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

RE: [Boost-Users] Problems with iterator_adaptors
by Ben Hutchings 11 Apr '03

11 Apr '03

Olaf Petzold wrote: > I try to write a cyclic Iterator using boost::iterator_adaptors. > Unfortunally, the advance function doesn't work for negative distances > correct: > template<class Iterator> > class cycle_iterator_policies > { > public: > cycle_iterator_policies(const Iterator& begin, const Iterator& end) > : m_begin( begin ), m_end( end ) > { > assert( m_begin != m_end ); > } > .... > template<class IteratorAdaptor, class DifferenceType> > void advance( IteratorAdaptor& x, DifferenceType n ) { > n %= std::distance( m_begin, m_end ); <snip> The problem is that the result of % and / (and by implication %= and /=) is implementation defined if either operand is negative. Most implementations round the quotient towards zero and produce a remainder with the same sign as the quotient. Evidently you want to have a non-negative remainder always. The simplest solution, I think, is: DifferenceType length = std::distance( m_begin, m_end ); n %= length; if (n < 0) n += length;

1 0

Re: thread lib using workshop CC on solaris
by Christopher Currie 11 Apr '03

11 Apr '03

I tried to send this earlier using the newsgroup, but it seems to have been dropped on the floor. Sorry for the duplicate if it shows up. The latest version of boost has some support for some versions of this compiler. The appropriate jamfile is "sunpro-tools.jam" (Sunpro is an old name for this compiler.) Thus: cd $BOOST_ROOT/libs/thread/build bjam -sTOOLS=sunpro That said, your mileage may vary on this compiler. I have found that it is notoriously buggy, especially wrt templates, and boost code will often cause it heartache in obscure ways. My results have found: * Up-to-date patches for the compiler and the operating system are a must. Find them here: http://access1.sun.com/sundev/index.html * Both versions of this compiler that I have tested have issues with Boost.TypeTraits, generating loads of spurious "Anacronism" warnings wrt the typename keyword being absent or present. Unfortunately, sun has not provided a flag to silence these warnings, and adding typename to the locations it complains about breaks other compilers, notably g++. * Both versions of this compiler will have problems linking some Boost unit tests, due to odd template RTTI issues. Adding -sBUILD="<cxxflags>-instances=static" will sometimes fix this problem. I don't recommend making this flag a default, because of the code bloat that entails. * Sun CC 5.4 ( Workshop/Forte Developer/SunOne Studio 7 ) will successfully build Boost.Threads. * Sun CC 5.3 update 2 ( W/FD/SOS 6u2 ) will only build Boost.Threads in release mode. This compiler generates assertions when attempting to compile the debug libraries. * I have not tested any versions earlier than Workshop 6 update 2. "jenny_qzhang" <jenny_qzhang(a)yahoo.com> wrote in message news:b722dr+7f4@eGroups.com... > Hi, All > > I am working on porting an application from MSVC to solaris > workshop(the applicatoin is using boost thread lib). I could not > locate the makefile or jam file for this compiler. Is there anybody > did the same thing before? > > If you finially made it, can you share your experience, makefile > with me? Actually I tried to write a makefile myself, but it reported > a lot of errors related to template parameters :(. > > > Thanks very much! > > Qiu

1 0

Problems with iterator_adaptors
by Olaf Petzold 11 Apr '03

11 Apr '03

Hello, I try to write a cyclic Iterator using boost::iterator_adaptors. Unfortunally, the advance function doesn't work for negative distances correct: template<class Iterator> class cycle_iterator_policies { public: cycle_iterator_policies(const Iterator& begin, const Iterator& end) : m_begin( begin ), m_end( end ) { assert( m_begin != m_end ); } .... template<class IteratorAdaptor, class DifferenceType> void advance( IteratorAdaptor& x, DifferenceType n ) { n %= std::distance( m_begin, m_end ); DifferenceType rest = std::distance( x.base(), m_end ); if( n < rest ) std::advance( x.base(), n ); else std::advance( x.base() = m_begin, n - rest ); } .... private: Iterator m_begin; Iterator m_end; }; It seems to be an algorithmic problem for me. How can I corrected it? Thanks Olaf

1 0

Signals and noncopyable
by svanechel 11 Apr '03

11 Apr '03

Hi, Is there a reason why siganls aren't copyable? As far as I can tell signal base derives from noncopyable, but it references its implementation using shared_ptr. So I'm guessing in theory a signal is perfectly copyable. I must admit I came to this conclusion by briefly looking at the code, so I might be completely wrong. I want to store signals into a vector of function<void()>'s, together with bound arguments, to capture events and replay them later. Something similar to (simplified): class recorder { public: // this is the handler for some event // it is connected to another signal void some_event( int i) { events_.push_back( bind( ev, i ) ); } void replay() { vector< function<void()> >::const_iterator itr = events_.begin(); while ( itr != events_.end() ) { (*itr)(); ++itr; } } signal<void(int)> ev; private: vector< function<void()> > events_; }; recorder rec; // connect the recorder to the signals is uses ... // now replay the events rec.ev.connect( some_handler ); rec.replay(); The example above may no be entirely correct since I'm coding this directly without running it through a compiler, but I hope the intent is clear. If I remove the derivation from noncopyable from signal_base, a sigal can be assigned to a function, although I haven't tested it yet to see if it actually works (currently I do not have something I can start testing with). Is it safe to remove the noncopyable base class? If so can it be done in the next release of the boost distribution as well? Thanks in advance, Sven Van Echelpoel

1 0

Problem with boost::pool_alloc
by Ilya Chvetsov(work) 11 Apr '03

11 Apr '03

Hi! I try to use boost::pool_alloc with STLport-4.5-0119 snapshot. STLPort call pool_alloc::allocate(const size_type n) with n = 0. STLPort naitive allocator returns in such case NULL, but pool_alloc tries return memory. Method simple_segregated_storage<SizeType>::try_malloc_n( void * & start, size_type n, const size_type partition_size) has problem too if parameter n is 0. Problem is in while condition: while (--n != 0)... This simple program demonstrates a problem: #include <boost/pool/poolfwd.hpp> #include <boost/pool/pool_alloc.hpp> typedef boost::pool_allocator<int> alloc_t; int main(int argc, char* argv[]) { alloc_t alloc; alloc.allocate(0); alloc.allocate(0); alloc.allocate(0); alloc.allocate(0); alloc.allocate(0); alloc.allocate(0); alloc.allocate(0); alloc.allocate(0); alloc.allocate(0); alloc.allocate(0); alloc.allocate(0); alloc.allocate(0); alloc.allocate(0); alloc.allocate(0); alloc.allocate(0); alloc.allocate(0); alloc.allocate(0); alloc.allocate(0); alloc.allocate(0); alloc.allocate(0); alloc.allocate(0); alloc.allocate(0); alloc.allocate(0); alloc.allocate(0); alloc.allocate(0); alloc.allocate(0); alloc.allocate(0); alloc.allocate(0); alloc.allocate(0); alloc.allocate(0); alloc.allocate(0); alloc.allocate(0); alloc.allocate(0); alloc.allocate(0); alloc.allocate(0); alloc.allocate(0); return 0; } How can I solve this problem? Thanks -Ilya

1 0

Boost random's min < max assert
by Glen Knowles 10 Apr '03

10 Apr '03

Uniform int contains an assert if min < max. If I ask for a random between 1 and 1 I'm expecting a 1, not an assert. Is there some reason this was deemed inappropriate? Thanks, Glen [Non-text portions of this message have been removed]

3 2

preprocessor: convert to capitals
by Toon Knapen 10 Apr '03

10 Apr '03

In the ublas forum we're having a discussion on the best way to provide signatures in C for fortran libraries (like BLAS and LAPACK). For linking C with fortran, you sometimes need to add an underscore to the function in C, but on other platforms fortran will store the function-name in capitals. Thus on different platforms, we should preprocess our source to take the platforms' convention into account and we thus would need a macro to convert an identifier automatically to uppercase. Is that possible ? toon -- Toon Knapen http://www.fft.be Free Field Technologies tel:+32 10 45 12 26 16 Place de l'Universite fax:+32 10 45 46 26 1348 Louvain-la-Neuve http://www.femtown.org Belgium

3 5

thread lib using workshop CC on solaris
by jenny_qzhang 10 Apr '03

10 Apr '03

Hi, All I am working on porting an application from MSVC to solaris workshop(the applicatoin is using boost thread lib). I could not locate the makefile or jam file for this compiler. Is there anybody did the same thing before? If you finially made it, can you share your experience, makefile with me? Actually I tried to write a makefile myself, but it reported a lot of errors related to template parameters :(. Thanks very much! Qiu

2 2

ublas matrix slice bug?
by Stephan Puchegger 08 Apr '03

08 Apr '03

Dear ublas developers! (I am using boost 1.30 & gcc 3.2) the following program: #include <boost/numeric/ublas/matrix.hpp> #include <boost/numeric/ublas/io.hpp> int main () { using namespace boost::numeric::ublas; matrix<double> m (4, 4); for (unsigned int i = 0; i < m.size1 (); ++ i) for (unsigned int j = 0; j < m.size2 (); ++ j) m(i, j) = 1; std::cout << m << std::endl; // multiply a slice of the matrix with 2.... project (m, slice (2, 1, m.size1()-2), slice (1, 0, m.size1()-2)) *= 2; // interesting result ... the slice has been multiplied with 4 std::cout << m << std::endl; } gives the following output: [4,4]((1,1,1,1),(1,1,1,1),(1,1,1,1),(1,1,1,1)) [4,4]((1,1,1,1),(1,1,1,1),(1,4,1,1),(1,4,1,1)) Why is the slice multiplied with 4 and not with 2?? Do I make some fundamental mistake? Everything works fine if I am using a matrix_vector_slice. My second (stupid?) question is the following: Is it intentional, that it is not possible to compile: #include <boost/numeric/ublas/matrix.hpp> #include <boost/numeric/ublas/io.hpp> int main () { using namespace boost::numeric::ublas; matrix<double> m (4, 4); for (unsigned int i = 0; i < m.size1 (); ++ i) for (unsigned int j = 0; j < m.size2 (); ++ j) m(i, j) = 4*i+j; std::cout << m << std::endl; swap(row(m,0), row(m,2)); std::cout << m << std::endl; } whereas matrix_row<matrix<double> > l(m, 0); matrix_row<matrix<double> > r(m, 2); swap(l, r); /* swap the rows */ works fine. MTL has support for a similar syntax: swap(rows(m)[0], rows(m)[2]); best regards and thanks for your help! Stephan

3 7

Spirit PDF or something
by Edward Diener 08 Apr '03

08 Apr '03

I would like to print out the Spirit documentation as a whole for offline study. However Boost 1.30 Spirit documentation comes as a series of HTML pages. Is there an easy way to print out the documentation rather than cycling through all of it manually and printing out a page at a time ?

2 3

condition::wait() - Alertable in Win32
by Simon Bailey 08 Apr '03

08 Apr '03

What is the latest thinking on making condition::wait() and condition::timed_wait() "alertable" in the Win32 environment? Please consider this a "me too" request for such a feature! I'm trying to implement a cross-platform (Unix, Windows) implementation of a Proactor pattern that manages asynchronous sockets. It seems that the way to implement this pattern with as much code common to both platforms as possible is to use 1. Overlapped sockets that call back (not post messages) in Windows. 2. Signals in Unix (similar to the Windows call backs but with less- clear semantics in my mind). For this scheme to work, the wait functions need to be alertable in Windows, but the Unix system shouldn't need any changes. >From the research that I have done so far, in my particular case, I would only need wait and timed_wait to be alertable, and it would be appropriate that they were always altertable - I would not need a way to enable or disable this feature.

1 0

Building (runtime-link-static) thread lib on vc7
by Taras Tielkes 07 Apr '03

07 Apr '03

Hi, Some newbie boost thread lib questions: 1) I used to build the 1.29 thread library with the runtime-link-static\threading-multi build configuration. This would yield (in release mode) a 467 KB import library in the destination directory. When I build boost 1.30 (with runtime-link-static\threading-multi), the target directory for the thread library contains: -a boost_thread.dll of 116 KB -boost_thread.lib import libarary of 27KB Programs compile against the import library have a runtime dependency on the boost thread dll (as I would expect from the size of the import .lib) What's the right way to compile a static-link, multi-threaded boost.threads library (using the vc7 toolset) in 1.30? 2) What happended to the threadmon.dll? Is it gone in 1.30? Thanks for any pointers..

1 0

RE: [Boost-Users] Re: Please, please help!!!
by scleary＠jerviswebb.com 07 Apr '03

07 Apr '03

> From: agiatojon [mailto:JonAgiato@nyc.rr.com] > > Thanks a lot for the info. I saw that but did not know exceptions > were disabled in some compilers by default, which as you suggested, I > found out this one does. I enabled them, and now it compiles. > > I have another question for you if I may. I attempted to simply > stream the numbers to the standard output but it complains that there > is no operator defined for this object. Is there a file I can include > that contains the overloaded output operator functionality? I am sure > this must be available, I am just not sure of its location. I know > that ublas uses the file "IO" in order to bring those definitions in. Make sure you're doing: std::cout << dev(); and not: std::cout << dev; If that isn't the problem, could you post a code sample? -Steve

2 1

specializing new lexical_cast
by Tom Matelich 07 Apr '03

07 Apr '03

I'm going to embark on it, but I thought I'd ask in parallel. One of the things about Terje's proposed lexical_cast was a lot of documentation of how to specialize it. That doesn't seem to be addressed in 1.30.0's version. I see in the documentation "Where non-stream-based conversions are required, lexical_cast is the wrong tool for the job and is not special-cased for such scenarios." I understand this sentiment, and I tried to search for discussions of this on the mailing list, but I did not conquer ASPN's search tool. The reason I'm asking is that we did some profiling and strtod kicked lexical_cast's butt for string -> double. I currently use lexical_cast in some template'd xml code, so I could overload my function, but I'd have to do it in about 5 spots due to an evolving interface. I'd like to just do it once for lexical_cast. I suppose I could use Terje's implementation, but I've stuck with boost canon for a long time now intentionally. Any suggestions? ----------------------------------------------------------------------- DISCLAIMER: Information contained in this message and/or attachment(s) may contain confidential information of Zetec, Inc. If you have received this transmission in error, please notify the sender by return email. -----------------------------------------------------------------------

4 4

New BGL warning in 1.30.0
by Steven Solie 07 Apr '03

07 Apr '03

I'm now getting the following warning from BGL 1.30.0: /gg/boost/boost/pending/property.hpp:71: warning: sorry: semantics of inline function static data `const bool match' are wrong (you'll wind up with multiple copies) /gg/boost/boost/pending/property.hpp:71: warning: you can work around this by removing the initializer The code that generates the above warning looks something like this: Territory* t1 = new Territory(...); Vertex vertex = add_vertex(m_graph); put(vertex_name, m_graph, vertex, t1); // generates warning The graph declaration looks like this: typedef adjacency_list<vecS, vecS, undirectedS, property<vertex_name_t, Territory*, property<vertex_distance_t, unsigned long> > > Graph; I don't have any problems with BGL 1.29.0 and I don't understand what the warning is trying to tell me. Any ideas? I'm using GCC 2.95.3 with STLport. --Steven

2 1

Spirit - rule generating functions
by brockpeabody 06 Apr '03

06 Apr '03

Does anyone know the best way to return a rule from a function by value? I know that the following does not work, but what's the best workaround? struct calculator : grammar<calculator> { template <typename ScannerT> struct definition { definition(const calculator&) { r = make_complex_rule('A') | make_complex_rule('B'); } typedef rule<ScannerT> rule_type; rule_type const& start() { return r;} private: rule_type make_complex_rule(char c) { return c >> real_p; //boom } rule_type r; }; }; my current workaround is the following nastiness: //inside struct calculator::definition<> rule_type& anonymous_rule() { anonymous_rules.push_back( boost::shared_ptr<rule_type>(new rule_type)); return *anonymous_rules.back(); } rule_type& make_complex_rule(char c) { return anonymous_rule() = c >> real_p; } std::vector<boost::shared_ptr<rule_type> > anonymous_rules; Is there a better way? Also, I've got an easy one... Why is a grammar passed by const reference to its definition? Does Spirit make copies of the grammars internally? I ask because otherwise the grammar class seems like the ideal place to store state information (like numbers on a calculator's stack). Thanks ahead of time! Brock Peabody brock.peabody(a)npcinternational.com

3 2

Tuple compatibility with VS.Net
by toddbfleming 06 Apr '03

06 Apr '03

I downloaded boost 1.30.0 but am having trouble with the Tuple library with Visial Studio.Net. The following code produces the errors: #include <boost/tuple/tuple.hpp> boost::tuple<int, int> f() { return boost::tuple<int, int>(1, 1); } c:\cpp\boost_1_30_0 \boost\tuple\detail\tuple_basic_no_partial_spec.hpp(411) : error C2039: 'map_tuple_to_cons' : is not a member of 'boost::detail' This code is: class tuple : public detail::map_tuple_to_cons<T1, T2, T3, T4, T5, T6, T7, T8, T9, T10>::cons1 This appears to be a compiler bug. The code at this line is in the boost::tuples namespace but the compiler seems to be ignoring boost::tuples::detail. I replaced detail with tuples::detail in many places in this code. Next error: c:\cpp\boost_1_30_0 \boost\tuple\detail\tuple_basic_no_partial_spec.hpp(456) : error C2629: unexpected 'boost::tuple<T1,T2,T3,T4,T5,T6,T7,T8,T9,T10> (' This code is: template<typename Head, typename Tail> explicit tuple(const cons<Head, Tail>& other) : cons1(other.head, other.tail) { } Visual C++ doesn't seem to understand template constructors. I replaced the template constructor with a copy constructor. VC++ failed to create the default copy constructor :( c:\cpp\boost_1_30_0 \boost\tuple\detail\tuple_basic_no_partial_spec.hpp(470) : error C2143: syntax error : missing ',' before '<' The code is: template<typename Head, typename Tail> self_type& operator=(const cons<Head, Tail>& other) Lookup failed on cons. I added tuples:: prefix. Has anyone else run into these or might I have a configuration problem? Thanks! Todd Fleming

1 0

documentation
by luise_mang 05 Apr '03

05 Apr '03

Hello, Is there some centralized documentation for offline-studies available (PDF, some platform-specific help docs, whatever - just download once, view offline many) ? thanx, Luise

1 0

RE: [Boost-Users] Please, please help!!!
by scleary＠jerviswebb.com 05 Apr '03

05 Apr '03

Your error message indicates that exceptions are disabled. Try not disabling them. I think some versions of GCC might disable exceptions by default based on the language its compiling for -- e.g., if it's invoked as 'gcc' or the filename extension designates a C file. I didn't see this behaviour under Cygwin, though. I can get it to compile under Cygwin just by adding the Boost include with -I and defining __linux__ with -D. HTH, -Steve > -----Original Message----- > From: Jon Agiato [mailto:JonAgiato@nyc.rr.com] > Sent: Friday, April 04, 2003 11:56 AM > To: Boost-Users(a)yahoogroups.com > Subject: RE: [Boost-Users] Please, please help!!! > > > Hi Steve, > Thanks again for the help. I did that and now the > file is recognized, but I am receiving the following > error upon compilation: > > In file included from main.cpp:3: > /usr/include/c++/3.2/boost/libs/random/random_device.cpp: > In member function > `void boost::random_device::impl::error(const > std::string&)': > /usr/include/c++/3.2/boost/libs/random/random_device.cpp:92: > exception handling > disabled, use -fexceptions to enable > gmake: *** [main.o] Error 1 > > The program is the same except now I am including > #include <boost/libs/random/random_device.cpp>. > > Thanks again! > > Jon Agiato > JonAgiato(a)nyc.rr.com > > --- Jon Agiato <jonagiato(a)nyc.rr.com> wrote: > > >> > >> ----- Original Message ----- > >> From: scleary(a)jerviswebb.com > >> To: Boost-Users(a)yahoogroups.com > >> Sent: Friday, April 04, 2003 10:49 AM > >> Subject: RE: [Boost-Users] Please, please help!!! > >> boost::random_device > >> > >> > >> There is a source file under the libs/random > >> directory that contains the > >> implementation of random_device. You'll have to > >> include it in your project > >> in order to link. > >> > >> -Steve > >> > > > > > >>> > -----Original Message----- > >>> > From: Jon Agiato [mailto:JonAgiato@nyc.rr.com] > >>> > Sent: Thursday, April 03, 2003 10:02 PM > >>> > To: Boost-Users(a)yahoogroups.com > >>> > Subject: [Boost-Users] Please, please help!!! > >> > >> > >> boost::random_device > > > > > >>> > > >>> > > >>> > Hello all, > >>> > > >>> > I am attempting to use the non-deterministic > >> > >> > >> random number generator > > > > > >>> > class contined in boost::random and have > >> > >> > >> encountered the following > > > > > >>> > errors upon running the program. The program does > >> > >> > >> compile without > > > > > >>> > errors, and I am using KDevelop 2.1.3 in the > >> > >> > >> RedHat Linux 8.0 > > > > > >>> > environment. > >>> > > >>> > Here's the offending code: > >>> > > >>> > #include <iostream> > >>> > #include <cstdlib> > >>> > #include <boost/nondet_random.hpp> > >>> > > >>> > int main() > >>> > { > >>> > boost::random_device dev; > >>> > > >>> > return EXIT_SUCCESS; > >>> > } > >>> > > >>> > I receive the following error message: > >>> > /home/JAgiato/boosttest/boosttest/main.cpp:6: > >> > >> > >> undefined reference to > > > > > >>> > `boost::random_device::default_token' > >>> > /home/JAgiato/boosttest/boosttest/main.cpp:6: > >> > >> > >> undefined reference to > > > > > >>> > > >> > >> > >> > > > > > `boost::random_device::random_device[in-charge](std::basic_str > > >> > ing<char, > >> > std::char_traits<char>, std::allocator<char> > > > > > > > const&)' > > >> > /home/JAgiato/boosttest/boosttest/main.cpp:8: > > > > > > undefined reference to > > >> > `boost::random_device::~random_device > > > > > > [in-charge]()' > > >> > > >> > My goal is to use this class to produce > > > > > > non-deterministic 32 bit > > >> > integers which will be streamed to a binary file. > > > > > > Any help is greatly > > > > apprechiated! Thank you! > > > > > > Jon Agiato > > > JonAgiato(a)nyc.rr.com > > > > ------------------------ Yahoo! Groups Sponsor > ---------------------~--> > Get 128 Bit SSL Encryption! > http://us.click.yahoo.com/xaxhjB/hdqFAA/VygGAA/EbFolB/TM > -------------------------------------------------------------- > -------~-> > > Info: <http://www.boost.org> > Wiki: <http://www.crystalclearsoftware.com/cgi-bin/boost_wiki/wiki.pl> > Unsubscribe: <mailto:boost-users-unsubscribe@yahoogroups.com> > > > Your use of Yahoo! Groups is subject to > http://docs.yahoo.com/info/terms/ > >

2 1

Extending max. arity of tuples
by Max Motovilov 04 Apr '03

04 Apr '03

I ran against the arity limitation of 10 and to resolve it, modified tuple_basic.hpp to use the preprocessor package for all variable length argument lists (with macro BOOST_TUPLE_MAX_ARITY serving as a single parameter configurable from config/user.hpp). I'm not sure if this modification will be of general interest, but if so, where do I send it? Modified code has been tested with MSVC 2003 (7.1). Regards, ...Max...

1 0

problems building boost thread with Linux7.2
by ZYA 04 Apr '03

04 Apr '03

Dear, How to build boost thread libs step by step? zya! Info: <http://www.boost.org> Wiki: <http://www.crystalclearsoftware.com/cgi-bin/boost_wiki/wiki.pl> Unsubscribe: <mailto:boost-users-unsubscribe@yahoogroups.com> Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/ [Non-text portions of this message have been removed]

2 1

RE: [Boost-Users] problems building boost thread with Linux7.2
by Jim.Hyslop 04 Apr '03

04 Apr '03

ZYA [mailto:Zhao.Yan@inventec.com] wrote: > How to build boost thread libs step by step? Have a look at these pages: http://www.boost.org/tools/build/index.html http://www.boost.org/libs/thread/doc/build.html The first page gives a fairly detailed set of instructions on building bjam, the tool which is used to build the thread library (and other Boost libraries). The second page tells you how to use bjam to build the library itself. If these pages don't help, feel free to post any additional questions (and I promise to read the subject line as well :=) -- Jim

1 0

Re: [Boost-Users] Re: Understanding MPL
by Tanton Gibbs 04 Apr '03

04 Apr '03

"Paul Mensonides" <yg-boost-users(a)m.gmane.org> wrote in message news:b6ifnh$5l0$1@main.gmane.org... > Edward Diener wrote: > > >> What specifically would you have me do? The pp-lib docs do not > >> attempt to explain the implementation of the pp-lib itself. > > > > I don't want an explanation of the inner workings of the pp-lib. I > > would like to see an explanation of the pp-lib which attempts to > > explain the functionality of it from the user's point of view. > > Something like a grouping of the macros in it from the point of view > > of general areas of functionality, with the specific macros grouped > > and explained within each area. It honestly sounds to me like Edward wants a tutorial -- something that the PP library does not currently have. I have used the PP lib numerous times for many different types of problems, and have always come away with a good feeling towards the docs. They are easy to navigate and are structured in such a way that you can find what you are looking for quickly. However, I also understand Edward's point about a tutorial. If I had not come to the PP lib with a particular problem, but instead just wanted to get a "warm, fuzzy" feeling about it, then it would be natural to want a tutorial to walk me through each macro giving an actual program and describing why each macro used in the program was chosen as well as what it expands to. I realize that this would be a huge undertaking, but perhaps there are people out there like myself, who have used and enjoy the PP lib that would help you in this endeavor if you choose to undertake it. I definately think it would be worthwhile to have a tutorial to help the unitiated get their feet wet. Basically, give them a problem to solve, show them how to solve it, then describe in detail how the solution was derived. The tutorial would then not get in the way of the initiated, but would bring the uninitiated up to speed. Thanks! Tanton [Non-text portions of this message have been removed]

1 0

boost::lexical_cast extension for boost::any ...
by Electric Monk 04 Apr '03

04 Apr '03

I have a need to cast a boost::any with an arbitrary contained type to another type, using a boost::lexical_cast for example. Consider the following: #include <boost/any.hpp> class obj { public: int val0; double val1; void setval(int i, boost:any const & val) { switch (i) { case 0: val0 = boost::any_cast<int>(val); break; case 1: val1 = boost::any_cast<double>(val); break; } } }; int main(int, char*[]) { obj o; o.setval(0, 1); o.setval(1, 2); return 0; } As expected, this throws a boost::bad_any_cast() at runtime. I'd like to write code similar to (from obj::setval above): case 0: val0 = boost::lexical_cast<int>(val); break; case 1: val1 = boost::lexical_cast<double>(val); break; Unfortunately boost::lexical_cast doesn't handle a parameter of type boost::any, since there isn't an overload of 'std::ostream operator <<()' for type boost::any. Can anyone suggest a simple way around this? My own musings on the subject have lead me to thinking that there needs to be a simple way to extract the contained value, or write it to a stream, without knowing its type. The simplest method I can see would be to add a function to boost::any::placeholder which either returns a string representation of the held value or writes the held value to a std::basic_ostream. This should allow a simple boost::lexical_cast conversion for boost::any to any arbitrary type. Comments? Have I missed anything major? Oh, and yes, I know that this goes beyond the intent of boost::any. I could easily create my own variant of boost::any which acts in the way I want, but I think this kind of functionality is useful in a number of cases, so it'd be nice to have the best solution possible... which is best discussed by the people here, who are much better versed in the boost library than I :) -- Corey Murtagh The Electric Monk "Quidquid latine dictum sit, altum viditur!"

2 1

RE: [Boost-Users] problems building boost thread with Linux7.2
by Jim.Hyslop 04 Apr '03

04 Apr '03

I wrote: > ZYA [mailto:Zhao.Yan@inventec.com] wrote: > > Dear, > > How to build boost thread libs step by step? > Step 1: tell us which compiler and platform you are using. My apologies, I didn't read the subject line. I use Windows, so I can't help. -- Jim

1 0