February 2005 - Boost-users - lists.preview.boost.org

RE: [Boost-users] Custom containers
by Johnson, Donald W MVK Contractor 08 Feb '05

08 Feb '05

In your example size() returns 4 capacity() returns 6 end() returns a pointer to the 5th element When in doubt about standard container design examing std::vector is a good place to start. Good Luck -----Original Message----- From: Foster, Gareth [mailto:gareth.foster@siemens.com] Sent: Tuesday, February 08, 2005 8:36 AM To: 'boost-users(a)lists.boost.org' Subject: [Boost-users] Custom containers Hello all, If you were to write a container that had a certain size, say 6, but may not always contain 6 items, say it contains 4, what would it be best to have the following methods do? size() - returns 6 or 4? end() - returns an iterator on the 5th or 7th element? I'm not sure if the standard way of writing such things dictates a certain approach, or if it is purely down to the intended use. Thanks for any information. Gaz ******************************* * http://planetearthworm.com/ * *******************************

1 0

boost::signals block/unblock
by frantz.maerten＠igeoss.com 08 Feb '05

08 Feb '05

Hi all, Maybe a stupid question: I wonder there's a way to connect a slot to signal (which gives a connection), temporarily disconnect this connection, and reconnect it later (like a block function) My problem is that I need to emit a signal to the connected slots given a condition. If a connection satisfies the precondition, it will be notified, otherwise it is temporarily disconnected. Thanks for any help

1 0

msvc vc 6.0 SP6 internal heap limit reached
by Ira 08 Feb '05

08 Feb '05

All, I am a newbie who is trying out the serialization library. In that article, there is a demo program (demo.cpp) that i am trying to build. Along with several warnings ('dynamic_cast' used on polymorphic type...), the compiler is generating the following error: fatal error C1076: compiler limit : internal heap limit reached; use /Zm to specify a higher limit I have tried various values for /Zm (200 is recomended) and either i still receive this error, or if increase the value too much, the compiler generates the following error: fatal error C1204: compiler limit : internal structure overflow Since this seesaw response doesn't seem correct, i am wondering if there is a particular flag or combination thereof that would correct this problem? Thank you.

6 5

Custom containers
by Foster, Gareth 08 Feb '05

08 Feb '05

Hello all, If you were to write a container that had a certain size, say 6, but may not always contain 6 items, say it contains 4, what would it be best to have the following methods do? size() - returns 6 or 4? end() - returns an iterator on the 5th or 7th element? I'm not sure if the standard way of writing such things dictates a certain approach, or if it is purely down to the intended use. Thanks for any information. Gaz ******************************* * http://planetearthworm.com/ * *******************************

2 1

Boost libraries and no exceptions
by David Ward 08 Feb '05

08 Feb '05

Boosters: In our embedded environment, we compile everything with -fno-exceptions. I'm getting errors in the new program_options library because it has naked throw calls in it. I did see a definition of boost::throw_exception and changed the naked throw calls to use that. That seems to work. However, naked throw calls are littered throughout many of the libraries. Is there some generic way to get these libraries to compile with -fno-exceptions or do I have to go through and change all calls to throw to boost::throw_exception? Thanks, David Ward Acterna

3 2

Bind question
by Surya Kiran Gullapalli 07 Feb '05

07 Feb '05

Hi all, This is a follow up to my previous message on boost::bind question. Please go thru the attached program. I've 2 structures X an Y, and i'm calling stable_sort on vectors of both the objects. But in one stable sort, i'm using the member variable directly, and in the other i'm using a member function to access that. When i use boost::bind with member variable directly, it has no problem (structure Y), But if i use member function to access the member variable, which is private, i'm getting compilation errors, The essence of the compilation error is given below.. ===================== ./boost/include/boost/lambda/detail/actions.hpp:86: error: no matching function for call to `boost::lambda::function_adaptor<int (X::*)()>::apply(int (X::*const&)(), const X&)' ./boost/include/boost/lambda/detail/function_adaptors.hpp:193: note: candidates are: static Result boost::lambda::function_adaptor<Result (Object::*)()>::apply(Result (Object::*)(), Object*) [with RET = int, Object = X, Result = int] ./boost/include/boost/lambda/detail/function_adaptors.hpp:197: note: static Result boost::lambda::function_adaptor<Result (Object::*)()>::apply(Result (Object::*)(), Object&) [with RET = int, Object = X, Result = int] ===================== what am i doing wrong. thanks in advance, Surya

2 1

Bind question
by Surya Kiran Gullapalli 07 Feb '05

07 Feb '05

Hi all, I'm having this particular class class File { public: int get_size(); std::string get_name(); private: int size, std::string name; } ; and a vector of objects of this class. vector <File> files ; and if want to sort the elements of the container i use stable_sort (files.begin(), files.end()) // Invokes operator < if its defined. otherwise shows up an error. there's another form of stable_sort, in which we can call a sorting function of our own. Now my problem is the sorting criteria i use is dependent on runtime. say at times i want to sort by size and at other times i want to sort them by name of the file. The simple solution could be write 2 functions, sort_on_size and sort_on_filename which does the work. Now, the sorting order is also decided at runtime. So i write 2 more functions for that. The problem is the number of functions i use increase like any thing, which does not look good. The best way should be to use bind (or am i wrong here ?) from Boost library. But i'm not getting the syntax correctly here. the crude way could be stable_sort (files.begin(), files.end(), (bind (&File::get_size, _1) > bind (&File::get_size, _2))) ; that is call get_size() on first argument, and call get_size() on second argument, and compare those values, and return a boolean value. Can anybody please help me out with the syntax here. ?? Thanks in advance, Surya

3 3

[signals] Painfully simple signals question
by Tony Juricic 07 Feb '05

07 Feb '05

Shame be where it belongs, after rading documentation, implementation, examples and test code, I am unable to write what should be simple signals code: --------------------------------------------- *.h typedef boost::function<void (void)> Callback; struct Signal { static void Connect(Callback f); }; ------------------------------------------------ *.cpp boost::signal<void (void)> mySignal; void Signal::Connect(Callback f) { // iterate over signal slots, if slot equals input f do nothing // otherwise connect f } Tony

2 2

Signals and fuction documentation really old?
by Tony Juricic 07 Feb '05

07 Feb '05

..or am I having a day of profound stupidity. documentation: Boost.Function 3.5 Comparing Boost.Function objects me: okay, boost function object wrapper objects are comparable. documentation: Chapter 5 FAQ 5.1 Comparison between boost::function objects cannot be implemented well, and therefore will not be implemented. me:okay so it is not implemented documentation: 5.1 This essentially negates the usefulness of operator==in the context in which it is most desired:multitarget callbacks. me: great, multitarget callbacks is exactly why I am reading all this documentation documentation: 5.1 The Signals library has a way around this. me: great, if I only had an example of that way that wouldn't contain 2 free functions declared in the same file as the signal itself. Any help in clearing up my confusion greatly appreciated Tony

2 1

deepcpy_ptr
by gast128 06 Feb '05

06 Feb '05

Dear all, I couldn't find functionality a la deep copy pointer in boost. Isn't this an idea to extend the smart_ptr library with it? Basically it has the same functionality as if the object was hold by value, but sometimes you want to dismiss the header file and only have a fwd decl: #include "bla.hpp" struct Foo { Bla m_Bla; }; deep copy: class Bla; //fwd decl. is sufficient { deepcpy_ptr<Bla> m_ptr; }; Or do I overlook something? Wkr, me

2 2

Re: Compiling boost on tru64 (Richard Hadsell)
by Mark Helzer 06 Feb '05

06 Feb '05

>Mark Helzer wrote: >>I'm getting this error trying to build boost >>tru64cxx65-C++-action >>bin/boost/libs/date_time/build/libboost_date_time.so/tru64 >>cxx65/debug/greg_month.o >>cxx: Error: /adp/local/src/boost_1_32_0/boost/cstdint.hpp, line 87: #5 >>could >> not open source file "inttypes.h" >> >>Quite true, it doesn't exist. >>Here's my command line: >>bjam -sTOOLS=tru64cxx65 install >> >>Compiler is >> >> >>cxx -V >> >> >>Compaq C++ V6.5-041 for Digital UNIX V4.0F (Rev. 1229) >>Compiler Driver V6.5-041 (cxx) cxx Driver >> >>Is this a configuration issue? >> >> >For now, remove '|| defined(__osf__)' from line 85 of cstdint.hpp. >There may be a more general fix in the current CVS version, but this >change works for me (and my 4.0F system). Kewl, that works, I got a little further down the path. Next problem: cxx: Error: /adp/local/src/boost_1_32_0/boost/bind.hpp, line 206: #266 "value" is ambiguous detected during: instantiation of class "boost::_bi::list1<A1> [with A1=boost::_bi::list_av_1<boost::arg<1>>::B1]" at line 161 of "/adp/local/src/boost_1_32_0/boost/bind/bind_template.hpp " instantiation of class "boost::_bi::bind_t<R, F, L> [with R=std::wstring, F=std::wstring (*)(const std::string &), L=boost::_bi::list1<boost::_bi::list_av_1<boost::arg<1>>: :B1>]" at line 71 of "/adp/local/src/boost_1_32_0/libs/program_options/build/. ./src/parsers.cpp" template<class T> T & operator[] (value<T> & v) const { return v.get(); } --------------------------------------^ I see value defined as a class template at line 90 of boost/bind.hpp

3 3

linker problems with gcc 3.3.5 and 3.4.3
by Derek Scherger 06 Feb '05

06 Feb '05

Hi, I've recently started getting thousands of messages like the following, while linking with -lboost_regex -lboost_date_time -lboost_filesystem. /usr/lib/gcc-lib/i686-pc-linux-gnu/3.3.5/../../../../i686-pc-linux-gnu/bin/ld: `.gnu.linkonce.t._ZN5boost2io6detail22parse_printf_directiveIcSt11char_traitsIcEN9__gnu_cxx17__normal_iteratorIPKcSsEENS0_16basic_outsstreamIcS4_EEEEbRT1_RKSC_PNS1_11format_itemIT_T0_EERT2_h' referenced in section `.rodata' of monotone-commands.o: defined in discarded section `.gnu.linkonce.t._ZN5boost2io6detail22parse_printf_directiveIcSt11char_traitsIcEN9__gnu_cxx17__normal_iteratorIPKcSsEENS0_16basic_outsstreamIcS4_EEEEbRT1_RKSC_PNS1_11format_itemIT_T0_EERT2_h' of monotone-commands.o Linking monotone (http://www.venge.net/monotone) produces ~5000 or so messages like this. I'm running on gentoo linux and I believe this started happening when I recently upgraded from gcc 3.3.4 to 3.3.5. Originally I was using boost 1.31.0 but I have tried upgrading to 1.32.0 and also upgrading gcc to 3.4.3. I have not yet tried downgrading gcc to 3.3.4 but that's pretty much next on the list. I really don't know for sure what is causing this, but it really is quite annoying. It could be gentoo's packaging of gcc or something else, or it could be gcc or boost, or ... I do see a few similar sounding posts such as http://www.mail-archive.com/lyx-devel@lists.lyx.org/msg74258.html but so far haven't been able to fix the problem. Any help or insight would be greatly appreciated. Cheers, Derek

2 2

Linker Problem with Regex Libraries VC7
by Nicholas Cardi 06 Feb '05

06 Feb '05

The linker is complaining that it can't find the file libboost_regex-vc71-mt-sgd-1_32.lib This file was not built when I made the regex library. Here is the exsact linker error: fatal error LNK1104: cannot open file 'libboost_regex-vc71-mt-sgd-1_32.lib' The file libboost_regex-vc7-mt-sgd-1_32.lib does exist on my machine and it is in my include directory. >From the install document Finally when you use regex++ it is only necessary for you to add the <boost> root directory to your list of include directories for that project. It is not necessary for you to manually add a .lib file to the project; the headers will automatically select the correct .lib file for your build mode and tell the linker to include it. Is it possible the headers are referancing the wrong lib file? Thanks Nicholas Cardi

1 1

(no subject)
by M L 05 Feb '05

05 Feb '05

unsub boost-users

1 0

[variant] is_a/can_be
by Yuval Ronen 05 Feb '05

05 Feb '05

Hi. I encountered the need to query a boost::variant to which type it currently contain. I'm talking about something like: variant<int, string> v(3); assert(v.is_a<int>()); assert(!v.is_a<string>()); This can be done using the which() method, but its drawback is that I rely on the order of the types in the variant declaration rather on the type, which is really what I'm interested in. I can implement the variant_is_a<>() feature using a visitor such as: template <typename T> struct is_a_visitor : public boost::static_visitor<bool> { template <typename U> bool operator()(const U &) const { return false; } bool operator()(const T &) const { return true; } }; template <typename T, class Variant> bool variant_is_a(const Variant &v) { return v.apply_visitor(is_a_visitor<T>()); } but this way has a problem: calling variant_is_a<double>() on a variant of type variant<int, string> (no double for this variant) will simply return false, rather than just not compile, which is what I prefer. Making this not compile will require a new feature from variant: a boolean constant 'can_be'. With this, the variant_is_a() function could be written with additional line: BOOST_STATIC_ASSERT(Variant::can_be<T>::value); IMO, the is_a/can_be features are general and useful enough to be included in the library. Wouldn't you agree? Yuval

4 7

RE: [Boost-users] Re: Bug in boost::lambda using operator <<?
by Richard_Cox＠Dell.com 05 Feb '05

05 Feb '05

On Friday, February 04, 2005 5:47 PM Kevin Wheatley wrote: > > Richard_Cox(a)Dell.com wrote: > > > > The program below outputs > > > > String: <First string> Second string> Fourth string> Fifth string> > > > > All but the first "<" is missing. Any ideas? > > std::for_each(coll.begin(), coll.end(), out << gt << _1 > << L"> "); > > > does this work? > > std::for_each(coll.begin(), coll.end(), out << var(gt) << _1 << L"> > "); > > not sure of the fix, but out << gt associates left to right and is > evaluated once and has the type of out, you want it lazy evaluated so > that the whole of the expression is considered as one... see the > delaying constants and variables section of the lambda documents. > > Kevin Of course, "obvious" now :-( out << gt would evaluate returning the wostream which the lambda expression created from the 2nd and 3rd <<'s then operates on. Many thanks. Richard All opinions are mine and do not represent in any way the position of Dell Corporation.

1 0

implementing an iterator
by Chris Goller 04 Feb '05

04 Feb '05

I'm using iterator_facade to implement an iterator over file based data. So, my question is, since deference returns a reference to a value does that reference need to persist when the iterator is incremented or decremented? Because this is file based data I can't always keep a reference around because the data is stored on disk. Thanks, Chris

2 1

boost::function as drop-in replacment for ordinary function pointers?
by Oliver Kowalke 04 Feb '05

04 Feb '05

Hi, is it possible to use boost::function as drop-in replacment for ordinary function pointers - or are there some other solutions? I can't get the code below to work. thx and best regards, Oliver typedef void (*FunctionPointer) ( int); void my_fun( int i) { std::cout << i << std::endl; } class MyFun { public: void operator()( int i) { std::cout << i << std::endl; } }; class Test { public: void f( FunctionPointer fp) { fp( 1); } }; void main() { Test tst; MyFun mf; boost::function< void(int) > f; f = boost::ref( mf); // tst.f( my_fun); // works tst.f( f); // ??? - doesn't compile } -- DSL Komplett von GMX +++ Superg�nstig und stressfrei einsteigen! AKTION "Kein Einrichtungspreis" nutzen: http://www.gmx.net/de/go/dsl

2 1

Bug in boost::lambda using operator <<?
by Richard_Cox＠Dell.com 04 Feb '05

04 Feb '05

The program below outputs String: <First string> Second string> Fourth string> Fifth string> All but the first "<" is missing. Any ideas? (On vc++7.1 using Boost 1.32.0.) Richard All opinions are mine and do not represent in any way the position of Dell Corporation. #include <ostream> #include <iostream> #include <string> #include <vector> #include <algorithm> #include <boost/lambda/lambda.hpp> using std::wstring; using std::wcout; using std::endl; using boost::lambda::_1; typedef std::vector<wstring> StringCollection; wstring gt(L"<"); std::wostream& operator << (std::wostream& out, StringCollection const& coll) { std::for_each(coll.begin(), coll.end(), out << gt << _1 << L"> "); return out; } int main(int, char*[]) { StringCollection s; s.push_back(L"First string"); s.push_back(L"Second string"); s.push_back(L"Fourth string"); s.push_back(L"Fifth string"); wcout << L"String: " << s << endl; return 0; }

2 1

Use of dollar signs in variable names when using boost::archive::xml_iarchive
by lloyd.reed＠leica-microsystems.com 04 Feb '05

04 Feb '05

I have found a bug or undocumented limitation in the boost serialization library. I'm using VC++ 2003, which allows you to use variable names which include dollar signs ($). Variables with names containing dollar signs can be saved to an archive using the boost::archive::xml_oarchive class. When boost::archive::xml_iarchive is used to reload the variable, however, the parsing of the variable name fails, and basic_xml_iarchive<Archive>::load_start(const char *name) calls boost::throw_exception( archive_exception(archive_exception::stream_error) ). For example, std::string filename( "C:\\dollar.txt" ); int $(0); { std::ofstream archive_file( filename.c_str() ); if( archive_file.good() ) { boost::archive::xml_oarchive output_archive(archive_file); output_archive << BOOST_SERIALIZATION_NVP( $ ); } } produces the following file: <?xml version="1.0" encoding="UTF-8" standalone="yes" ?> <!DOCTYPE boost_serialization> <boost_serialization signature="serialization::archive" version="3"> <$>0</$> </boost_serialization> but trying to load the file causes a boost::archive::archive_exception to be thrown: { std::ifstream archive_file( filename.c_str() ); if( archive_file.good() ) { boost::archive::xml_iarchive input_archive(archive_file); input_archive >> BOOST_SERIALIZATION_NVP( $ ); // throws } } I think the spirit definitions for the names of variables are in C:\boost_1_32_0\libs\serialization\src\basic_xml_grammar.ipp and C:\boost_1_32_0\libs\serialization\src\xml_grammar.cpp. Name = (Letter | '_' | ':') >> *(NameChar); and template<> void xml_grammar::init_chset(){ Char = chset_t("\x9\xA\xD\x20-\xFF"); Letter = chset_t("\x41-\x5A\x61-\x7A\xC0-\xD6\xD8-\xF6\xF8-\xFF"); Digit = chset_t("0-9"); Extender = chset_t('\xB7'); Sch = chset_t("\x20\x9\xD\xA"); NameChar = Letter | Digit | chset_p("._:-") | Extender ; } ______________________________________________________________________ This email has been scanned by the MessageLabs Email Security System. For more information please visit http://www.messagelabs.com/email ______________________________________________________________________

2 2

shared_ptr on HP-UX - patch
by Jaroslav Gresula 04 Feb '05

04 Feb '05

I tried to compile the smart pointer regression tests on HP-UX and found out that the compiler takes the 'advanced' version of get_deleter() template function (defined in shared_ptr.hpp) which is evidently beyond its limits (we use aCC, version HP ANSI C++ B3910B A.03.35). In case anyone is interested, I attached a 'patch' against shared_ptr.hpp, revision 1.52 - it just redirects the compiler to the already existing workaround version. /Jarda *** shared_ptr.hpp.orig Wed Feb 2 22:12:50 2005 --- shared_ptr.hpp Wed Feb 2 21:49:23 2005 *************** *** 442,448 **** // get_deleter (experimental) ! #if (defined(__GNUC__) && (__GNUC__ < 3)) || (defined(__EDG_VERSION__) && (__EDG_VERSION__ <= 238)) // g++ 2.9x doesn't allow static_cast<X const *>(void *) // apparently EDG 2.38 also doesn't accept it --- 442,450 ---- // get_deleter (experimental) ! #if (defined(__GNUC__) && (__GNUC__ < 3)) ||\ ! (defined(__EDG_VERSION__) && (__EDG_VERSION__ <= 238)) ||\ ! (defined(__HP_aCC) && (__HP_aCC <=33500)) // g++ 2.9x doesn't allow static_cast<X const *>(void *) // apparently EDG 2.38 also doesn't accept it

3 2

weak_ptr_test - non-portable?
by Jaroslav Gresula 04 Feb '05

04 Feb '05

When I run the weak_ptr_test (a part of the regression tests) on HP-UX (aCC version HP ANSI C++ B3910B A.03.35) I'm getting a lot of failures. The reason is that IMHO the weak_ptr_test.cpp relies on a non-portable construct. There is copy_assignment() template function which takes shared_ptr<> by value. All caller sites of this function pass on an unnamed shared_ptr<> temporary. The first line of copy_assignment() checks if shared_ptr<>::unique() is true for the passed object. This test succeeds for all compilers we use in our shop but for the aCC. The reason is that upon entering copy_assignment() shared_ptr::use_count() is equal 2 as the compiler creates a temporary on the caller site and copy-ctor is used to pass on the object to copy_assignment() which seems to be in concert with the C++ standard [12.2]. Is my observation correct? Thanks, Jarda

2 1

Installation log
by Nicolas Busca 04 Feb '05

04 Feb '05

Hi! Is there any way to see the installation log of boost? When I installed it I got a message in the end that ~50 out of ~1000 links failed to be created. How could I know which they were? Thanks, Nicolas Busca.

2 1

BGL + SubGraph + Remove
by Gordon Smith 04 Feb '05

04 Feb '05

Is there anyway to remove a subgraph once created? Gordon.

1 0

[mpl::transform] How to make metafunction from transform?
by Vyacheslav E. Andrejev 04 Feb '05

04 Feb '05

Hello All! I want to parametrize mpl::transform to form a new metafunction. But I have confused with mpl syntax and notations. Imagine, I have following vector typedef boost::mpl::vector<Type1, Type2, Type3, Type4> types; and following transformation: typedef boost::mpl::transform< types , boost::is_convertible<Type3, boost::mpl::_1> >::type result; The result of the tansformation is a bool vector that indicates whether Type3 is convertible to any of type from types vector. Now I want to form metafunction that will construct such array not from Type3, but from its argument. Of course, I may use straightforward method: struct WhoConverted { template <typename ConversionType> struct apply { typedef typename boost::mpl::transform< types , boost::is_convertible<ConversionType, boost::mpl::_1> >::type type; }; }; But it seems more like workaround rather than solution. Can it be done in more convenient and short manner by means of all mpl machinery like mpl::lambda, mpl::bind, placeholders, etc... ? -- Vyacheslav E. Andrejev System Architect, Excelsior, LLC

2 2