February 2010 - Boost-users

[proto] defining function call operators on expressions
by Manjunath Kudlur 02 Mar '10

02 Mar '10

I am trying to define operator()(...) for my expressions. I want to be able to form expression of the form Program_(a,b) and call Program_(a, b)(10,20). Here, 'a' and 'b' are terminals and I have a mechanism shown in the code below that somehow gets the type suitable for the parameter position. Suppose 'a' is terminal<Var<int> >, I want the first parameter position in Program_(a,b) (pos1, pos2) to be of the type 'int'. typename boost::result_of<call_param_type<M>(const Expr &)>::type gives me that type for the Mth argument in the expression of the form Program_(a1,a2,...). Now, I am trying to use this to define operator()(...) in my program_expr class. That doesn't seem to work. However, if I define a separate function callfunc, that takes the expression and the arguments, that seems to work. In the following code, enabling the if 0'ed out code results in compilation errors. How do I go about defining operator()(...) in my program_expr class with the above property? Thanks, Manjunath #include <boost/proto/proto.hpp> #include <boost/proto/proto_typeof.hpp> #include <boost/mpl/bool.hpp> #include <boost/mpl/assert.hpp> #include <iostream> #include <sstream> #include <string> namespace proto=boost::proto; namespace mpl=boost::mpl; unsigned int ids; template<typename VT> struct Var { unsigned int id; Var() { id = ++ids; } }; typedef proto::terminal<Var<int> >::type int32_; typedef proto::terminal<Var<float> >::type float_; struct program_ {}; template<class E> struct program_expr; struct program_domain : proto::domain<proto::pod_generator<program_expr> > { }; struct _var_type : proto::callable { template<typename Sig> struct result; template<typename This, typename T> struct result<This(const Var<T> &)> { typedef T type; }; }; struct var_type : proto::or_< proto::when<proto::terminal<Var<proto::_> >, _var_type(proto::_value)> > {}; template<int N> struct call_param_type : proto::or_< proto::when< proto::function< proto::terminal<program_>, proto::vararg<proto::terminal<Var<proto::_> > > >, var_type(proto::_child_c<N+1>)> > {}; template <typename Expr> struct program_expr { BOOST_PROTO_BASIC_EXTENDS(Expr, program_expr<Expr>, program_domain); BOOST_PROTO_EXTENDS_SUBSCRIPT(); typedef void result_type; #if 0 result_type operator()(typename boost::result_of<call_param_type<0>(const Expr &)>::type x) { std::cout << "program with one arg\n"; } result_type operator()(typename boost::result_of<call_param_type<0>(const Expr &)>::type x, typename boost::result_of<call_param_type<1>(const Expr &)>::type y) { std::cout << "program with two args\n"; } #endif }; template<typename Expr> void callfunc(const Expr &e, typename boost::result_of<call_param_type<0>(const Expr &)>::type x) { std::cout << "program with one arg\n"; } template<typename Expr> void callfunc(const Expr &e, typename boost::result_of<call_param_type<0>(const Expr &)>::type x, typename boost::result_of<call_param_type<1>(const Expr &)>::type y) { std::cout << "program with two args\n"; } template<typename A0> typename proto::result_of::make_expr<proto::tag::function , program_domain , program_ , A0 const &>::type const Program_(A0 const &a0) { return proto::make_expr<proto::tag::function, program_domain>( program_(), boost::ref(a0)); } template<typename A0, typename A1> typename proto::result_of::make_expr<proto::tag::function , program_domain , program_ , A0 const & , A1 const &>::type const Program_(A0 const &a0, A1 const &a1) { return proto::make_expr<proto::tag::function, program_domain>( program_(), boost::ref(a0), boost::ref(a1)); } int main() { int32_ a,b,c; callfunc(Program_(a, b), 10, 20); #if 0 Program_(a, b)(10, 20); #endif }

3 8

Boost.Test 1.42 --report_sink="path with spaces"
by Chris Hoeppler 01 Mar '10

01 Mar '10

Hi, when calling some my_unit_test module with my_unit_test --report_level=detailed --report_sink="foo bar" then the report isn't written to a file "foo bar" but to a file "foo". Is there any way to pass a path containing spaces to the report_sink option? Regards Chris

2 1

[GIL] Status of Ticket #2831?
by Olivier Tournaire 01 Mar '10

01 Mar '10

Hi, A colleague of mine opened a ticket more than a year ago ( https://svn.boost.org/trac/boost/ticket/2831) for a GIL defect. In fact, the custom color converter for color_converted_view on any_view is not plugged. The problem is also described in the title of the ticket. What is the status of this ticket? The changes are very simple, and are still not done in 1.42 version. Instead of: template <typename DstP, typename Result> struct color_converted_view_fn { typedef Result result_type; template <typename View> result_type operator()(const View& src) const { return result_type(color_converted_view<DstP>(src)); } }; There should be: template <typename DstP, typename Result, typename CC = default_color_converter> struct color_converted_view_fn { typedef Result result_type; color_converted_view_fn(CC cc = CC()): _cc(cc) {} template <typename View> result_type operator()(const View& src) const { return result_type(color_converted_view<DstP>(src, _cc)); } private: CC _cc; }; Best regards. Olivier

2 3

[regex] wildcard matching byte not character
by Richard Clokie 01 Mar '10

01 Mar '10

Hi all I'm having trouble with the behaviour of the wildcard character when using boost regex and unicode strings. I would expect a . to match a character, not a byte, but that's not the behaviour I'm seeing. I would have thought one wildcard would match any previous character, but for multi-byte characters in UTF-8 I have to use multiple wildcards to match them. I would appreciate it if someone could explain whether this is expected behaviour or not, or if there are flags that control this. What I'm trying to accomplish is to match a pattern (in UTF-8 ) against a string (in UTF-8). I'm creating icu UnicodeStrings since I'm having other problems with straight UTF-8 char*s and my platform doesn't support w_chars. I can show examples of the non-UnicodeString problems if desired. I'm using 1.42 Test program follows - output is: $ g++ regex2.cc -l icui18n -l icuuc -l icudata -lboost_regex -o example && ./example unicodeString tests failed Success! ---- #include <iostream> #include <boost/regex.hpp> #include <boost/regex/icu.hpp> using namespace boost; using namespace std; int main(){ static const char input[]={0xC2,0xA3, 0xC3,0x98, 0xC2,0xB2, 0 }; UnicodeString uInput(input); const char match1[] = {0xC2,0xA3,0x2E,0xC2,0xB2, 0} ; // one . const char match2[] = {0xC2,0xA3,0x2E,0x2E,0xC2,0xB2, 0} ; // two .s UnicodeString uMatch1(match1); UnicodeString uMatch2(match2); u16match what; cout << "unicodeString tests" << endl; if(u32regex_search( uInput , what, // one . fails make_u32regex(uMatch1,regex::extended))) { cout << "Success!" << endl; } else { cout << "failed" << endl; } if(u32regex_search( uInput , what, // two . succeeds make_u32regex(uMatch2,regex::extended))) { cout << "Success!" << endl; } else { cout << "failed" << endl; } }

2 2

[GIL new_io] Some questions about TIFF handling
by Olivier Tournaire 01 Mar '10

01 Mar '10

Hi all (and in particular Christian), I have some questions about TIFF reader/writer in GIL new_io. First, I do not understand why *bits32f* are scoped in the range [0.f ... 1.f]. It is for our applications a real limitation because we use float TIFF images not scoped in the range. We thus have to define a new type (e.g. bits32F) which is not scoped and simply typedef float bits32F. In addition, why double is not "typedefed" by default? We also have to define a new type: typedef double bits64f. Double however seems to be supported in the writer. Last, but not least, there is no support for *tiled TIFF*. Nothing about tiled TIFF is defined in tiff_tags.hpp. Maybe we could add in tiff_device_base a read_tile method, similar to read_scaline (sic), but which simply swap TIFFReadScanline with TIFFReadTile. So, do you plan to add this stuffs in the new version of GIL io? Maybe I could help. Let me know. Best regards, Olivier

3 2

[fusion] clarification on stateful unary function objects
by Matthias Vallentin 01 Mar '10

01 Mar '10

Hi there, While trying to use a stateful unary function objects, I stumbled over and old discussion from 2007 [1]. From the current Fusion documentation, I understand that the functor passed to the for_each algorithm must satisfy the Regular Callable Object concept. I presume that regularity, which I understand as replacing inputs with equal objects results in equal outputs, here implies that operator()(T& x) must be const. In other words, mutable function objects are not supported. As pointed out in [1], it is of course possible to bypass this restriction by defining the state that needs to changed as mutable. I was wondering if this is still the correct way to deal with mutable function objects, as illustrated in the small example below: #include <iostream> #include <string> #include <boost/bind.hpp> #include <boost/fusion/include/algorithm.hpp> #include <boost/fusion/include/container.hpp> struct container { struct printer { template <typename T> void operator()(T& x) const { std::cout << ++i << ": " << x << std::endl; } mutable unsigned i; }; template <typename T> void write(const T& row) { boost::fusion::for_each(row, printer()); } }; int main() { int i = 42; unsigned j = 7; std::string s = "foo"; container c; c.write(boost::fusion::vector_tie(i, j, s)); return 0; } Matthias [1] http://lists.boost.org/boost-users/2007/03/26338.php -- Matthias Vallentin vallentin(a)icsi.berkeley.edu http://www.icir.org/matthias

2 2

[Exceptions] Custom type in error_info value?
by Andrew Chinkoff 01 Mar '10

01 Mar '10

This is the test program I would like to be explained to me: #include <iostream> #include <string> #include <boost/exception.hpp> struct MyException: virtual boost::exception{ }; class MyClass { char* s; public: MyClass(const char* msg) : s(strdup(msg)){ } MyClass(const MyClass& c){ s = strdup(c.s); } ~MyClass(){ free(s); } }; void main() { MyException ex; ex << boost::error_info<struct my_tag1, MyClass>("qweqwe"); ex << boost::error_info<struct my_tag2, std::string>("asdasd"); std::cout << boost::diagnostic_information(ex); } This is the program output: > Throw in function (unknown) > Dynamic exception type: struct MyException > [PAUmy_tag1@?1?main@] = [ type: ?AV?$error_info@Umy_tag1@?1?main@VMyClass @@@boo > t@@, size: 8, dump: 50 69 10 01 b8 18 7f 00 ] > [PAUmy_tag2@?1?main@] = asdasd So, my question: what should I do with 'MyClass' to see the output like this: > Throw in function (unknown) > Dynamic exception type: struct MyException > [PAUmy_tag2@?1?main@] = qweqwe (Wow !!! Terrible dump is replaced with readable string here !!!???) > [PAUmy_tag2@?1?main@] = asdasd Thanks, Andrew.

2 2

boost test: How to make boost test a friend ?
by Avi Bahra 01 Mar '10

01 Mar '10

I am looking into boost test. I dont not want to expose some member functions as public. However I need to access them for testing, is there any way making boost test a friend of this class ? Any help appreciated. Best regards, Ta, Avi

3 2

asio guidance
by Paul Dugas 28 Feb '10

28 Feb '10

I'm writing code to implement an existing communication protocol that uses a packet with a fixed header and variable payload (depending on a header field). I started using a previously recorded file and fstream::read() calls; i.e. read(buf,1) looking for the start-of-packet, read(buf, HEADER_LEN), extract the payload size and read the remainder. I am now looking to hook up the receiver logic to the actual gear and need to handle timeouts and such. I'm new to asio but it looks like what I may need given that I have to support device connections via serial ports and sockets but I also need to retain the ability to read from files. I'm wondering if anyone could suggest an approach to support all three comm schemes with the same packet reception logic. The idea would be for the receive routine to return message pointers or throw on timeouts or partial packets Paul

3 3

[lambda]
by Marcin Zalewski 28 Feb '10

28 Feb '10

I am trying to use lambda's constructor [1] with fusion's fused [2]. Fused requires that the function object passed to it has a following valid expression: boost::result_of< F(T1 ...TN) >::type Unfortunately, constructor does not support result_of. Its implementation contains the following piece of code: template <class U> struct sig { typedef T type; }; This happens to be exactly what's necessary for constructor to support result_of, but sig would have to be renamed to result. I guess I do not really have a question here but rather a request to make constructor compatible with result_of. There seems to be no reason not to do it, and I could not figure out a much nicer way to construct objects with fusion sequences than combining lambda's constructor and fusion's fused. It comes so close, but if only sig was result... Cheers, Marcin [1] http://www.boost.org/doc/libs/1_42_0/doc/html/lambda/le_in_details.html#lam… [2] http://www.boost.org/doc/libs/1_42_0/libs/fusion/doc/html/fusion/functional…

4 9