- Boost-users - lists.boost.org

Re: [Boost-users] Test Tools Linking Problem
by Merrill Cornish 29 Jul '05

29 Jul '05

John, <sound of head banging on the keyboard> Yes, I knew the order was important, but I remembered the wrong order. :-( Thanks, Merrill

1 0

Re: [Boost-users] Test Tools Linking Problem
by Merrill Cornish 29 Jul '05

29 Jul '05

Gennadiy, I re-built the libs just as you said. I then copied the newly built static library to the stage/lib directory where I have all of the others. Unfortunately, I still got the same undefined. HOWEVER!... This is the first time I've noticed the boost_1_32_0/bin directory. I tried copying test_tools.obj into the directory with my own object files, added test_tools.obj to the list of my own object files in my make, and the link WORKED! To me, that says the Boost code if OK, but I've messed up makefile. (It's been many years since I've had to do battle with a raw makefile, and it wasn't GCC's make then.) I've appended an excerpt from my makefile. The "test" target is the one that gets the undefined referenced UNLESS test_tools.obj is included in the ${testobjs} list. (Yes, there is a TAB before ${CXX}.) What's wrong with the makefile? Merrill ---------------------------------------------------------------- basedir = E:/DigS boostdir = E:/boost_1_32_0 boostlibdir = ${boostdir}/stage/lib boostlibtype = -mgw-mt-1_32 boostlibs = ${boostlibdir}/boost_regex${boostlibtype}.dll boosttestlibs = ${boostlibdir}/libboost_unit_test_framework${boostlibtype}.lib . . . test : ${testobjs} ${CXX} -o test ${boosttestlibs} ${boostlibs} test.o ${testobjs}

3 2

bosst::bind problem
by Oleg Smolsky 28 Jul '05

28 Jul '05

Hi all, I've just started using boost::bind and found the following problem: class Handler { public: bool Test1(std::string sComponent); bool Test2(std::string sComponent, std::string sAddress); }; void Test() { std::vector<Handler> v; // This compiles std::string a, b; std::for_each(v.begin(), v.end(), boost::bind(&Handler::Test1, _1, a)); // This doesn't compile //std::for_each(v.begin(), v.end(), // boost::bind(&Handler::Test2, _1, _2, a, b)); } Environment: boost 1.32, VC8, WinXP. Any ideas? Thanks, Oleg.

3 6

Branch and Bound alogorithm
by Abhijit Deshmukh 28 Jul '05

28 Jul '05

Hi everyone, I'm a new user of BGL. So can anyone please tell me whether any direct Branch and Bound algorithm is already implemented in BGL? I haven't been able to find any. Basically I have to modify DFS using branch and bound scheme to avoid unnecessary computations. I think that the DFS Visitor interface can be used to do that. Is this the best way to proceed? Thanks in advance, with regards, Abhijit

2 1

[Boost Lambda] Simple problem using argument as object: part III
by Thaddeus Olczyk 28 Jul '05

28 Jul '05

Ok. I'm still not getting it. Sorry for the inconvenience. Another piece of sample code that doesn't compile. Thanks. #include <boost/lambda/lambda.hpp> #include <boost/lambda/bind.hpp> #include <boost/lambda/if.hpp> #include <vector> #include <set> #include <string> #include <iostream> using boost::lambda::_1; using boost::lambda::bind; class A { std::string s; public: A(std::string sIn):s(sIn){} void print_A() { std::cout<<"A="<<s<<std::endl; } bool isLong() { return s.length()>5; } std::string get(){return s;} }; typedef std::vector<A *> AList; typedef std::set<std::string> StrList; int main() { std::string in; AList alist; StrList excluded; while(in!="end") { in=""; std::getline(std::cin,in); alist.push_back(new A(in)); } std::for_each(alist.begin(),alist.end(), boost::lambda::if_then((bind(&StrList::find,&excluded,bind(&A::get,_1))) !=excluded.end()), bind(&StrList::insert,&excluded,bind(&A::get,_1))); std::for_each(alist.begin(),alist.end(), bind(&A::print_A,_1)); return 1; } ----------------------------------- main2.cpp:43: error: no matching function for call to `bind(<unknown type>, StrList*, const boost::lambda::lambda_functor<boost::lambda::lambda_functor_base<boost::lambda::action<2, boost::lambda::function_action<2, boost::lambda::detail::unspecified> >, boost::tuples::tuple<std::string (A::*const)(), const boost::lambda::lambda_functor<boost::lambda::placeholder<1> >, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type> > >)' main2.cpp:45: error: no matching function for call to `bind(<unknown type>, StrList*, const boost::lambda::lambda_functor<boost::lambda::lambda_functor_base<boost::lambda::action<2, boost::lambda::function_action<2, boost::lambda::detail::unspecified> >, boost::tuples::tuple<std::string (A::*const)(), const boost::lambda::lambda_functor<boost::lambda::placeholder<1> >, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type> > >)' -- Thad

2 1

Re: [Boost-users] Missing functions in date_time library
by Sean Rohead 28 Jul '05

28 Jul '05

I'm even more confused after looking at the URL and test file you suggested. All of the examples in those documents require you to explicitly construct and provide a time zone using a name or an offset from UTC. I don't want to have to hardcode a timezone in my program -- I just want to use whatever time zone that the user has configured on their computer. The c_local_adjustor did this for me in 1.32 (although as noted in my original email, it only solves half of the problem). Perhaps I missed something in the URL you sent, but I didn't see anything in there which simply uses the timezone in use on the local machine. Thanks, Sean -----Original Message----- From: boost-users-bounces(a)lists.boost.org [mailto:boost-users-bounces@lists.boost.org] On Behalf Of Jeff Garland Sent: Saturday, July 23, 2005 12:44 PM To: boost-users(a)lists.boost.org Subject: Re: [Boost-users] Missing functions in date_time library On Fri, 22 Jul 2005 18:09:10 -0600, Sean Rohead wrote > I am using the boost date_time library and noticed that there is a > from_ftime() function to convert a FILETIME into a ptime, but I could > not find the reverse operation to_ftime(). Am I missing it somewhere? No. FILETIME, as you're certainly aware, is a Windows only way to measure time. This function is there primarily to support conversion from legacy applications. If there's really some need it's not hard to support. > Likewise, I am using the c_local_adjustor::utc_to_local() function > and cannot find the corresponding local_to_utc() function in c_local_adjustor. In 1.33 there is vastly enhanced support for local time conversion via the timezone class and it's friends posix_timezone and tz_database. See <http://engineering.meta-comm.com/resources/cs-win32_metacomm/doc/html/d ate_time/local_time.html#date_time.local_time.time_zone_base> for more on this. If you still want to use the compile time approach supported in 1.32, have a look at libs/date_time/test/posix_time/testlocal_adjustor.cpp. There's code in there that show how to convert back and forth. Jeff _______________________________________________ Boost-users mailing list Boost-users(a)lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users

2 1

Re: [Boost-users] Missing functions in date_time library
by Sean Rohead 28 Jul '05

28 Jul '05

Just as a follow up to this, let me describe what I'm trying to do. I would like the current time in UTC with microsecond resolution. Well, the microsec_clock class only provides a local_time() method (unlike the second_clock class which also provides universal_time()). So, I need to convert the result of local_time() into UTC but I can't since the only class which knows the local timezone is c_local_adjustor and it does not provide local_to_utc() (only utc_to_local()). If there is a better way to get the current time in UTC with microsecond resolution, I'd be happy to use it instead. Thanks, Sean -----Original Message----- I'm even more confused after looking at the URL and test file you suggested. All of the examples in those documents require you to explicitly construct and provide a time zone using a name or an offset from UTC. I don't want to have to hardcode a timezone in my program -- I just want to use whatever time zone that the user has configured on their computer. The c_local_adjustor did this for me in 1.32 (although as noted in my original email, it only solves half of the problem). Perhaps I missed something in the URL you sent, but I didn't see anything in there which simply uses the timezone in use on the local machine. Thanks, Sean On Fri, 22 Jul 2005 18:09:10 -0600, Sean Rohead wrote > I am using the boost date_time library and noticed that there is a > from_ftime() function to convert a FILETIME into a ptime, but I could > not find the reverse operation to_ftime(). Am I missing it somewhere? No. FILETIME, as you're certainly aware, is a Windows only way to measure time. This function is there primarily to support conversion from legacy applications. If there's really some need it's not hard to support. > Likewise, I am using the c_local_adjustor::utc_to_local() function and > cannot find the corresponding local_to_utc() function in c_local_adjustor. In 1.33 there is vastly enhanced support for local time conversion via the timezone class and it's friends posix_timezone and tz_database. See <http://engineering.meta-comm.com/resources/cs-win32_metacomm/doc/html/d ate_time/local_time.html#date_time.local_time.time_zone_base> for more on this. If you still want to use the compile time approach supported in 1.32, have a look at libs/date_time/test/posix_time/testlocal_adjustor.cpp. There's code in there that show how to convert back and forth. Jeff _______________________________________________ Boost-users mailing list Boost-users(a)lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users

2 1

Re: [Boost-users] Test Tools Linking Problem
by Merrill Cornish 28 Jul '05

28 Jul '05

Gennadi, Thanks for the suggestions. I had reduced the failing test to a single BOOST_MESSAGE() macro call to simplify the error output. (All of the regex stuff links fine with the DLL.) When I try linking my entire test file, *ALL* of the unresolved references involve test_tools::tt_detail::... It's as if just the test tool details weren't created. If I remove the test tools library reference from the Makefile, then I get undefined test tools references for *OTHER* than details. So, at least something in the library is being used by the linker--just not the stuff from the details. I have no idea why the detail references would have been dropped. I have tried rebuilding the entire Boost library with bjam -sTOOLS=mingw -sMINGW_IMPLEMENTATION_DIR=E:\MinGW stage (I may have mis-remembered the exact name for the implementation directory, but I got it right.) This time I captured the output. There were error message about the serialization library because of wide characters, but no complaints about any of the others. The rebuilt library gives the same errors. I would certainly like to use the Boost test library, but for the moment, I'm going to have to go back to implementing it myself. Merrill

2 1

Question about date
by Larry 28 Jul '05

28 Jul '05

I searched the group and didn't see anything that answered my question about date although I may have just not recognized it when I read it. The date constructor will throw exceptions if you use an invalid date: date d(2005,2,29); Throws an exception because there is no 29th of February in 2005. If one does the following: string ds("20050229T000000"); date d(from_iso_string(ds)); there is no exception and like localtime() and gmtime() it actually changes the date silently to 20050301. (ptime also does this.) Was this done intentionally perhaps to be consistent with localtime() etc although that begs the question of why there was an exception in the first case of date construction? Or perhaps I forgot to do something?

1 0

Re: [Boost-users] bosst::bind problem
by Jasper, Alexander 28 Jul '05

28 Jul '05

Hi Oleg! your function Handler::Test2 takes 2 arguments + this. But your code passes 3 arguments + this to the function boost::bind(&Handler::Test2, _1, _2, a, b)). If I'm not completely mistaken that cannot work. What did you expect the _2 to be in this case?. Using the _2 also implies that the resulting function of the bind expects 2 arguments. That is you cannot simply use it with for_each in that kind of way. This supplies only one parameter. kind regards Alex > -----Ursprüngliche Nachricht----- > Von: boost-users-bounces(a)lists.boost.org > [mailto:boost-users-bounces@lists.boost.org]Im Auftrag von > Oleg Smolsky > Gesendet: Donnerstag, 28. Juli 2005 03:22 > An: boost-users(a)lists.boost.org > Betreff: [Boost-users] bosst::bind problem > > > Hi all, > > I've just started using boost::bind and found the following problem: > > class Handler > { > public: > bool Test1(std::string sComponent); > bool Test2(std::string sComponent, std::string sAddress); > }; > > void Test() > { > std::vector<Handler> v; > > // This compiles > std::string a, b; > std::for_each(v.begin(), v.end(), > boost::bind(&Handler::Test1, _1, a)); > > // This doesn't compile > //std::for_each(v.begin(), v.end(), > // boost::bind(&Handler::Test2, _1, _2, a, b)); > } > > Environment: boost 1.32, VC8, WinXP. > > Any ideas? > > Thanks, > Oleg. > > _______________________________________________ > Boost-users mailing list > Boost-users(a)lists.boost.org > http://lists.boost.org/mailman/listinfo.cgi/boost-users >

1 0