boost-users-bounces@list Subject: RE: [Boost-users] Problems with compiling Filesystem library on Solaris s.boost.org 08/10/2004 18:47 Please respond to boost-users > Hi, > I get the following error from the Filesystem library when > building boost 1_31 on Solaris with sunpro compiler: > > "/db/pub/eq/build/src/boost_1_31_0/boost/iterator/iterator_fac > ade.hpp", line 622: Error: Could not find a match for > boost::iterator_core_access::equal<boost::iterator_core_access > ::Facade1, boost::iterator_core_access::Facade2>(const > boost::filesystem::directory_iterator, const > boost::filesystem::directory_iterator, > boost::is_convertible<boost::filesystem::directory_iterator, > boost::filesystem::directory_iterator>). > > "/db/pub/eq/build/src/boost_1_31_0/libs/filesystem/build/../sr c/operations_posix_windows.cpp", line 176: Where: While instantiating > "boost::operator==<boost::filesystem::directory_iterator, > boost::filesystem::path, boost::single_pass_traversal_tag, > boost::filesystem::path&, int, > boost::filesystem::directory_iterator, > boost::filesystem::path, boost::single_pass_traversal_tag, > boost::filesystem::path&, int>(const > boost::iterator_facade<boost::filesystem::directory_iterator, > boost::filesystem::path, boost::single_pass_traversal_tag, > boost::filesystem::path&, int>&, const > boost::iterator_facade<boost::filesystem::directory_iterator, > boost::filesystem::path, boost::single_pass_traversal_tag, > boost::filesystem::path&, int>&)". > > "/db/pub/eq/build/src/boost_1_31_0/libs/filesystem/build/../sr c/operations_posix_windows.cpp", line 176: Where: Instantiated from non-template code. > > > > It looks suspiciously similar to bug "[ 976241 ] Compiler > error for tokenizer on Solaris" from 2004-06-20, with the > difference that I am using version 5.3 of sunpro rather than 5.5. I have submitted a similar problem and the proposed fix recently. I am just waiting on our developers to confirm that it works before submitting it to boost, but it seems to fix the tokenizer problem, too. For a quick workaround (before the patch is applied) do this: At the top of boost/type_traits/detail/bool_trait_def.hpp you will see something like this: #if defined(__SUNPRO_CC) # define BOOST_TT_AUX_BOOL_TRAIT_VALUE_DECL(C) \ typedef mpl::bool_< C > type; \ enum { value = type::value }; \ /**/ # define BOOST_TT_AUX_BOOL_C_BASE(C) //1 Comment out the line marked with //1. That seems to have fixed my problems with both 5.3 and 5.5. HTH, Tom -------------------------------------------------------- NOTICE: If received in error, please destroy and notify sender. Sender does not waive confidentiality or privilege, and use is prohibited. _______________________________________________ Boost-users mailing list Boost-users(a)lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users -- This e-mail may contain confidential and/or privileged information. If you are not the intended recipient (or have received this e-mail in error) please notify the sender immediately and destroy this e-mail. Any unauthorized copying, disclosure or distribution of the material in this e-mail is strictly forbidden.

1 0

Help ... BOOST::thread 's certain constuctor methods fails with STLPort
by Raj 09 Oct '04

09 Oct '04

Hi I have build boost libraries for a) regular gcc b) gcc-STLPort the following code works perfectly for "gcc" and throws up the following errors for "gcc-STLPort". // this is my function object that // takes in input parameters as part of its constructor. func_obj instance(data,m_regMap); boost::shared_ptr<thread> pThread(new boost::thread(instance)); ERROR: mystuff.cpp:101: undefined reference to `boost::thread::thread[in-charge] (boost::function0<void, _STL::allocator<boost::funct ion_base> > const&)' If I assume that the boost runtime library has not been linked, that is not the case.(ldd tells me that i have) whereever I use 'xtime','lock' or 'condition' it doesn't complain. WHAT IS STRANGE IS THE FOLLOWING COMPILES boost::thread thrd; boost::thread thrd(func_obj); // it is the actual fo not an instance. Any help would be greatly appreciated. This happens on Linux/with STLPort 4.6.2/gcc 3.2.3

1 0

How to compile BOOST on Linux with NPTL threads and STLPort
by Raj 09 Oct '04

09 Oct '04

Hi, I am trying to use BOOST thread and regex libraries on Linux. I am able to compile them with sTOOLSS=gcc-stlport option. But, I would also like to specify that I would like to use the NPTL thread lib instead of pthread. How can I do it ? Any help would be greatly appreciated. Thanks in Advance Raja

2 3

building graph library
by kyle huang 09 Oct '04

09 Oct '04

i just started looking at the boost libraries and thus i'm a total beginner. i'm trying to build the graph library in particular, and i haven't had any luck at all. first, is it a buildable library, or is it just a collection of templates? if so (collection of templates), how do i use it? i'm using windows xp, and building for Visual Studio 6 thanks!

3 2

BLL: Confusing sig example code in documentation
by Markus Larsson 09 Oct '04

09 Oct '04

from the BLL/bind documentaion on the cvs: struct A { // the return type equals the third argument type: template<class T1, T2, T3> T3 operator()(const T1& t1, const T2& t2, const T3& t3); template <class Args> class sig { // get the third argument type (4th element) typedef typename boost::tuples::element<3, Args>::type T3; public: typedef typename boost::remove_cv<T3>::type type; } }; To begin with this piece of code has 2 syntax errors. template<class T1, T2, T3> // should have class before all parameters struct sig { ... } // missing semicolon If a first time user copy/pastes the example, fixes the syntax errors, defines operator() and tries to call the function with: int i = 7; bind( A(), 1, 2, _1) (i); she gets a hard to parse error because operator() wasn't declared const. I see two possible solutions: 1) Add const to the declaration 2) mention the non constness and show how to call the function object using var(func), with a short note on why that is needed.

1 0

shared_ptr interactions
by eric lindvall 08 Oct '04

08 Oct '04

I have created a simple template to make instantiation of shared_ptrs a little prettier. Here is an abridged version of the template: template <class T> class ManagedPointer : public boost::enable_shared_from_this<T> { public: typedef boost::shared_ptr<T> shared_ptr; public: static shared_ptr create () { return (shared_ptr (new T ())); } template <class A1> static shared_ptr create (A1 a1) { return (shared_ptr (new T (a1))); } }; This allows for me to create a class: class TestClass : public ManagedPointer<TestClass> { public: TestClass(); }; and create an instance by: TestClass::shared_ptr test = TestClass::create(); I have noticed in certain cases I get very strange behaviour. Most recently I've seen things hanging in the shared_ptr destructor. Changing the code to: TestClass::shared_ptr test (new TestClass()); does "solve" the problem. Is there something subtle in how the templates and subclassing are interacting that I am not understanding? thanks. e.

2 2

boost implementation
by Gast 128 08 Oct '04

08 Oct '04

Dear all, Browsing through source files of boost, I see that some libraries reimplement common idioms found in other libraries. For example the tuple library reimplements its own compile time if and it's own typelist. I guess it would be better if high level boost libraries(e.g. regex, threads, to some extend also the tuples) would use the low level boost libraries (e.g. preprocessor, type_traits, mpl). Besides bugfixes and platform/compiler workarounds, it has also a documentary advantage. If you are familiar with mpl for example, one knows how to interpret the construct with if_. Also one can better interact with libraries if they are using the same base constructs. There almost no disadvantages, because all the low level boost libraries are template based. But probably this is a matter of time, because at first look at the variant library, I see that it uses the mpl. And people are talking about the fusion library for the tuples. Are there boost developers who can comment on this? Wkr, me.

2 1

RE: [Boost-users] Problems with compiling Filesystem library on Solaris
by Puverle, Tomas (IT) 08 Oct '04

08 Oct '04

> Hi, > I get the following error from the Filesystem library when > building boost 1_31 on Solaris with sunpro compiler: > > "/db/pub/eq/build/src/boost_1_31_0/boost/iterator/iterator_fac > ade.hpp", line 622: Error: Could not find a match for > boost::iterator_core_access::equal<boost::iterator_core_access > ::Facade1, boost::iterator_core_access::Facade2>(const > boost::filesystem::directory_iterator, const > boost::filesystem::directory_iterator, > boost::is_convertible<boost::filesystem::directory_iterator, > boost::filesystem::directory_iterator>). > > "/db/pub/eq/build/src/boost_1_31_0/libs/filesystem/build/../sr c/operations_posix_windows.cpp", line 176: Where: While instantiating > "boost::operator==<boost::filesystem::directory_iterator, > boost::filesystem::path, boost::single_pass_traversal_tag, > boost::filesystem::path&, int, > boost::filesystem::directory_iterator, > boost::filesystem::path, boost::single_pass_traversal_tag, > boost::filesystem::path&, int>(const > boost::iterator_facade<boost::filesystem::directory_iterator, > boost::filesystem::path, boost::single_pass_traversal_tag, > boost::filesystem::path&, int>&, const > boost::iterator_facade<boost::filesystem::directory_iterator, > boost::filesystem::path, boost::single_pass_traversal_tag, > boost::filesystem::path&, int>&)". > > "/db/pub/eq/build/src/boost_1_31_0/libs/filesystem/build/../sr c/operations_posix_windows.cpp", line 176: Where: Instantiated from non-template code. > > > > It looks suspiciously similar to bug "[ 976241 ] Compiler > error for tokenizer on Solaris" from 2004-06-20, with the > difference that I am using version 5.3 of sunpro rather than 5.5. I have submitted a similar problem and the proposed fix recently. I am just waiting on our developers to confirm that it works before submitting it to boost, but it seems to fix the tokenizer problem, too. For a quick workaround (before the patch is applied) do this: At the top of boost/type_traits/detail/bool_trait_def.hpp you will see something like this: #if defined(__SUNPRO_CC) # define BOOST_TT_AUX_BOOL_TRAIT_VALUE_DECL(C) \ typedef mpl::bool_< C > type; \ enum { value = type::value }; \ /**/ # define BOOST_TT_AUX_BOOL_C_BASE(C) //1 Comment out the line marked with //1. That seems to have fixed my problems with both 5.3 and 5.5. HTH, Tom -------------------------------------------------------- NOTICE: If received in error, please destroy and notify sender. Sender does not waive confidentiality or privilege, and use is prohibited.

1 0