May 2006 - Boost - lists.boost.org

Re: [boost] [BULK] Re: boost::function overhead?
by Sohail Somani 02 May '06

02 May '06

> -----Original Message----- > From: boost-bounces(a)lists.boost.org > [mailto:boost-bounces@lists.boost.org] On Behalf Of David Abrahams > Sent: Tuesday, May 02, 2006 2:24 PM > To: boost(a)lists.boost.org > Subject: Re: [boost] [BULK] Re: boost::function overhead? > > "Sohail Somani" <s.somani(a)fincad.com> writes: > > > Dave A. had a nice explanation in SD expo about how its done for > > shared_ptr. I don't know if the slides are available. > > Thanks. I did talk about that technique, but shared_ptr uses virtual > functions, and I don't think I'd have said otherwise. Nevermind, it must have been boost function. I can't recall exactly which, but I understood :) Something like fn_table[idx]=function<T>; // function<T>(void *)

1 0

Re: [boost] [BULK] Re: boost::function overhead?
by Sohail Somani 02 May '06

02 May '06

> -----Original Message----- > From: boost-bounces(a)lists.boost.org > [mailto:boost-bounces@lists.boost.org] On Behalf Of Arkadiy Vertleyb > Sent: Tuesday, May 02, 2006 11:02 AM > To: boost(a)lists.boost.org > Subject: [BULK] Re: [boost] boost::function overhead? > Importance: Low > > "Doug Gregor" <dgregor(a)cs.indiana.edu> wrote > > > boost::function doesn't use virtual functions, but the effect is the > > same. boost::function adds one additional indirect call through a > > function pointer. > > Sorry for the inaccurate response. > > Out of curiousity, is this technique described anywhere > (other than the > code)? It's hard to imagine the way to erase the type of a > functor without > some kind of a polymorphic adaptor... Dave A. had a nice explanation in SD expo about how its done for shared_ptr. I don't know if the slides are available.

2 1

[numeric/conversion] Patch for BCB2006
by Nicola Musatti 02 May '06

02 May '06

Hallo, The following patch appears to be necessary to get BCB2006 to pass the conversion tests. If nobody has any objections (Fernando?) , I'll commit it in a few days. Cheers, Nicola Musatti ******Index: cast.hpp =================================================================== RCS file: /cvsroot/boost/boost/boost/numeric/conversion/cast.hpp,v retrieving revision 1.8 diff -d -u -r1.8 cast.hpp --- cast.hpp 27 Jun 2005 15:05:12 -0000 1.8 +++ cast.hpp 27 Apr 2006 20:43:24 -0000 @@ -22,7 +22,7 @@ #include <boost/detail/workaround.hpp> -#if BOOST_WORKAROUND(BOOST_MSVC, < 1300) || BOOST_WORKAROUND(__BORLANDC__, == 0x564) +#if BOOST_WORKAROUND(BOOST_MSVC, < 1300) || BOOST_WORKAROUND(__BORLANDC__, BOOST_TESTED_AT(0x582)) # include<boost/numeric/conversion/detail/old_numeric_cast.hpp>

2 3

Re: [boost] [SoC][string_cvt] The proposal for stringconversionslibrary.
by Martin Bonner 02 May '06

02 May '06

----Original Message---- From: Alexander Nasonov > As I pointed out in other thread, I implemented integral2str but > it converts according to "C" locale. I'm going to add thousands_sep > after every group of 3 digits for locales with appropriate grouping. > I expect a slight performance degradation but it should be much > faster then ostringstream anyway. > As I don't know anything about locales with other grouping, I can't > say how efficient integral2str might be for those locales. The culturally correct formating in (most of) India would be something like "12,34,56,789" grouping() returns a vector with .size()=2, [0]='\3', and [1]='\2'. I don't know how much that helps ... -- Martin Bonner Martin.Bonner(a)Pitechnology.com Pi Technology, Milton Hall, Ely Road, Milton, Cambridge, CB4 6WZ, ENGLAND Tel: +44 (0)1223 203894

2 1

[Boost-bugs] [ boost-Feature Requests-1480523 ] thread identification and getting the id
by SourceForge.net 02 May '06

02 May '06

Feature Requests item #1480523, was opened at 2006-05-02 09:48 Message generated for change (Tracker Item Submitted) made by Item Submitter You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=357586&aid=1480523&group_… Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: thread Group: None Status: Open Resolution: None Priority: 5 Submitted By: Nobody/Anonymous (nobody) Assigned to: William E. Kempf (bill_kempf) Summary: thread identification and getting the id Initial Comment: There are no identification for created threads in thread_group (at least working correctly under Linux). If it will be possible to set an id in function "create_thread" and have a function to get the id of current thread, then it will be a way to connect certain thread with some other resources (to do monitoring for example and for other purposes) ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=357586&aid=1480523&group_… ------------------------------------------------------- Using Tomcat but need to do more? Need to support web services, security? Get stuff done quickly with pre-integrated technology to make your job easier Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642 _______________________________________________ Boost-bugs mailing list Boost-bugs(a)lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/boost-bugs

1 0

Re: [boost] Including thread.hpp causes coredump on AIX
by chris.grayson＠database-brothers.com 02 May '06

02 May '06

How about these linker errors on AIX? Do others see these with Boost on AIX? (As previously noted, the code I'm linking in this example doesn't even make Boost calls...) chris.grayson(a)database-brothers.com wrote: > I do get these linker errors: > ld: 0711-224 WARNING: Duplicate symbol: ._GLOBAL__I_thread_creds_t > ld: 0711-224 WARNING: Duplicate symbol: ._GLOBAL__D_thread_creds_t > ld: 0711-224 WARNING: Duplicate symbol: _GLOBAL__F_thread_creds_t > ld: 0711-224 WARNING: Duplicate symbol: thread_creds_t > ld: 0711-224 WARNING: Duplicate symbol: _GLOBAL__I_thread_creds_t > ld: 0711-224 WARNING: Duplicate symbol: _GLOBAL__D_thread_creds_t > ld: 0711-345 Use the -bloadmap or -bnoquiet option to obtain more > information. P.S. Apologies again for the empty messages. Apparently mailing from Thunderbird on Windows doesn't work quite right. :-(

1 0

Re: [boost] Including thread.hpp causes coredump on AIX
by chris.grayson＠database-brothers.com 02 May '06

02 May '06

I don't know why my posts aren't coming through - here's my third try at sending my original message, with the three source files (again). I'm trying to use the Boost threads library in an application on AIX 5.3, but I'm having serious trouble. I've narrowed it down to the test code below. When both "main.cpp" and "Class1.cpp" include <boost/thread/thread.hpp>, the compiled application coredumps whenever an exception is thrown in Class1. When the thread.hpp include statement is commented out - in either file - the exception is thrown and caught fine, with no coredump. Note that neither main.cpp nor Class1.cpp *uses* any Boost threads code. This problem occurs with both gcc 3.3.2 (as provided by IBM in their "AIX Toolbox") and gcc 3.4.6 (as built from source). In both cases, the AIX linker "ld" is used, per both IBM and GNU's recommendations. This seems to only happen on AIX 5.3. The same code on Linux, Solaris and AIX 5.1, compiled the same way, does not crash in this way. This also occurs with both Boost 1.33.0 and 1.33.1. I'm currently compiling against the latest Boost source (1.33.1), though it's not built, since I'm only including headers, and not linking anything. ---- main.cpp ---- #include <iostream> #include "Class1.h" #include <boost/thread/thread.hpp> main () { std::cout << "starting main()" << std::endl; Class1 c; c.noop(); std::cout << "exiting main() normally" << std::endl; } ---- Class1.h ---- class Class1 { public: void noop(); }; ---- Class1.cpp ---- #include <iostream> #include "Class1.h" #include <boost/thread/thread.hpp> void Class1::noop () { std::cout << "entering Class1::noop()" << std::endl; try { std::cout << "..throwing exception" << std::endl; throw (2.1); } catch (...) { std::cout << "..caught exception in Class1.noop()" << std::endl; } std::cout << "..leaving Class1.noop()" << std::endl; return; } The previous files are built and run with these commands: > g++ -o crasher -I$HOME/boost/boost_1_33_1 main.cpp Class1.cpp > ./crasher I do get these linker errors: ld: 0711-224 WARNING: Duplicate symbol: ._GLOBAL__I_thread_creds_t ld: 0711-224 WARNING: Duplicate symbol: ._GLOBAL__D_thread_creds_t ld: 0711-224 WARNING: Duplicate symbol: _GLOBAL__F_thread_creds_t ld: 0711-224 WARNING: Duplicate symbol: thread_creds_t ld: 0711-224 WARNING: Duplicate symbol: _GLOBAL__I_thread_creds_t ld: 0711-224 WARNING: Duplicate symbol: _GLOBAL__D_thread_creds_t ld: 0711-345 Use the -bloadmap or -bnoquiet option to obtain more information. Any help will be greatly appreciated. Thanks, Chris Grayson Database-Brothers, Inc.

2 1

Re: [boost] [SoC] Specialized containers with Boost.MultiIndex
by JOAQUIN LOPEZ MU?Z 02 May '06

02 May '06

----- Mensaje original ----- De: Matias Capeletto <matias.capeletto(a)gmail.com> Fecha: Lunes, Mayo 1, 2006 6:59 am Asunto: Re: [boost] [SoC] Specialized containers with Boost.MultiIndex > I am finishing my application to the (M)ulti (I)ndex (S)pecialized > (C)ontainers (from now on MISC, do you like it?) for the SoC. > I am constructing a quickbook with a formal proposal, interface > design, project plan and personal information. I want to share it with > the list. I will really appreciate some feedback before sending the > application, specially from Joaquin. Can i post it here when it is > mature enough? Sure, I'm eager to read it. Best, Joaquín M López Muñoz Telefónica, Investigación y Desarrollo

4 6

[fusion] How does is_pointer< Sequence > work?
by Andreas Huber 02 May '06

02 May '06

Hi Joel & Dan Going through the fusion quick start, I'm wondering how the expression ... for_each(filter_if<boost::is_pointer<Sequence> >(seq), print_xml()); ... could ever work? Shouldn't that rather read similar to ... for_each(filter_if<boost::is_pointer<_> >(seq), print_xml()); ... so that the _ placeholder can be replaced with the individual types in the sequence? Thanks, -- Andreas Huber When replying by private email, please remove the words spam and trap from the address shown in the header.

4 3

Re: [boost] [Boost Review] Do property_tree and program_options overlap?"
by JOAQUIN LOPEZ MU?Z 01 May '06

01 May '06

----- Mensaje original ----- De: Gennadiy Rozental <gennadiy.rozental(a)thomson.com> Fecha: Lunes, Mayo 1, 2006 6:43 pm Asunto: Re: [boost] [Boost Review] Do property_tree and program_options overlap?" > "bwood" <brass(a)mailvault.com> wrote in message > > I read those and was sympathetic to the terminology > > arguments you made. L?pez Mu?oz believes it is too late > > to make those sort of changes. I disagree with him on that, > > but it isn't clear enough yet to me what would be better > > terminology. I didn't follow the other matter you raised > > in those posts. > > My point was that we need another level of indirection an in-Order > indexing > (along with different orders for the same data). > Hello Gennadiy, It'd be great if you can resume that discussion of ours where we left: http://lists.boost.org/Archives/boost/2006/04/103441.php This last response I sent on Apr 20th, after some inactivity days, so maybe you just didn't notice it. Best regards, Joaquín M López Muñoz Telefónica, Investigación y Desarrollo

1 0