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

Re: [Boost-users] bind/function: composition
by Johan Torp 03 Oct '06

03 Oct '06

Monday, October 2, 2006, 10:05:23 PM, you wrote: burlbo> Johan Torp wrote: >> I often tend to have interfaces which take a boost::function<void()> >> that is to be called when an event occurs. Quite often I need make two >> separate function calls as response to one event. In these cases I >> usually solve the problem using the following technique: >> >> void foo(); >> void bar(); >> >> void call_two(const boost::function<void ()>& one, const >> boost::function<void ()>& two) >> { one(); two(); } >> >> boost::function<void ()> on_event = boost::bind(&call_two, >> boost::bind(&foo), boost::bind(&bar)); burlbo> Why not just use Boost Signals to have a single signal just call your burlbo> two different slots of foo and bar ? It seems you are unnecessarily burlbo> trying to re-invent a wheel which has already been re-invented by the burlbo> signals library. Good point. Actually I have developed some classes which closely resemble the signals library which I had already used extensively when I found signals. Maybe I will switch to boost signals when some shortcoming of my "library" becomes appearent, but for now it works nicely and I prefer to not mix APIs. In this particular case, boost::lambda solved the problem. I have a strong functional programming background so doing things like this (higher order programming) is much more natural to me than having to think about concepts such as signals and slots. -- Best regards, Johan Torp mailto:johan.torp@touchtable.se

1 0

Re: [Boost-users] [BGL] Boost poor performance (against LEDA)
by Sohail Somani 02 Oct '06

02 Oct '06

> -----Original Message----- > From: boost-users-bounces(a)lists.boost.org > [mailto:boost-users-bounces@lists.boost.org] On Behalf Of Leandro Melo > Sent: Monday, October 02, 2006 3:38 PM > To: boost-users(a)lists.boost.org > Subject: Re: [Boost-users] [BGL] Boost poor performance (against LEDA) > > Ok Sohail. I agree with that. So we may be all talking the same thing > but in different words. > The way you formulate your sentence is very nice. Hey thanks!! Usually I'm incomprehensible :)

1 0

Re: [Boost-users] [BGL] Boost poor performance (against LEDA)
by Sohail Somani 02 Oct '06

02 Oct '06

> -----Original Message----- > [mailto:boost-users-bounces@lists.boost.org] On Behalf Of Leandro Melo > > Hi David. > I think there's a misunderstanding here. I believe that I know many of > the benefits of Generic Programming. But my poiny is that I don't > agree that algorithms implemented generically can be considered to be > faster just because of that. No, but its easier to make algorithms faster because of it. It ends up almost working like multiple dispatch. I'm speaking generically about generic programming, not specifically BGL. Eventually you hit an "aha" moment where runtime code can be moved to compile time through generic techniques. Atleast that's what I've found. I'm sure a true artist can skip the "aha" moment.

2 1

Whitespace ins boost::filesystem::path
by shookie＠email.de 02 Oct '06

02 Oct '06

All I want to know is, wether there is a specific reason the path class can't deal with a normal space in any path. path p("a b"); throws filesystem errors. I'd like to point out, that this omission basically defeats the purpose of this library to be usabel cross-platform, since the two most used systems have directory names with spaces in it for very commonly used paths. Neither windows with "My Documents" or "Program Files" nor OS X with "Application Support" will allow to use the boost::filesystem::path class. And if you can't use it for those platforms, what is the point of using it if you still need to write platform specific code? Regards Jan Schukat

2 1

bug in boost::multi_index::sequenced splice (patch)
by Hubert Schmid 02 Oct '06

02 Oct '06

Hi, I guess there is a bug in boost/multi_index/sequenced_index.hpp. The following program fails to compile, because there is no matching operator== for class C. But this operator shouldn't be required for splice. #include "boost/multi_index/sequenced_index.hpp" #include "boost/multi_index_container.hpp" struct C { }; int main() { using namespace boost; using namespace boost::multi_index; typedef multi_index_container<C, indexed_by<sequenced<> > > Cs; Cs cs; cs.push_back(C()); cs.splice(cs.end(), cs, cs.begin()); // XXX: compile error } The following program compiles, but both assertions fail at runtime: #include "boost/multi_index/sequenced_index.hpp" #include "boost/multi_index_container.hpp" int main() { using namespace boost; using namespace boost::multi_index; typedef multi_index_container<int, indexed_by<sequenced<> > > ints; ints foo; foo.push_back(42); ints bar; bar.push_back(42); foo.splice(foo.end(), bar, bar.begin()); assert(foo.size() == 2); assert(bar.size() == 0); } I have made both tests with the boost library in Debian/unstable (version 1.33.1-7). I have also compared the source file sequenced_index.hpp with the CVS-MAIN version (1.12). I guess the following patch should fix this bug: --- sequenced_index.hpp.orig 2006-09-29 11:38:06.000000000 +0200 +++ sequenced_index.hpp 2006-09-29 11:39:43.000000000 +0200 @@ -355,7 +355,7 @@ BOOST_MULTI_INDEX_CHECK_DEREFERENCEABLE_ITERATOR(i); BOOST_MULTI_INDEX_CHECK_IS_OWNER(i,x); BOOST_MULTI_INDEX_SEQ_INDEX_CHECK_INVARIANT; - if(x==*this){ + if(&x==this){ if(position!=i)relink(position.get_node(),i.get_node()); } else{ @@ -388,7 +388,7 @@ BOOST_MULTI_INDEX_CHECK_IS_OWNER(last,x); BOOST_MULTI_INDEX_CHECK_VALID_RANGE(first,last); BOOST_MULTI_INDEX_SEQ_INDEX_CHECK_INVARIANT; - if(x==*this){ + if(&x==this){ BOOST_MULTI_INDEX_CHECK_OUTSIDE_RANGE(position,first,last); if(position!=last)relink( position.get_node(),first.get_node(),last.get_node()); regards, Hubert

5 5

Re: [Boost-users] Whitespace ins boost::filesystem::path
by shookie＠email.de 02 Oct '06

02 Oct '06

Ok, found the name_checking constructors, and now slap me for all that it's worth. I apologize for my 4:00 in the morning 15 h straight coding session mishaps. I should have slept before before crying out loud ;) Regards Jan Schukat

1 0

Re: [Boost-users] [function] How do you bind two boost::functionobjects together?
by Sohail Somani 02 Oct '06

02 Oct '06

> -----Original Message----- > From: boost-users-bounces(a)lists.boost.org > [mailto:boost-users-bounces@lists.boost.org] On Behalf Of Jason House > > I want the result of the bound expression to have one input parameter > that would be decided later. For instance: > > bool some_function(char x); > boost::function<bool (char) some_function_object(some_function); > boost::function<bool (int)> f = convert(some_function_object) > std::cout << f(32); I think what you're looking for is composition of functions? i.e. g o f (x) => g(f(x))? The right nomenclature might help... What it seems you want to do is to generate a function that does this: bool the_result(int i) { return some_function(convert_to_char(int)); } Is that right? I know you can curry with boost function + lambda but I haven't tried composition myself (though it should be possible!)

2 2

How to pass data to thread
by Monica Gretzer 02 Oct '06

02 Oct '06

Hi, I want to pass data to a thread function. My question is- what is wrong with this code? The compiler error output is: "error: function returning a function". In the code below I am trying to pass an int to the thread function. // ------------- code: how to pass data to the thread ? ---------------- #include <boost/thread/thread.hpp> #include <boost/bind.hpp> #include <iostream> using namespace std; class TestThread { public: void thread(int a); void CreateThreads(void); }; void TestThread::thread(int a) { cout << "I am thread " << a << endl; } void TestThread::CreateThreads() { boost::thread t1( boost::bind( &TestThread::thread, 10)); boost::thread t2( boost::bind( &TestThread::thread, 20)); // wait for the thread to finish t1.join(); t2.join(); } int main() { TestThread t; t.CreateThreads(); }

4 5

bind/function: composition
by Johan Torp 02 Oct '06

02 Oct '06

I often tend to have interfaces which take a boost::function<void()> that is to be called when an event occurs. Quite often I need make two separate function calls as response to one event. In these cases I usually solve the problem using the following technique: void foo(); void bar(); void call_two(const boost::function<void ()>& one, const boost::function<void ()>& two) { one(); two(); } boost::function<void ()> on_event = boost::bind(&call_two, boost::bind(&foo), boost::bind(&bar)); Is there any other way to acheive this, beside having to redeclare and define a function such as call_two? It seems quite easy to define something like: template <typename R1, typename R2> void nullary_compose boost::function<R1 ()>& one, const boost::function<R2 ()>& two); template <typename R1, typename R2, typename R3> void nullary_compose boost::function<R1 ()>& one, const boost::function<R2 ()>& two, const boost::function<R3 ()>& three); ... Is anybody else interested in seeing such templates in boost? Best Regards, Johan Torp -- System Developer TouchTable AB Tel: +46 (0)31 773 68 12 johan.torp(a)touchtable.se

3 2

Inverse matrix
by Karin Thomas 02 Oct '06

02 Oct '06

I would like to know whether Boost library is convenient for calculating an inverse matrix. Or does anyone suggest another way to do it? Kind regards, Karin Thomas

1 0