- Boost - lists.boost.org

[signal] performance
by Vladimir Prus 23 Dec '05

23 Dec '05

Hi, I've just tried to measure performance of Boost.Function and Boost.Signals using the following simple program: #include <boost/function.hpp> #include <boost/signals.hpp> using namespace std; void do_nothing() {} int main() { boost::function<void (void)> f = do_nothing; void (*pf)(void) = &do_nothing; boost::signal<void (void)> s; s.connect(&do_nothing); for(int i = 0; i < 10000000; ++i) s(); return 0; } and trying calls to 's()', 'pf()' and 'f()'. The timings are, on an Athlon 1700 with g++ 3.3.5 and -O3: Function pointer: 0.05 secs Boost.Function: 0.2 secs Boost.Signals: 9.8 secs Boost.Function is 4 times slower than function pointer, which seems decent for me (not that I mind further improvement). But Boost.Signals is nearly 50 times slower that Boost.Function, and nearly 200 times slower than function pointer. Any chance this will be improved? Say, I don't care about combining the results of signals invocation, or any connection tracking -- I just want to use boost::signals as more convenient variant of vector<boost::function>. Is it possible to have some preprocessor define that throws away all advanced functionality? Or a base class, say "lightweight_signal" that has only basics? TIA, Volodya

7 10

[spirit] [1.33.1] file_iterator needs workaround for VC8
by MB 23 Dec '05

23 Dec '05

Hi, Under VC8 Express Edition: #include <functional> #include <boost/spirit/iterator/file_iterator.hpp> // greeting namespace your { struct plus { }; } namespace my { using namespace your; struct A : plus { }; // error C2872: 'plus' : ambiguous symbol } I don't know what happens, but 'using namespace std;' at <iterator/impl/file_iterator.ipp:line41> does something bad. Simplified Code: namespace standard { struct plus { }; } // almost like 'using namespace standard' under vc8? namespace fileiter_impl { // 'using-directive' here using namespace standard; // 'class template' here template< typename CharT > class std_file_iterator { }; } namespace your { struct plus { }; } namespace my { using namespace your; struct A : plus { }; // error C2872: 'plus' : ambiguous symbol } Regards, MB p-stade.sourceforge.net

1 0

bug in signals library? - segmentation fault (only linked) -> exit(11) called
by Oliver.Kowalke＠infineon.com 23 Dec '05

23 Dec '05

Hello, I get an segmentation fault if I try to link against libboost_signals-gcc-1_33_1 or libboost_signals-gcc-mt-1_33_1! The application consist only of an empty main function and the binary is linked against the signals library. I've appended an output (tusc.out) of tusc (trace unix system calls) which indicates, that exit(11) is called and so a core dump created. System: HP/UX 11.11 PA-RISC Compiler: gcc-3.4.4 (with posix thread support) Code: int main( int argc, char* argv[]) { return 0; } Compiled with: if g++ -DHAVE_CONFIG_H -I. -I/eda-home/kowalke/Projects/test/src -I.. -I/opt/boost/include/boost-1_33_1 -O0 -g3 -MT test.o -MD -MP -MF ".deps/test.Tpo" -c -o test.o /eda-home/kowalke/Projects/test/src/test.cpp; then mv -f ".deps/test.Tpo" ".deps/test.Po"; else rm -f ".deps/test.Tpo"; exit 1; fi /bin/sh ../libtool --tag=CXX --mode=link g++ -O0 -g3 -o test -L/opt/boost/lib test.o -lboost_signals-gcc-1_33_1 mkdir .libs g++ -O0 -g3 -o test test.o -L/opt/boost/lib -lboost_signals-gcc-1_33_1 <<tusc.out>>

2 1

asio pooled_list/hash_map implementation
by christopher baus 23 Dec '05

23 Dec '05

I've been playing around with ways to remove some of the hits to the global allocator from asio/detail/reactor_op_queue.hpp, and decided to try an alternative approach to pooling. Instead of overriding STL allocators or global new/delete, I decided to implement pooled lists in terms of lists themselves. I take advantage of the fact that list::splice doesn't realloc nodes to move value data between the active and free lists during insert and erase operations. Chris's hash_map stores the value data in a std::list. I made that a template parameter, which allowed me to optionally specify the pooled_list. The implementation is here: http://baus.net/static/hash_map.hpp http://baus.net/static/pooled_list.hpp http://baus.net/static/pooled_hash_map.hpp This doesn't address the bigger problem of the handler allocation, but pre-allocing of the hash_map on *nix makes me a bit happier with the implementation. christopher

2 1

Re: [boost] Typeofand mpl(was Interest in dimension/unit-checking library?)
by Andy Little 23 Dec '05

23 Dec '05

"Andy Little" <andy(a)servocomm.freeserve.co.uk> wrote in message news:... > Hi Arkadiy, > > Its late and I'll have to reply on the rest of your post later. Meanwhile > enclosed is a small sample of the error output, which happens when I comment > out > registration for boost::mpl::math::reduced_rational in my registration Sorry I meant boost::mpl::math::simplified_rational rather than reduced_rational in my last post regards Andy Little

1 0

asio review
by Jody Hagins 22 Dec '05

22 Dec '05

I know it was started at 2 weeks. Any chance on an extension?

2 1

[review] asio review
by Brock Peabody 22 Dec '05

22 Dec '05

What is your evaluation of the design? I think the design is good. Asio has an easy to use, modern interface that doesn't compromise on portability or performance. What is your evaluation of the implementation? I haven't inspected the code but it sounds pretty impressive that it selects the optimal library for your platform while maintaining a consistent interface. What is your evaluation of the documentation? I found the tutorial easy enough to follow, but I haven't actually implemented anything with the library yet so I can't comment much on the reference section. What is your evaluation of the potential usefulness of the library? Highly. Everyone wants a cross-platform networking library for C++. Did you try to use the library? With what compiler? Did you have any problems? I compiled and ran some of the examples in the tutorial under VC 7.1 with no problems. How much effort did you put into your evaluation? A glance? A quick reading? In-depth study? I've been following discussions from other reviews. I read the tutorial and glanced through some of the rest of the documentation. Are you knowledgeable about the problem domain? I've written my own cross platform synchronous networking library. I've written countless client and server applications mostly under windows, all synchronous. I have no experience actually doing asynchronous IO. I vote to accept asio. Christopher has done a great job, thanks a lot! I'm excited to start using asynchronous IO. I've got server applications that will be pushing the limit on the number of threads soon. I will happily throw out all of my own networking code in favor of this library if it is accepted. My one small concern is that asio rewrites part of Boost.Thread. In the last couple of versions boost has become so easy to install and link against that I don't see the benefit of a header only implementation. Brock

1 0

who's using boost page
by Max Khesin 22 Dec '05

22 Dec '05

Just a style note: the "who's using" page is not (yet ;) big enough to justify a page with just categories (open source, shrink-wrap, etc). If the categorization is really desired it could be done with anchors within a page. Right now it just adds an extra click, and fragments the list that is more impressive as a single entity. 2c. This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. If you have received this email in error please notify the system manager. This message contains confidential information and is intended only for the individual named. If you are not the named addressee you should not disseminate, distribute or copy this e-mail.

1 0

Interest in a function object to sort strings more naturally?
by Andy Tompkins 22 Dec '05

22 Dec '05

I have written a function object to sort string alpha-numerically. If a string contains a number, that part is converted to an int and compared as ints. Thus "2" comes before "10". For example the list {a1, a10, a11, a2} would be sorted as {a1, a2, a10, a11} Example use code: { vector<string> my_list; //populate list sort(my_list.begin(), my_list.end(), lt_alphanum<char>()); } The class lt_alphanum splits the strings into sections containing digits, and non-digits. It then compares successive sections such that if two sections both contain digits, they are converted to ints and compared. Otherwise the sections are compared normally. I hope I am being clear. Andy Tompkins

1 0

[MPL.Math] compile-time double macro and printing
by Chris Weed 22 Dec '05

22 Dec '05

Hi, I am trying to simply print a compile-time double using gcc 4.0.0 on Sun. The following prints out : 1.9036e+185 #include <boost/metamath/double.hpp> #include <boost/metamath/double_macros.hpp> #include <iostream> int main() { typedef BOOST_DOUBLE(3.3) D; D d; std::cout << (double)d << std::endl; return 0; } Am I doing something wrong? Chris

4 5