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

Boost.Thread : catching all exceptions
by Adrian Mawbey 13 Jun '05

13 Jun '05

Hi folks. Apologies if this has been discussed frequently before. I'm puzzled as to why the boost.thread thread_proxy function catches all exceptions and then quietly ends the thread. For example:- class MyThreadFunctor { public: void operator()() const { // Do some processing... throw very_bad_exception(); // ...never get to do some processing. } }; int main(int argc, char * argv[]) { MyThreadFunctor my_functor; boost::thread my_thread(my_functor); // Do lots of processing and then join to my_thread. // ... } In boost 1.32, the exception that is thrown in the created thread is caught in the library's thread_proxy function and then the thread exits quietly. Is there a compelling reason for this behaviour? >From my point of view, it seems more flexible to not catch exceptions in the thread_proxy function. If I wanted to, I could still get the current behaviour by putting a try/catch-all block in the operator() function. I would have expected the normal C++ uncaught exception actions (i.e. calling terminate()) to occur if an exception is thrown out of the user provided thread code. I noticed this when (oh horror!) an error in my code threw an exception out of operator(). I could have identified where the problem was from the core file or Dr Watson log that would have been created if the exception had not been caught by thread_proxy(). Am I missing something obvious here? Thanks, Adrian.

1 0

Tokenizer Question
by Tom Browder 13 Jun '05

13 Jun '05

I have used my own C++ tokenizer in the past, but I would like to use Boost's instead. The predominant use of tokenizing for me is to split on white space, but Boost's default is to use white space AND punctuation. Is there any possibility to have either the default changed, or another TokenizerFunction added such as ws_separator, or something similar? I know I can use boost::char_separator<char> sep(" \n\t"); (but do I need to add "\v" to the char set?) but I would rather have something like boost::ws_separator sep; and, better, make the ws_separator be the default TokenizerFunction for tokenizer. Thanks for listening. Tom Browder

2 2

Text_archives don't support integer types?
by Marcin Tustin 13 Jun '05

13 Jun '05

I'm using the CVS version of Boost from 2005-06-06, and it seems that text archives don't support anything except strings, characters, and the like, specifically not integer types like int. I've grepped the archive includes, and it seems that there are no save methods specified for the integer types (except char) by the text archives, and I get compiler errors like: c:\D2\API\NewClient\EnhancedMessageHandler.h(370) : error C2679: binary '<<' : no operator found which takes a right-hand operand of type 'int' (or there is no acceptable conversion) (with VC7.1) It works when the RHS of << is a variable, but not when it is a function. I take it that this is because with variables, a string is automatically constructed to wrap the int types. This is very confusing, and might well be worth mentioning in the docs. Marcin Tustin IT Innovation Centre 2 Venture Road Chilworth Science Park Southampton, SO16 7NP, UK tel: +44 23 8076 0834 fax: +44 23 8076 0833 mailto:mt@it-innovation.soton.ac.uk http://www.it-innovation.soton.ac.uk

2 1

qualified name error on vc7-stlport
by Johannes Strömberg 13 Jun '05

13 Jun '05

Hi, I am trying to build boost for vc7-stlport (stlport version 4.6.2) and I get most libraries to compile but when compiling the date_time library I get errors like the following: C:\wspace\dev\ssdk\ssdk\stlport\src\stlport\stl\_iterator_base.h(98) : error C2825: '_Iterator::iterator_category': cannot form a qualified name C:\wspace\dev\ssdk\ssdk\boost\src\boost\date_time\date_parsing.hpp(252) : see reference to class template instantiation '_STL::iterator_traits<_Iterator>' being compiled with [ _Iterator=_STL::basic_string<char,_STL::char_traits<char>,_STL::allocator<char>>::iterator ] C:\wspace\dev\ssdk\ssdk\boost\src\boost\date_time\gregorian\parsers.hpp(73) : see reference to function template instantiation 'boost::date_time::period<point_rep,duration_rep> boost::date_time::from_simple_string_type<boost::gregorian::date,char>(const _STL::basic_string<_CharT,_Traits,_Alloc> &)' being compiled with [ point_rep=boost::gregorian::date, duration_rep=boost::gregorian::date_duration, _CharT=char, _Traits=_STL::char_traits<char>, _Alloc=_STL::allocator<char> ] C:\wspace\dev\ssdk\ssdk\stlport\src\stlport\stl\_iterator_base.h(98) : error C2039: 'iterator_category' : is not a member of 'operator``global namespace''' C:\wspace\dev\ssdk\ssdk\stlport\src\stlport\stl\_iterator_base.h(98) : fatal error C1507: previous user errors and subsequent error recovery halt further compilation I know it has to do with a compiler deficiency but how can I fix it? /Johannes

1 0

multi_array operator= and copy constructor have problem
by Yan Zhang 13 Jun '05

13 Jun '05

I'm testing multi_array with a simple example: typedef boost::multi_array<double, 3> array; array A(boost::extents[3][4][2]); A[0][0][0] = 3.14; array B; B = A; When execute the last statement B=A, you encounter assertion: assert(std::equal(other.shape(),other.shape()+this->num_dimensions(), this->shape())); in file multi_array_ref.hpp (line 476). I used VC++ 7.1 with debug build. Please advice.

2 1

Re: [Boost-users] How to inherit from boost::multi_array
by Yan Zhang 13 Jun '05

13 Jun '05

I'm read the posting by Thomas and replay by Ronald on how to inherit from boost::multi_array as following: On Oct 2, 2004, at 10:26 AM, Thomas Mensch wrote: Dear List, I am using boost 1.31.0 on Linux (with gcc 3.3.2) I would like to create a 2D array class, which inherits from boost::multi_array<double, 2> class. However I have problems with the copy constructor of my class Array. Below is the code, which I am trying to run. #include <boost/multi_array.hpp> class Array : public boost::multi_array<double,2> { public: Array (void) : boost::multi_array<double, 2> () {} Array (const int nrow, const int ncol) : boost::multi_array<double, 2> (boost::extents[nrow][ncol]) {} Array (const Array& other) : boost::multi_array<double, 2> (other) {} ~Array (void) {} Array& operator= (const Array& rhs) { if (&rhs != this) { boost::multi_array<double, 2>::operator= (rhs); } return *this; } }; /* class Array */ Ronald's replay: Greetings Thomas, Though you did not supply the error message from your compiler, I suspect that the problem is this: multi_array has a templated constructor of the form: template <class T> multi_array(T const&); that will match your Array type better than the constructor with the signature: multi_array(multi_array<double,2> const&); In short, overload resolution is not selecting the constructor that you desire. You'll need to coerce your Array reference to a multi_array<double,2> reference before passing it to the constructor. Hope this helps. I had the same problem as Thomas'. I don't quit understand Ronald's comment. How to you overload template <class T> multi_array(T const&) instead of multi_array(multi_array<double,2> const&)? I thought Thomas' code is overloading template <class T> multi_array(T const&). Thanks in advance, Yan

2 1

[BGL] edge_index_t
by martin f krafft 13 Jun '05

13 Jun '05

I see the edge_index_t symbol used all over the place, but I cannot find its defintion. So now I am stuck: error: `egde_index_t' undeclared in namespace `boost' What header file do I have to include to be able to access my graph's edge index property? -- martin; (greetings from the heart of the sun.) \____ echo mailto: !#^."<*>"|tr "<*> mailto:" net@madduck invalid/expired pgp subkeys? use subkeys.pgp.net as keyserver! spamtraps: madduck.bogus(a)madduck.net echo '9,J8HD,fDGG8B@?:536FC5=8@I;C5?@H5B0D@5GBIELD54DL>@8L?:5GDEJ8LDG1' |\ sed ss,s50EBsg | tr 0-M 'p.wBt SgiIlxmLhan:o,erDsduv/cyP'

2 2

bug on /libs/graph/doc/using_adjacency_list.html
by martin f krafft 13 Jun '05

13 Jun '05

In this text: NOTE: The Boost Graph Library supports two interchangeable methods for specifying interior properties: bundled properties and property lists. The former is easier to use and requires less effort, whereas the latter is compatible with older, broken compilers and is backward-compatible with Boost versions prior to 1.32.0. If you absolutely require these compatibility features, read on to learn about property lists. Otherwise, we strongly suggest that you read about the bundled properties mechanism. I really think that "former" and "latter" got swapped around. :) -- martin; (greetings from the heart of the sun.) \____ echo mailto: !#^."<*>"|tr "<*> mailto:" net@madduck invalid/expired pgp subkeys? use subkeys.pgp.net as keyserver! spamtraps: madduck.bogus(a)madduck.net "a cigarette is the perfect type of pleasure. it is exquisite, and it leaves one unsatisfied." -- oscar wilde

1 0

No << or & operator defined (a different problem?)
by Marcin Tustin 13 Jun '05

13 Jun '05

I'm getting substantially the same error as before: c:\D2\API\NewClient\EnhancedMessageHandler.h(371) : error C2679: binary '&' : no operator found which takes a right-hand operand of type 'int' (or there is no acceptable conversion) Or, alternatively binary '<<' when using the << operator, in the code: template<class Archive > static void save(Archive & ar, const IplJrsMediaTimePoint& g, const unsigned int version) { ar << ( const int )(g.GetYear()); } - But also when I use static_cast or for type uint32 when I don't have a cast. Any suggestions very gratefully received. Marcin Tustin IT Innovation Centre 2 Venture Road Chilworth Science Park Southampton, SO16 7NP, UK tel: +44 23 8076 0834 fax: +44 23 8076 0833 mailto:mt@it-innovation.soton.ac.uk http://www.it-innovation.soton.ac.uk

1 0

Re: [Boost-users] Boost.Serialization and Pointers to ObjectsofDerived Classes
by Richard Jennings 13 Jun '05

13 Jun '05

Thanks for your reply. I included the void_cast_register call because the base class is not being serialised by the derived class nor is the derived class being serialised and hence my understanding was that it was needed in this special case. It works with the paramaters to void_cast_register swapped. The include file void_cast.hpp does show the parameters as you state, however the documentation and examples have them in the order <base, derived> (special.html). Can I suggest that this is updated please? Regards, Richard -----Original Message----- From: boost-users-bounces(a)lists.boost.org [mailto:boost-users-bounces@lists.boost.org] On Behalf Of Robert Ramey Sent: 10 June 2005 16:52 To: boost-users(a)lists.boost.org Subject: [Boost-users] Re: Boost.Serialization and Pointers to ObjectsofDerived Classes void_cast_register should be necessary only in the most unusual cases. It should happen automatically as part of the "base_object" wrappers call. Also it is used void_cast_register<Derived, Base> which is the opposite you've of what you've done below. So I would remove it from your example. The export should make up for this. There are number of code paths related to borland 5.64 and borland 5.51. I don't know anything about borland 6. void_cast/export/template instantiation are areas which many compilers have exhibited unexpected behavior. So its quite possible that the code as is can't accomodate BC 6. As far as I know - we've never seen borland 6 or tested it. So if you want to use any part of boost with a new compiler whose ancestor has a history of special requirements, you should be prepared to run the test suite for that compiler. Please do this and post the results. Robert Ramey

1 0