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

boost::pool + shared memory
by Tom Matelich 07 Jan '02

07 Jan '02

I need to build a system for having miscellaneous objects in a shared memory segment. Essentially, we need a few inter-process singletons. These classes have STL containers, so I was looking at using pool_allocator and writing a UserAllocator to use ACE_Mem_Map ( http://doc.ece.uci.edu/Doxygen/Current/html/ace/classACE__Mem__Map.html#_det ails ) or something. I just wanted to see if anyone had done anything resembling this and could give me any pointers or tips. Thanks, Tom Matelich PS Why can't I see who is a member of this group? I wanted to check to see if Steve Cleary was a member. Sorry if you are Steve.

1 0

overflow of clock_t with clock(), timing longer processes
by gigu74 07 Jan '02

07 Jan '02

hello hackers first a hi to everyone. just signed today as a new member to this mailinglist. well.. i've got a problem and hope someone can help me. i run a c++ program (mostly on a sunblade 100 or a sunblade 1000) about genetic algorithms which takes quite some time to solve for certain problems (up to 100 hours) because its important to me to compare different jobs i need to measure the elapsed cpu time used for the particular jobs. well, i could use the standart system time, but because my running job is not always the only one i dont get really worthy results. so i used the std::clock(). but measuring 1/1000000 of a seconds with that and with the clock_t beeing a long int of course this results in a overflow after around 36 minutes. thought to find something about that in the boost library but timer.hpp just gets also de maximum elapsed time out.. doesnt really help is there some way this can be achieved? would be really neat to get an answer from anyone soon. thanks uf

1 0

bind question
by Duane Murphy 05 Jan '02

05 Jan '02

Is there a syntax for making boost::bind bind to functions that are templates? Here is an (extremally simplified) example that doesnt compile: #include <boost/bind.hpp> template < class Function > void f( Function func ) { func( 46 ); } void g( int num ) { num++; } int main( void ) { boost::bind( f, g )(); boost::bind( f, _1 )( g ); return 0; } The two bind lines in main both get an error "call of non-function". Is there a way to make this work? ..Duane p.s. CodeWarrior 7 on a Macintosh (OS X)

2 3

Another Bind Question
by Duane Murphy 04 Jan '02

04 Jan '02

Bind has been working very well for some of the things I have been wanting to do. I have run into a rather puzzling problem this time and was wondering what was happening: Here is the example: void f( int ) { } struct array_obj { int operator[]( int ){ return NULL; } }; int main() { array_obj a; int array[1]; int value = a[0]; boost::bind( f, _1 )( 3 ); boost::bind( f, _1 )( value ); boost::bind( f, _1 )( array[0] ); boost::bind( f, _1 )( a[0] ); return 0; } The first three bind statements compile just fine. The third one issues an error: function call 'operator()(int)' does not match. Along with all 20 variations of operator() generated by bind. Obviously there is something amiss with going through the return value of operator[](). I also get the same problem using an iterator which would be going through operator*(). I am using CodeWarrior 7 on Mac on OS X. Could this be a compiler specific problem? I can work around the problem using bind1st but it seems odd that bind1st works and bind doesnt? Could CW have included some magic in bind1st to work around this? Thanks for the help, ..Duane

2 1

Using boost::thread under Visual C++
by John Maddock 03 Jan '02

03 Jan '02

>I'm still flummoxed. If you have more ideas, I would really appreciate >them. Why not just use your IDE to build threadmon.cpp as a dll, and the rest of the source as a static lib, using whatever your project options are? - John Maddock http://ourworld.compuserve.com/homepages/john_maddock/

1 0

iterator_adaptors on Tru64 version V6.1-027
by vvasta 02 Jan '02

02 Jan '02

Hi all, I'm trying to use the iterator_adaptors class on Tru64. To make it work I had to do the following changes: In concept_check.hpp I had to change the code from this: template <class Concept> inline void function_requires() { #if !defined(NDEBUG) void (Concept::*x)() = BOOST_FPTR Concept::constraints; ignore_unused_variable_warning(x); #endif } to this: template <class Concept> inline void function_requires(type<Concept> t = type<Concept>()) { #if !defined(NDEBUG) void (Concept::*x)() = BOOST_FPTR Concept::constraints; ignore_unused_variable_warning(x); #endif } because this compiler generates an error if a template parameter is not used in the function declaration. In iterator_adaptors.hpp I had to remove the parenthesis around all the if_true bool template parameter from this: typedef typename if_true<( // original version ::boost::is_same<Value,typename iterator_traits<Iterator>::value_type>::value )>::template then< typename iterator_traits<Iterator>::pointer, Value* >::type pointer; to this: typedef typename if_true< // Changed version no parenthesis ::boost::is_same<Value,typename iterator_traits<Iterator>::value_type>::value >::template then< typename iterator_traits<Iterator>::pointer, Value* >::type pointer; It is possible to put this changes in the original distribution in order to avoid to put this changes in any new boost release I use ? Thanks for your help Vincenzo

1 0

Building nested FSM with BGL...
by Herman Bruyninckx 31 Dec '01

31 Dec '01

Has anyone already implemented a nested Finite State Machine with the Boost Graph Library? (Or a Petri net for that matter.) The thing that really puzzles me is how to implement the nesting, i.e., all nodes in a subset of the graph react to the same ``events'' from the ``outside world'', but I want to avoid making all the edges from the ``outside world'' nodes to the subset nodes... Herman

1 0

Re: new component, extent_set<>
by Nathan Myers 30 Dec '01

30 Dec '01

"David A. Greene" wrote: > Nathan Myers wrote: > > boost::extent_set<> implements a set of half-open ranges on a scalar > > type, with automatic splitting and merging, appropriate for managing > > ranges of disk blocks or video frames. In use it looks like: > > Cool! Why "extent_set" though? What about range_set or interval_set? "Extent" is the standard term of art when speaking of ranges of disk blocks, particularly as part of a file system. Also, since it's not really a set of ranges, range_set seemed inappropriate. > Any way this could be adapted to allow non-scalar ranges? What are > the requirements on the range element type? I would think anything > less-than-comparable should be useable but I didn't implement the > class. :) The current version uses closed ranges, BTW. The requirements are minimal: as listed in the header file, the Bound argument must be copyable, assignable, equality-comparable, comparable using the Compare argument, and incrementable and decrementable using the Succ and Pred arguments. It imposes no other requirements, so a non-scalar could work. Unfortunately I don't know how to define a useful decrement operator for a string, or a Compare operator for a complex number. Did you have some particular non-scalar in mind? The original version didn't need the increment and decrement operations, internally, but the client was still obliged to apply the equivalent of an increment operation. That made it slippery to use on strings. You could define increment to mean "append a space", so that a dictionary page might occupy the range "lost" to "love " (note the space). That assumes a lot about the character encoding, which the client (but not extent_set<>) is allowed to do. Unfortunately there is no simple equivalent for decrementing, so the trick doesn't work with closed ranges. It is conceivable that I might be able to continue storing half-open ranges internally, and thus avoid needing a decrement operation. It seems a bit fragile though. > Is there a way to control when ranges are combined (maybe sometimes > I want [10, 20) and [20, 30) to stay separate)? A policy for this > would be nice. It doesn't provide that in the current version. In fact, the presence of such a pair of extents would violate an invariant, breaking the insert and erase members as written. (They could probably be changed to accommodate it, at some cost in complexity.) How would you use such a feature? Nathan Myers ncm at cantrip dot org

1 0

Presentation Boost at Fosdem
by Toon Knapen 27 Dec '01

27 Dec '01

Dear Boost Developers, Currently, a presentation of Boost is scheduled at FOSDEM (www.fosdem.org) Fosdem is an anual free and open-source developer meeting in Brussels, Belgium that is gaining a lot of momentum, with an audience from all over europe and many international speakers of highly visible open-source projects. As has been discussed on the list before, Boost is well known to the standard commitee but need also be more visible to a wider public. That's also why I proposed to the Fosdem organiser (Raphael Bauduin) to schedule a Boost presentation. Boost developers and users that are planning to attend, contribute, ... please let me know. Toon

1 0

new component, extent_set<>
by Nathan Myers 27 Dec '01

27 Dec '01

I have uploaded to the Boost files section couple of archives containing an implementation of a class template extent_set; look in the folder "extent_set". The zip and tgz archives there are identical. Included in each is a header file and a test suite. boost::extent_set<> implements a set of half-open ranges on a scalar type, with automatic splitting and merging, appropriate for managing ranges of disk blocks or video frames. In use it looks like: boost::extent_set<int> set; set.insert(10, 20); // now contains [10,20) set.insert(30, 50); // now contains [10,20) [30,50) set.erase(40, 45); // now contains [10,20) [30,40) [45,50) set.insert(20, 30); // now contains [10,40) [45,50) std::cout << "(" << set.begin()->first << "," << set.begin()->second << ")" << std::endl; // writes "(10,40)" It is implemented using std::map. This is the initial submission. Please 'ave a look. (I am only subscribed to the -users list, so far.) Nathan Myers ncm at cantrip dot org

2 1