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

Thread within thread
by Leon Mergen 14 Dec '04

14 Dec '04

Hello everyone, I'm currently attempting to create a thread within a thread, and for some reason, the constructor seems to block and things just don't work very well. Consider a class ``BackupHandler'' and a class ``BackupHandlerThread'' - the thread is created for once in the 50ms polling to an output file which might be used by other processes. Now, in BackupHandler, I do this: --- boost::thread thread ( BackupHandlerThread (this->backupFile, this->data) ); --- Which, as far as I am aware of, should first create a BackupHandlerThread object, and then "launch" the thread with this object. Now, since I print out some debug info, it seems like it reaches the BackupHandlerThread constructor once, and the destructor twice. The function BackupHandlerThread::operator() () never is reached even once, and the execution of the initial thread blocks at the creation. Also, according to gdb there is no new thread being created, and it adds any use, this is the stacktrace of the thread that should create the new thread: --- [Switching to thread 3 (Thread 16386 (LWP 24085))]#0 0x401d2b94 in __pthread_sigsuspend () from /lib/i686/libpthread.so.0 (gdb) bt #0 0x401d2b94 in __pthread_sigsuspend () from /lib/i686/libpthread.so.0 #1 0x401d29d8 in __pthread_wait_for_restart_signal () from /lib/i686/libpthread.so.0 #2 0x401d2151 in pthread_create@@GLIBC_2.1 () from /lib/i686/libpthread.so.0 #3 0x400a5fc6 in boost::thread::thread(boost::function0<void, std::allocator<boost::function_base> > const&) () from /usr/local/lib/libboost_thread-gcc-mt-1_31.so.1.31.0 #4 0x080846c2 in BackupHandler::addData(std::string) (this=0x40bd2cfc, data= {static npos = 4294967295, _M_dataplus = {<allocator<char>> = {<No data fields>}, _M_p = 0x8084f10 "U\211�\203�\030�D$\f"}, static _S_empty_rep_storage = {0, 0, 0, 0}}) at function_template.hpp:427 #5 0x0808455d in BackupHandler::addData(std::string*) (this=0xfffffffc, data=0xfffffffc) at BackupHandler.cc:36 #6 0x0805a759 in Normaliser::normalise() (this=0x80bd0d8) at Normaliser.cc:81 #7 0x08079c9d in HandleClient::operator()() (this=0x80b9d38) at HandleClient.cc:37 #8 0x08052c81 in boost::detail::function::void_function_obj_invoker0<HandleClient, void>::invoke(boost::detail::function::any_pointer) (function_obj_ptr= {obj_ptr = 0xfffffffc, const_obj_ptr = 0xfffffffc, func_ptr = 0xfffffffc, data = "�"}) at function_template.hpp:128 #9 0x400a69f4 in boost::thread_group::join_all() () from /usr/local/lib/libboost_thread-gcc-mt-1_31.so.1.31.0 #10 0x401cff60 in pthread_start_thread () from /lib/i686/libpthread.so.0 #11 0x401d00fe in pthread_start_thread_event () from /lib/i686/libpthread.so.0 #12 0x402f7327 in clone () from /lib/i686/libc.so.6 (gdb) --- Thanks in advance for any ideas! Regards, Leon Mergen

2 3

RE: [Boost-users] Re: Iterators and maps [was Re: Where should ST L suggestionsgo?]
by Foster, Gareth 14 Dec '04

14 Dec '04

"I believe that iterator theories are as central to Computer Science as theories of rings or Banach spaces are central to Mathematics." -- An Interview with A. Stepanov http://www.stlport.org/resources/StepanovUSA.html <http://www.stlport.org/resources/StepanovUSA.html> that's just the tip of the ice berg, i found that interview to be a real eye opener having been through university where only OO programming got extensive coverage. Always seemed odd to me that what is such a big thing in the C++ community seems to get such a 'bad deal'. I hear things are changing though ... </ramble> gaz -----Original Message----- From: boost-users-bounces(a)lists.boost.org [mailto:boost-users-bounces@lists.boost.org] On Behalf Of Jim Lear Sent: 14 December 2004 16:30 To: boost-users(a)lists.boost.org Subject: Re: [Boost-users] Re: Iterators and maps [was Re: Where should STL suggestionsgo?] David Abrahams wrote: Jim Lear wrote: So in the absolute most abhorently poor code example (let's just call it "meta-code"), the map operator would behave like: data_type &operator[](const iterator &i) { return i->second; } Am I just too ignorant to make any sense? :-) Maybe I'll crawl back into my hole. :-) You're making sense, but consider this: why should the map get involved in that operation at all when you can do it all with the iterator? Seems like a design flaw to get another object involved needlessly. Not only might it be a design flaw, but it may invite inefficiency. But really it's just consistency with other languages which iterate over the keys (like awk) and code clarity. Someone unfamiliar with STL would find the following obvious: for (m_type::iterator i=m.begin();...) for (mi_type::iterator j=m[i].begin();...) cout << m[i][j] << endl; But they may find for (m_type::iterator i=m.begin();...) for (mi_type::iterator j=i->second.begin();...) cout << j->second << endl; obtuse. Perhaps folks intimately familiar with iterators may find this is hard to believe. :-) An alternative question might be, why should one who uses maps and vectors also be required to be familiar with the intricacies of iterators? To me, that seems like a design flaw. I have never encountered another language that requires the user to understand iterators to the extent STL does. Ease of use and learning ought to be a priority in the design of these objects, to the extent it is efficient and safe. To use iterators in lieu of keys or indices seems to me like a great simplification for users. It completely obviates the need to understand iterators in this case (while still allowing existing iterator functionality). -- Jim Lear

1 0

Re: Re: Boost-1-32 multi_index compilation error on msvc6.
by Jia Pu 14 Dec '04

14 Dec '04

I forgot to include ordered_index.hpp. I thought including multi_index_container.hpp will give me everything I need. Thank you. > From: Joaquin M Lopez Munoz <joaquin(a)tid.es> > Subject: [Boost-users] Re: Boost-1-32 multi_index compilation error on > msvc6. > To: boost-users(a)lists.boost.org > Message-ID: <loom.20041214T083314-505(a)post.gmane.org> > Content-Type: text/plain; charset=utf-8 > > Jia Pu <very.funny.j <at> gmail.com> writes: > > > > > I am using MSVC 6 with Boost 1-32. Below is my sample code using two indeices: > > > [...] > > When I compile it, I have following errors. The actual message is much > > longer than this. I don't want paste them here to piss off every one > > on this mailing list. :) > > > > Anyway, I am not a template expert. I couldn't figure it out myself. > > Any help will be appreciated. Thanks. > > > > Hi, > > I cannot see what could be going wrong with your example. > The attached complete program compiles OK here > with MSVC 6.0 + SP5. Maybe you can post some complete code > that exhibits the problem? > > JoaquÃn M LÃ³pez MuÃ±oz > TelefÃ³nica, InvestigaciÃ³n y Desarrollo > > #include <boost/multi_index_container.hpp> > #include <boost/multi_index/key_extractors.hpp> > #include <boost/multi_index/ordered_index.hpp> > #include <string> > > struct by_id{}; > struct by_symbol{}; > > struct StringIdPair > { > StringIdPair(int _id, const std::string & _string): > id(_id), > string(_string) > {} > > int id; > std::string string; > }; > > using namespace boost::multi_index; > > typedef multi_index_container< > StringIdPair, > indexed_by< > ordered_unique< > tag<by_id>, > BOOST_MULTI_INDEX_MEMBER(StringIdPair, int, id) >, > ordered_unique< > tag<by_symbol>, > BOOST_MULTI_INDEX_MEMBER(StringIdPair, std::string, string) > > > > > StringIdTable; > > int main() > { > StringIdTable string_id_table; > > string_id_table.insert(StringIdPair(0,"zero")); > string_id_table.insert(StringIdPair(1,"one")); > string_id_table.insert(StringIdPair(2,"two")); > > return 0; > }

1 0

Re: newbie boost sample compile program
by Brian Gaber 14 Dec '04

14 Dec '04

"Martin Lemieux" < <mailto:lemieux@cae.com> lemieux<at>cae.com> wrote: > > As I suspected, those are linker errors! That means steps 1 and 2 worked. Now > (step 3), you need to specify the directory where to find the boost libraries, > as well as the lirbary file(s) for the linker to link with. At this point you > have plenty of choices: static/dynamic, multi-thread or not, debug/release. > > 3a) You might have some of the following library files in your AIX > [...]/boost/lib directory: > > boost_filesystem-AIX-mt-1_31.so > boost_filesystem-AIX-mt-1_31.lib > boost_filesystem-AIX-mt-gd-1_31.so > boost_filesystem-AIX-mt-gd-1_31.lib On AIX v5.2 here are the boost_filesystem* libraries I have under /usr/local/lib libboost_filesystem-xlc-1_32.a libboost_filesystem-xlc-1_32.so libboost_filesystem-xlc-d-1_32.a libboost_filesystem-xlc-d-1_32.so libboost_filesystem-xlc-d.a libboost_filesystem-xlc-d.so libboost_filesystem-xlc-mt-1_32.a libboost_filesystem-xlc-mt-1_32.so libboost_filesystem-xlc-mt-d-1_32.a libboost_filesystem-xlc-mt-d-1_32.so libboost_filesystem-xlc-mt-d.a libboost_filesystem-xlc-mt-d.so libboost_filesystem-xlc-mt.a libboost_filesystem-xlc-mt.so libboost_filesystem-xlc.a libboost_filesystem-xlc.so > On your compiler command line, add "-L /usr/local/include/boost-1_32/boost/lib" > (or make a symbolic link as you did for the include directory > towards /usr/local/include/boost/lib and use this one instead, as you wish). On AIX (maybe all unix's) boost libraries are found in /usr/local/lib > Then (usually at the end of the command line) add the library file you like. > May I the a static one (.lib). Your command line should look like: No boost libraries with a .lib extension > xlc++ -o simple_ls simple_ls.cpp -I /usr/local/include -L /usr/local/include/boost-1_32/boost/lib boost_filesystem-AIX-mt-1_31.lib the actual xlc++ compiler command line that worked was: xlc++ -o simple_ls simple_ls.cpp -I /usr/local/include -lboost_filesystem-xlc-1_32 -L/usr/local/lib NOTE: after the -l the three letters "lib" are removed from the front of the actual library name. > > 2b) On MSDEV6, add "c:\boost\include\boost-1_32" in your project settings. More specifically in the "C++" tab, under "Preprocessor" category (drop-down list), and at "Additional include directories:". > > > > I did this and I can now build, but I get a linker error: > > > > LINK : fatal error LNK1104: cannot open file "libboost_filesystem-vc6-sgd- > 1_32.lib" > > 3b) On MSDEV6, > > i) add "libboost_filesystem-vc6-sgd-1_32.lib" in your project settings. More specifically in the "Link" tab, under "Input" category (drop-down list), and at "Object/Library modules:". > > ii) and add "c:\boost\include\boost-1_32\boost\lib" in "Additional library path:". On my Windows 2000 the file "libboost_filesystem-vc6-sgd-1_32.lib" is found in c:\boost\lib Martin, thanks for your help. I would have never got this working without your assistane.

2 1

Bug in Boost.Range + Boost.String Algorithms
by Raul Huertas Diaz 14 Dec '04

14 Dec '04

I think I have found a bug mixing together this two libraries in Boost 1.32.0 Simply doing: #include <boost/algorithm/string.hpp> #include <boost/range.hpp> it's a bug because boost::algorithm::iterator_range<IteratorT> is redefined. I have attached a small demo program. ¿Am I wrong? ¿Is there a way to bypass this, apart from changing Boost code? Thanks in advance. P.D.: In case it is interesting, I'am working with: - Linux (Mandrake 10.0) - gcc 3.2.1 P.P.D: Boost.String Algorithms is great !!! Boost.Range is superb !! ;) -- Raúl Huertas Díaz ************************************************************************************************************************************************ *La información contenida en este mensaje de correo electrónico es confidencial y puede revestir el carácter de reservada. * *Está dirigida exclusivamente a la persona destinataria. * *El acceso o cualquier uso por parte de cualquier otra persona de este mensaje no están autorizados y pueden ser ilegales.* *Si no es Ud. la persona destinataria, le rogamos que proceda a borrarlo. * *The information in this e-mail is confidential and may be legally privileged. * *It is intended solely for the addressee. * *Access or any use by any other person to this Internet e-mail is not authorised and may be unlawful. * *If you are not the intended recipient, please delete this e-mail. * ************************************************************************************************************************************************

2 1

Antwort: Re: [Boost-users] Using zero characters in regular expessions
by Keno.Basedow＠gedis.rohde-schwarz.com 14 Dec '04

14 Dec '04

Thanks a lot! Works very fine. --- Keno Basedow GEDIS GmbH Sophienblatt 100 D-24114 Kiel Fon: +49-431-60051-24 (-0) Fax: +49-431-60051-11 basedow(a)gedis-online.de |---------+-----------------------------------> | | "John Maddock" | | | <john(a)johnmaddock.co.uk>| | | Gesendet von: | | | boost-users-bounces@list| | | s.boost.org | | | | | | | | | 14.12.2004 13:58 | | | Bitte antworten an | | | boost-users | | | | |---------+-----------------------------------> >------------------------------------------------------------------------------------------------------------------------------| | | | An: <boost-users(a)lists.boost.org> | | Kopie: | | Thema: Re: [Boost-users] Using zero characters in regular expessions | >------------------------------------------------------------------------------------------------------------------------------| > i try to use the regex library for binary data. > I tried this code: > > string h("\x81\x05\x00\x48"); > regex e2("\x81\x05\x00(.)"); > smatch m; > if (regex_match(h, m, e2)) > { > cout << "0. matched=" << (m[0].matched?"true":"false") << endl; > cout << "0. match length=" << m[0].str().size() << endl; > cout << "0. match character=" << m[0].str() << endl; > cout << "1. matched=" << (m[1].matched?"true":"false") << endl; > cout << "1. match length=" << m[1].str().size() << endl; > cout << "1. match character=" << m[1].str() << endl; > } The thing is if you pass a const char* to either basic_string or boost::regex then it will assume that the string is null-terminated, which as you have found does the wrong thing in this case. In both cases use the constructor that takes a string length as well as a pointer: string h("\x81\x05\x00\x48", 4); regex e2("\x81\x05\x00(.)", 7, regex::perl); Hope this helps, John. _______________________________________________ Boost-users mailing list Boost-users(a)lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users

1 0

Using zero characters in regular expessions
by Keno.Basedow＠gedis.rohde-schwarz.com 14 Dec '04

14 Dec '04

Hello, i try to use the regex library for binary data. I tried this code: string h("\x81\x05\x00\x48"); regex e2("\x81\x05\x00(.)"); smatch m; if (regex_match(h, m, e2)) { cout << "0. matched=" << (m[0].matched?"true":"false") << endl; cout << "0. match length=" << m[0].str().size() << endl; cout << "0. match character=" << m[0].str() << endl; cout << "1. matched=" << (m[1].matched?"true":"false") << endl; cout << "1. match length=" << m[1].str().size() << endl; cout << "1. match character=" << m[1].str() << endl; } But regex seems to have problems with the zero character because the result is: 0. matched=true 0. match length=2 0. match character=ü♣ 1. matched=false 1. match length=0 1. match character= If i change the code above to: string h("\x81\x05\x01\x48"); regex e2("\x81\x05\x01(.)"); and now the result is like intended: 0. matched=true 0. match length=4 0. match character=ü♣☺H 1. matched=true 1. match length=1 1. match character=H Technical it should be possible to work with zero characters in basic_string. Is this a bug of regex or is this technical not possible? Keno Basedow GEDIS GmbH Sophienblatt 100 D-24114 Kiel Fon: +49-431-60051-24 (-0) Fax: +49-431-60051-11 basedow(a)gedis-online.de

2 1

result_of example?
by Fedor Pikus 14 Dec '04

14 Dec '04

I was trying to come up with a solution for the following problem: say, I want to declare this class template <typename T1, typename T2> class A { public: A( T1 x, T2 y ) : xy( x + y ) {} private: The_Right_Type xy; }; Or I want to declare a function template: template <typename T1, typename T2> The_Right_Type f( T1 x, T2 y ) { return x + y; } The problem is, of course, what is The_Right_Type? I could not get a solution better than something which would require me to "register" all types in some way (then I could get types of all expressions involving registered types). Then Scott Meyers suggested I look at boost::result_of. It looks very cool except I can't figure out how to use it :( I searched the Boost mailing lists, found some incomplete examples and some broken tinyurl links, but nothing which would show me how to use it. Looking at the proposal, I could see how it is used inside a template where I already have types, but not how to get a type out of an expression in the first place. Now the question: can result_of solve the problem above, and would someone be kind enough to show me an example of how to use it? Thanks in advance, Fedor -- Fedor G Pikus (fpikus(a)gmail.com) http://www.pikus.net http://wild-light.com

2 1

Boost-1-32 multi_index compilation error on msvc6.
by Jia Pu 14 Dec '04

14 Dec '04

I am using MSVC 6 with Boost 1-32. Below is my sample code using two indeices: struct by_id{}; struct by_symbol{}; struct StringIdPair { StringIdPair(int _id, const std::string & _string): id(_id), string(_string) {} int id; std::string string; }; typedef multi_index_container< StringIdPair, indexed_by< ordered_unique<tag<by_id>, BOOST_MULTI_INDEX_MEMBER(StringIdPair, int, id) >, ordered_unique<tag<by_symbol>, BOOST_MULTI_INDEX_MEMBER(StringIdPair, std::string, string) > > > StringIdTable; StringIdTable string_id_table; When I compile it, I have following errors. The actual message is much longer than this. I don't want paste them here to piss off every one on this mailing list. :) Anyway, I am not a template expert. I couldn't figure it out myself. Any help will be appreciated. Thanks. c:\boost\include\boost_1_32\boost\multi_index\detail\msvc_index_specifier.hpp(45) : error C2039: 'node_class' : is not a member of 'fake_index_type<0>' c:\boost\include\boost_1_32\boost\multi_index\detail\node_type.hpp(42) : see reference to class template instantiation 'boost::multi_index::detail::msvc_index_specifier<struct boost::multi_index::ordered_unique<struct boost::multi_index::tag <struct by_symbol,struct boost::mpl::na,struct boost::mpl::na>,struct boost::multi_index::member_offset<struct StringIdPair,class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> >,4>,struct boost::multi_index::detail: :null_arg> >::result_node_class<struct boost::multi_index::detail::index_node_base<struct StringIdPair> >' being compiled c:\boost\include\boost_1_32\boost\mpl\aux_\preprocessed\msvc60\apply_wrap.hpp(102) : see reference to class template instantiation 'boost::multi_index::detail::index_node_applier::apply<struct boost::mpl::v_iter<struct boost::mpl::vector2<st ruct boost::multi_index::ordered_unique<struct boost::multi_index::tag<struct by_id,struct boost::mpl::na,struct boost::mpl::na>,struct boost::multi_index::member_offset<struct StringIdPair,int,0>,struct boost::multi_index::detail::null_arg>,struct boost::multi_index::ordered_unique<struct boost::multi_index::tag<struct by_symbol,struct boost::mpl::na,struct boost::mpl::na>,struct boost::multi_index::member_offset<struct StringIdPair,class std::basic_string<char,struct std::char_traits<char>,c lass std::allocator<char> >,4>,struct boost::multi_index::detail::null_arg> >,1>,struct boost::multi_index::detail::index_node_base<struct StringIdPair> >' being compiled

2 1

boost::threads and dual processor Macs
by Stephen Carson 13 Dec '04

13 Dec '04

We are having a problem that we have narrowed down to a problem with boost::threads on dual processor Macs. The very same software works fine, with boost::threads, on a single processor Mac. It also works fine with native Mac threads on a dual processor Mac. The crash is repeatable though, as threaded programs are wont to do, it crashes in slightly different places each time. The program does seem to crash during memory allocation. We have CodeWarrior 9 on its way to see if a compiler upgrade might address the problem, (there are some known problems with CodeWarrior 8 and threads, but only in debug builds. We see our crash in release builds as well). Our questions for the boost crew: -Are there any known issues with boost::threads and multi-processor machines? -Has boost::threads been tested on multi-processor machines? On multi-processor Macs? Details: Compiler: CodeWarrior 8.3 Computer: Dual processor 1.8Ghz Macintosh G5 OS: MacOS X 10.3.6 Boost: v1.31 -- Stephen W. Carson, Software Engineer DirXion Software in St. Louis, MO (636) 717-2367 "Premature optimization is the root of all evil." -Donald Knuth

1 0