August 2006 - Boost-users - lists.preview.boost.org

shared_ptr segfaults on 64bit AMD
by Paul Davis 27 Aug '06

27 Aug '06

Howdy, I've come across an odd segfault that originates in some of the boost code. (By originate, meaning thats where the stack trace points, I'm not sure if its me or boost thats wrong) Anyway, the weird part is that its only on my 64 bit machine. I took a look at where its segfaulting in boost::detail::atomic_exchange_and_add(). Its scary inline assembly stuff. Well, mostly I just don't know assembly so I haven't the slightest idea if its right or wrong. And obviously, its a platform specific header and what not so I imagine its limited to this area. I have had other weird segfaults that come and go from the atomic_* set of methods. I can't pin down exactly whats causing it. They mostly seem to be coming from storing shared_ptr's in STL containers. I've never had any problems with it before so I'm assuming its just a relatively untested section of code. And if its relevant, I also tried downloading CVS and running the shared_ptr_test. It ran without segfaulting just fine. Some relevant info: Segfaults on: 2.6.15-25-amd64-generic #1 SMP PREEMPT Wed Jun 14 11:28:03 UTC 2006 x86_64 GNU/Linux g++ (GCC) 4.0.3 (Ubuntu 4.0.3-1ubuntu5) boost/version.hpp #define BOOST_VERSION 103301 Does NOT segfault on: 2.6.15-1.1831_FC4smp #1 SMP Tue Feb 7 13:48:31 EST 2006 i686 i686 i386 GNU/Linux g++ (GCC) 4.0.2 20051125 (Red Hat 4.0.2-8) boost/version.hpp : #define BOOST_VERSION 103200 Stack trace: Program received signal SIGSEGV, Segmentation fault. [Switching to Thread 46912529528592 (LWP 27560)] 0x000000000040a2f8 in boost::detail::atomic_exchange_and_add (pw=0x98, dv=-1) at sp_counted_base_gcc_x86.hpp:50 50 ); (gdb) bt #0 0x000000000040a2f8 in boost::detail::atomic_exchange_and_add (pw=0x98, dv=-1) at sp_counted_base_gcc_x86.hpp:50 #1 0x000000000040a46c in boost::detail::sp_counted_base::release (this=0x90) at sp_counted_base_gcc_x86.hpp:143 #2 0x000000000040a4d0 in ~shared_count (this=0x2514218) at shared_count.hpp:159 #3 0x00002aaaaccb849b in ~shared_ptr (this=0x2514210) at shared_ptr.hpp:106 #4 0x00002aaaaccb88e1 in __gnu_cxx::new_allocator<boost::shared_ptr<min_heap> >::destroy (this=0x7fffffc2b410, __p=0x2514210) at new_allocator.h:107 #5 0x00002aaaaccb9f66 in std::vector<boost::shared_ptr<min_heap>, std::allocator<boost::shared_ptr<min_heap> > >::erase ( this=0x7fffffc2b410, __position={_M_current = 0x2514220}) at vector.tcc:115 #6 0x00002aaaaccb7add in label_map::CloseToAlive (this=0x7fffffc2b3e0, p=@0x7fffffc2b1a0) at build/Modules/Marching/label_map.cc:273 #7 0x00002aaaaccb326f in fast_march::march (this=0x7fffffc2b3c0, mask=@0x527d30, gradient=@0x527d10, labeled=@0x527d40) at build/Modules/Marching/fast_march.cc:76 #8 0x00002aaaaccbec64 in marching::run (this=0x527c50, cli=@0x7fffffc2b5e0, options=@0x7fffffc2b5d0, args=@0x7fffffc2b5c0) at build/Modules/Marching/marching.cc:110 #9 0x00000000004074bd in process (args=@0x7fffffc2c460) at build/DDS/dds.cc:259 #10 0x0000000000408ffd in main (argc=2, argv=0x7fffffc2c5a8) at build/DDS/dds.cc:531 Paul Davis

4 7

Re: [Boost-users] on-line documentation, boost 1.34, etc.
by JOAQUIN LOPEZ MU?Z 27 Aug '06

27 Aug '06

Andriy Tylychko ha escrito: > Hi. > > Several question, slightly unrelated: > > 1) To read the last version of documentation I go to > http://www.boost.org/doc/html/index.html. But this list > counts only 15 libraries. Next, going to > http://www.boost.org/doc/html/libraries.html, we see the > same list of 15 libraries with sub-links, but just after > that "Libraries Listed Alphabetically" contains already > much more libraries. Just a bug? The page http://www.boost.org/doc/html/index.html contains those Boost libs which are documented using a tool called BoostBook. Not all Boost libs are documented with BoostBook; the complete list can be obtained at: http://boost.org/libs/libraries.htm which is where you get to from the main Boost page link "Documentation". As to http://www.boost.org/doc/html/libraries.html, this BoostBook-produced page attempts to list every library in Boost, be it documented with BoostBook or not, but in fact it's missing some entries, so you better rely on the link I give above. I don't know if this lack of some entries must e categorized as a bug or as an unmaintained spot to be removed. > > 2) Where can I see documentation of "coming soon" boost > 1.34? Especially documentation for new libraries... I > tried to find it in CVS, but in > http://boost.cvs.sourceforge.net/boost/boost/doc/html/ I > didn't find even "index.html". Using the CVS web browser, go to libs/libraries.htm under the branch RC_1_34_0: http://boost.cvs.sourceforge.net/boost/boost/libs/libraries.htm? view=log&pathrev=RC_1_34_0 There you can find the complete list of all libraries coming into Boost 1.34. If you're interested in new libraries, go to the main index.htm under RC_1_34_0: http://boost.cvs.sourceforge.net/boost/boost/index.htm? view=log&pathrev=RC_1_34_0 Also, Boost Consulting kindly hosts the entire Boost site in its current CVS form at http://boost-consulting.com/boost/ It is the main trunk of the CVS that is published here, which strictly speaking does not have to be the same as RC_1_34_0 (for instance, it may contain material scheduled for 1.35) for in general differences are few. > Why some new libraries > already have nice documentation, but this documentation > cannot be found in logical (?) locations (ASIO e.g., > http://boost.cvs.sourceforge.net/boost/boost/libs/asio/doc/ > is empty)? ASIO has not yet been integrated into Boost. I guess it'll be by Boost 1.35. > > 3) What is the easiest way (better "what is the right > place") to investigate with libraries (and their docs) that > are planned for inclusion to boost 1.34? Hopefully answered above. > > Thank you in advance! > > Best regards, > Andriy Tylychko, > telia(a)vichnavich.com > www.vichnavich.com Joaquín M López Muñoz Telefónica, Investigación y Desarrollo

1 0

on-line documentation, boost 1.34, etc.
by Andriy Tylychko (mail.ru) 27 Aug '06

27 Aug '06

Hi. Several question, slightly unrelated: 1) To read the last version of documentation I go to <http://www.boost.org/doc/html/index.html> http://www.boost.org/doc/html/index.html. But this list counts only 15 libraries. Next, going to <http://www.boost.org/doc/html/libraries.html> http://www.boost.org/doc/html/libraries.html, we see the same list of 15 libraries with sub-links, but just after that "Libraries Listed Alphabetically" contains already much more libraries. Just a bug? 2) Where can I see documentation of "coming soon" boost 1.34? Especially documentation for new libraries... I tried to find it in CVS, but in http://boost.cvs.sourceforge.net/boost/boost/doc/html/ I didn't find even "index.html". Why some new libraries already have nice documentation, but this documentation cannot be found in logical (?) locations (ASIO e.g., http://boost.cvs.sourceforge.net/boost/boost/libs/asio/doc/ is empty)? 3) What is the easiest way (better "what is the right place") to investigate with libraries (and their docs) that are planned for inclusion to boost 1.34? Thank you in advance! Best regards, Andriy Tylychko, telia(a)vichnavich.com www.vichnavich.com <http://www.vichnavich.com/>

1 0

Re: [Boost-users] [Boost.Test] How to handle std::valarray<>w/BOOST_CHECK or BOOST_CHECK_EQUAL
by Marc Viala 26 Aug '06

26 Aug '06

After few days of vacation, I've got your reply... Thanks for your response... Marc Viala ________________________________ De : boost-users-bounces(a)lists.boost.org [mailto:boost-users-bounces@lists.boost.org] De la part de Gennadiy Rozental Envoyé : vendredi 18 août 2006 06:17 À : boost-users(a)lists.boost.org Objet : Re: [Boost-users] [Boost.Test] How to handle std::valarray<>w/BOOST_CHECK or BOOST_CHECK_EQUAL operator==( std::valarray<double> const&, std::valarray<double> const& ) returns std::valarray<bool> which is not bool convertible value You will either have to use by element comparison or write your own predicate that compare the valarrays BOOS_CHECK_PREDICATE( my_valarray_comp, (v1)(v2) ); Gennadiy "Marc Viala" <mviala(a)acticm.com> wrote in message news:D0D608FD34E14C4BA93A30AC162651E11F3B97@SERVEUR2003.acticm.com... Hello, I am using the Boost.Test framework, but I not able to use std::valarray<> with BOOST_CHECK or BOOST_CHECK_EQUAL macro. For example: <\code> static void print_valarray(std::ostream& os, std::valarray<double> const& v) { // TODO } namespace std { std::ostream& operator<< (std::ostream& os, std::valarray<double> const& v) { print_valarray(os, v) ; return os ; } } ; std::valarray<double> v(10) ; BOOST_CHECK_EQUAL(v, v); <\code> Failed with the following compilation error: c:\Dev\MViala\ADV\Library\boost-1_33_1\boost\test\predicate_result.hpp(51): error C2664: 'boost::test_tools::predicate_result::readonly_property57::readonly_property57(const boost::test_tools::predicate_result::readonly_property57::write_param_t)' : impossible de convertir le paramètre 1 de 'std::_Boolarray' en 'const boost::test_tools::predicate_result::readonly_property57::write_param_t' Configuration: Platform: Win32 Compiler: Microsoft Visual C++ version 7.1 STL : Dinkumware standard library version 313 Boost : 1.33.1 Any help will be appreciated... Marc Viala ________________________________ _______________________________________________ Boost-users mailing list Boost-users(a)lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users

1 0

[mpl] how to create a map using an algorithm
by Yuval Ronen 26 Aug '06

26 Aug '06

Lets say I have several types t1, t2, t3, and so on. Each of them has a typedef x within it. I want to be able to get a type y, and find the type t which holds is_same<y, t::x>. I thought of doing it using mpl::map< pair<t1::x, t1>, pair<t2::x, t2>, ...>. The problem is how to create the map. I can create it manually, just like in the previous sentence, but I hope there's a better way. Is there some transform algorithm that I can pass it an mpl::set<t1, t2, ...> and a meta-function that yields t::x for any given t, and this algorithm would produce the desired map? Thanks, Yuval

2 3

[boost::function_types] Create function type from member function type
by Peter Soetens 25 Aug '06

25 Aug '06

Hi, I was looking for a way to convert a member function type: bool (X::*)(int) or bool (X::)(int) to the 'C' equivalent, without the class pointer: bool (*)(int) or bool(int) Now, I used this helper class: // Converts R (X::)(Args) to R (Args) template<class F> class UnMember { typedef boost::function_type_signature<F> member_signature; typedef typename boost::mpl::erase<typename member_signature::types, typename boost::mpl::next<typename boost::mpl::begin<member_signature>::type>::type>::type non_member_signature; public: typedef typename boost::function_type<boost::plain_function,non_member_signature>::type type; }; And use it like: typedef UnMember<bool(X::)(int)>::type PlainFunctionType; It gets even complexer when one wants to convert a member function type to a function type which takes the member as first argument: // Converts R (X::)(Args) to R (X::*, Args) template<class F> class MemberToFirstArg { typedef boost::function_type_signature<F> member_signature; // remove the class from the arg list: typedef typename boost::mpl::erase<typename member_signature::types, typename boost::mpl::next<typename boost::mpl::begin<member_signature>::type>::type>::type non_member_signature; // insert it as first argument. typedef typename boost::mpl::insert<non_member_signature, typename boost::mpl::next<typename boost::mpl::begin<non_member_signature>::type>::type, typename boost::add_pointer<typename boost::mpl::at<typename member_signature::types,boost::mpl::int_<1> >::type>::type >::type arg_signature; public: typedef typename boost::function_type<boost::plain_function,arg_signature>::type type; }; Is there an easier way to do this ? Does a boost library have a mechanism to convert member function types to non-member function types and vice versa ? Peter -- Peter Soetens -- FMTC -- <http://www.fmtc.be>

3 8

Re: [Boost-users] boost.test and Cruise Control
by Ferentinos, Jerry 25 Aug '06

25 Aug '06

Matt Schuckmann wrote: > Up to using NAnt that sounds pretty much like what I want to do. Right > now I'm pretty much just having CC.NET do the build via the Dev Studio > command line interface. We are still on .NET 2003 right now but the 2005 > disks should be here next week. > What does using NAnt buy you that CC.NET doesn't have? > Do you have CC.NET do anything with the results of the unit tests, I.E > have it display the tests/test_suites run, have it display errors and > maybe even notify when people when test fail? I know that Boost.Test can > produce xml output but I don't know how to begin to have CC.NET consume it. The benefit of NAnt is that it is able to launch build of Visual Studio solutions, there's no having to mess with command lines and the like. Another benefit is NAnt integration with our version control system (ClearCase). Basically we have CruiseControl observe the ClearCase integration stream, issuing NAnt builds when new deliveries appear in the stream. NAnt is used to control the whole build process as well as running the tests and reacting to their results. Using NAnt we're able to perform many of the administrative tasks involved in version control and configuration management as well as promote the baselines as they pass through our build/smoke/integration tests. For CC.NET to consume Boost.Test output you will have to write an XSL transform to convert Boost.Test XML output to a format recognized by CC.NET. Jerry Ferentinos Basic Research Informatics Merck Frosst Center for Therapeutic Research www.merckfrosstlab.ca ------------------------------------------------------------------------------ Notice: This e-mail message, together with any attachments, contains information of Merck & Co., Inc. (One Merck Drive, Whitehouse Station, New Jersey, USA 08889), and/or its affiliates (which may be known outside the United States as Merck Frosst, Merck Sharp & Dohme or MSD and in Japan, as Banyu - direct contact information for affiliates is available at http://www.merck.com/contact/contacts.html) that may be confidential, proprietary copyrighted and/or legally privileged. It is intended solely for the use of the individual or entity named on this message. If you are not the intended recipient, and have received this message in error, please notify us immediately by reply e-mail and then delete it from your system. ------------------------------------------------------------------------------

1 0

scoped_ptr and swap question?
by Michael Nicolella 25 Aug '06

25 Aug '06

Looking at boost::scoped_ptr, as I understand it it's meant to model sole-ownership without transfer-of-ownership semantics. If that's the case, what is the utility for having it support swap() operations ? What functionality is lost if the swap() capability is removed? Thanks, Mike

2 1

Re: [Boost-users] Boost serialization : Stream error exception
by Nat Goodspeed 25 Aug '06

25 Aug '06

> -----Original Message----- > From: boost-users-bounces(a)lists.boost.org [mailto:boost-users- > bounces(a)lists.boost.org] On Behalf Of smith smithson > Sent: Friday, August 25, 2006 10:43 AM > To: boost-users(a)lists.boost.org > Subject: Re: [Boost-users] Boost serialization : Stream error exception > > I figured out what the problem is; I use a buffered > version of iostream/stream_buf to buffer data from my > mmap'd file and pass it out when requested. For some > reason if the very first character in the buffer is > character 255 (an extended ascii character) in any of > the buffers, I get the stream error exception. Any > suggestions why this might happen? Is this reserved in > some way? I see the issue in both text mode and binary > mode. [Nat] Do you have access to the source code to your buffered version of stream_buf? Is the buffer declared as an array of char? Does char default to signed with your compiler? If so, char 255 == 0xFF can get returned as the int value -1 -- which is still used as EOF in the stream code I've poked through. Sorry if this is way off base, but I've been bitten before by not using unsigned char for buffers containing bytes with high-order bits set.

1 0

Boost serialization : Stream error exception
by smith smithson 25 Aug '06

25 Aug '06

I'm trying to use boost serialization to serialize/deserialize data to and from a mmap'd file. I have my own ostream/istream classes that essentially read/write bytes from a mmap'd file. The process works fine except on some rare occasions (With different objects/number of objects), in which case boost throws a "stream error" exception. Any ideas what could trigger this? A search of the Boost archives indicated other people had the same problem, where the solution was "use binary mode when you open the file"... but I'm not using a file on disk, I'm using a mmap'd file. (As a side note, the mmap file is large enough to hold all my data and does not need to be expanded or remmap'd etc) The OS i'm using is Linux. Thanks for any suggestions... __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com

4 4