May 2003 - Boost-users - lists.boost.org

date_time link or compile problem
by Jimmy Pritts 08 May '03

08 May '03

I have boost 1.29 and gcc 3.2.2 on Linux. Here is the simple code I am trying to compile: #include <string> #include "boost/date_time/gregorian/gregorian.hpp" int main() { using namespace std; using namespace boost::gregorian; string ds("2002/1/25"); date d1(from_string(ds)); return 0; } I get the following link error: make g++ main.cc -lstdc++ -o test /tmp/ccph1inq.o(.gnu.linkonce.t._ZN5boost9date_time4dateINS_9gregorian4dateENS2_18gregorian_calendarENS0_13date_durationIlEEEC2ERKNS0_19year_month_day_baseINS2_9greg_yearENS2_10greg_monthENS2_8greg_dayEEE+0x11): In function `boost::date_time::date<boost::gregorian::date, boost::gregorian::gregorian_calendar, boost::date_time::date_duration<long> >::date[not-in-charge](boost::date_time::year_month_day_base<boost::gregorian::greg_year, boost::gregorian::greg_month, boost::gregorian::greg_day> const&)': : undefined reference to `boost::date_time::gregorian_calendar_base<boost::date_time::year_month_day_base<boost::gregorian::greg_year, boost::gregorian::greg_month, boost::gregorian::greg_day>, unsigned long>::day_number(boost::date_time::year_month_day_base<boost::gregorian::greg_year, boost::gregorian::greg_month, boost::gregorian::greg_day> const&)' collect2: ld returned 1 exit status make: *** [test] Error 1 The missing symbol does not seem to be in the static library libboost_date_time.a, so I don't think I need to link to that. Any suggestions? Thanks in advance, Jimmy.

2 1

RE: [Boost-Users] call_traits and MSVC 7.1
by Arjun Bijanki 08 May '03

08 May '03

> From: Randy Bowen [mailto:rbowen@stamps.com] > Sent: Wednesday, May 07, 2003 5:46 PM > To: Boost-Users(a)yahoogroups.com > Subject: [Boost-Users] call_traits and MSVC 7.1 > > We have been using call_traits from Boost version 1.29.0 in > conjunction with MSVC 7.0. When attempting to recompile this code > under MSVC 7.1, using either Boost 1.29.0 or 1.30.0, we now > encounter compilation errors. A simple example of the usage pattern > is: <code snipped> This is illegal code. That it compiled in MSVC 7.0 was a bug. The template parameter '_Source' can not be deduced from the callsite of doSomething. 14.8.2.4 in the ISO/IEC 14882 lists the forms in which template parameters can be deduced, but a qualified name involving nested types is not one of them. I suspect it's because of something like this: template<typename T> struct S { typedef T type; }; template<> struct S<char> { typedef int type; }; template<typename T> void foo(typename S<T>::type); int main() { foo(3); //what is T deduced to be? } I'd guess the documentation for call_traits discusses typical usage patterns, and may help. You can also explicitly specify the template argument. Hope this helps. Arjun Bijanki Microsoft Visual C++

1 0

ublas: identity_matrix and row()
by Alexander G Winbow 08 May '03

08 May '03

It seems that identity_matrix does not work with row(); is this supposed to be the case? For example: ----------- code excerpt ----------- #include <boost/numeric/ublas/matrix.hpp> #include <boost/numeric/ublas/vector.hpp> int main () { using namespace boost::numeric::ublas; //matrix<double> im(3,3); // this works identity_matrix<double> im(3); // this does not boost::numeric::ublas::vector<double> v = row(im,1); } ------------------------------------ Using matrix<> im, the code works fine, but using identity_matrix<> im, g++ blows up with four or five errors of the form included below, on lines 529, 61, 100, 104, and 84 of matrix_proxy.hpp. I've got quite a lot of questions on ublas, even after combing through the ublas and boost documentation, wiki, and mailing lists at great length. Is this an appropriate forum to pepper with questions, or is there some other manual I can RTFM first? Thanks, -Alex -------- errors excerpt compiling the above --------- /home/linux/phcomp/ph141w/ph141wam/libraries/boost_1_30_0/boost/numeric/ublas/matrix_proxy.hpp: In instantiation of `boost::numeric::ublas::matrix_row<boost::numeric::ublas::identity_matrix<double> >': rowtest.cpp:10: instantiated from here /home/linux/phcomp/ph141w/ph141wam/libraries/boost_1_30_0/boost/numeric/ublas/matrix_proxy.hpp:529: no type named `closure_type' in `class boost::numeric::ublas::identity_matrix<double>' /home/linux/phcomp/ph141w/ph141wam/libraries/boost_1_30_0/boost/numeric/ublas/matrix_proxy.hpp:529: template argument 3 is invalid -----------------------------------------------------------

2 1

Why isn't the "last access time" part of the filesystem library?
by richard_fanta 07 May '03

07 May '03

Looking at the filesystem library and the attributes work in boost- sandbox, I don't immediately see "last access time" being available for a file or directory. "Last access time" was included in Dietmar Kuehl's original dir_it. Can someone kindly tell me why this was removed? It's highly useful, and part of every filesystem that I've seen. Thanks, Rick

2 2

RegEx Split function dies.
by tfandango 07 May '03

07 May '03

Hi- I am attempting to parse email headers using the boost regex library. I would prefer that each email address in the TO field (delimited by a comma) would be placed into a vector<string>. Currently, I am able to obtain the entire to line (all the to addresses). I can use another regular expression to seperate it out but I'd rather do it all in one. However, If I try to add another more to this I get an exception: "Max regex search depth exceeded." I am assuming that this is because I've created a infinite loop. I'm not sure how to extend this expression to place every email address in the vector seperatly. Thanks here's my code snippet: char * expression_text = //Any possible leading whitespace: "^[ ]*" //The "To" field name in any case: "TO" //Any more whitespace: "[ ]*" //Then a colon: ":" //More white space potentially "[ ]*" //Everything here is an address up to "\r\n"; "([-(a)_\\.\"<>A-Za-z0-9]+),?"; //Create the expression class to use: RegEx expression(expression_text, true); //Attempt to grep the mail_buffer: int num_recv = 0; try { num_recv = expression.Split(to, mail_buffer); } catch (...){ //throw some exception } }

2 2

bjam and windows command line problem.
by Fernando Cacciola 07 May '03

07 May '03

Hi, I'm having a hard time trying to use bjam to build boost 1.29.0. I have Win98 in Spanish, and I use Borland C++ Builder 5.0 (its bcc32.exe corresponds to is BCC5.5.1) I have boost 1.29.0 installed at: "C:\Libraries\boost\Repository\boost" I've downloaded bjam for Microsoft Windows: >bjam -v Boost.Jam Version 3.1.0. OS=NT. Copyright 1993-2002 Christopher Seiwald and Perforce Software, Inc. Copyright 2001 David Turner. Copyright 2001-2002 David Abrahams. I've temporarily modified the root 'jamfile' so that it builds only the Test library. and I've done: subst d: "c:\archivos de programa\borland\cbuilder5" set BCCROOT=d: (you'll see why in a moment) If I try: c:\>cd C:\Libraries\boost\Repository\boost" C:\Libraries\boost\Repository\boost> bjam -sTOOLS=borland I receive messages of the form: borland-C++-action libs\test\build\bin\libpem.lib\borland\debug\runtime-link-static\threading-s ingle\execution_monitor.obj "d:\bin\bcc32" -j5 -q -c -P -w -Ve -Vx -a8 -v -Od -vi- -tWC -WM- -w-8001 -I"libs\test\build" -I"C:\Libraries\boost\Repository\boost" -I"d:\includ e" -o"libs\test\build\bin\libpem.lib\borland\debug\runtime-link-static\thre ading-single\execution_moni Error E2266: No file names given tor.obj" "libs\test\build\../src\execution_monitor.cpp" C:\WINDOWS\TEMP\JAMTMP00.BAT [2] Unknown command "TOR.OBJ" "LIBS\TEST\BUILD\.." "d:\bin\bcc32" -j5 -q -c -P -w -Ve -Vx -a8 -v -Od -vi- -tWC -WM- -w-8001 -I"libs\test\build" -I"C:\Libraries\boost\Repository\boost" -I"d:\includ e" -o"libs\test\build\bin\libpem.lib\borland\debug\runtime-link-static\thre ading-single\execution_monitor.obj" "libs\test\build\../src\execution_monitor.cpp" ...failed borland-C++-action libs\test\build\bin\libpem.lib\borland\debug\runtime-link-static\threading-s ingle\execution_monitor.obj... If I'm interpreting this output correctly, the problem is that the command line feed to bcc32 exceeds 256 characters. What can be done about this? Is there a way for 'borland-tools.jam' to put the bcc32 parameters in a tmp response file and use: bcc32.exe @tmprsp.rsp? Any help will be appreciated. TIA, -- Fernando Cacciola

4 22

RE: PP Library...
by Eric Robert 07 May '03

07 May '03

Ok, I see that there are 3 ways of doing this now. What is the recommendation for something small like 2 or 3 lines for each native types i.e. more or less 12 iterations? Eric "Paul Mensonides" <yg-boost-users(a)m.gmane.org> wrote > Tanton Gibbs wrote: > > You can use BOOST_PP_LIST_FOR_EACH > > > > #define ARG_LIST (char, (int, (unsigned int, (long, ...)))) > > #define DO_IT(r, data, elem) void foo( elem ); > > > > BOOST_PP_LIST_FOR_EACH( DO_IT, , ARG_LIST ) > ^ > This empty argument is undefined in C++. In C99, it is allowed and called a > "placemarker." Instead of passing nothing, just pass anything and ignore it. > > // ----- // > > #include <boost/preprocessor/list/for_each.hpp> > > #define ARG_LIST \ > (char, (int, (unsigned, (long, BOOST_PP_NIL)))) \ > /**/ > #define DO_IT(r, ignored, elem) \ > void foo( elem ); \ > /**/ > > BOOST_PP_LIST_FOR_EACH( DO_IT, ?, ARG_LIST ) > > #undef ARG_LIST > #undef DO_IT > > // ----- // > > Note that there are other ways to do this also. For instance, use of > "sequences" makes the definition of ARG_LIST easier: > > // ----- // > > #include <boost/preprocessor/seq/for_each.hpp> > > #define ARGS \ > (char)(int)(unsigned)(long) \ > /**/ > #define DO_IT(r, ignored, elem) \ > void foo( elem ); \ > /**/ > > BOOST_PP_SEQ_FOR_EACH( DO_IT, ?, ARGS ) > > #undef ARGS > #undef DO_IT > > // ----- // > > Of course, if what you want to generate is much larger, you might want to use > some form of vertical repetition: > > // ----- // > > #include <boost/preprocessor/iteration/local.hpp> > #include <boost/preprocessor/seq/elem.hpp> > > #define ARGS \ > (char)(int)(unsigned)(long) \ > /**/ > > #define BOOST_PP_LOCAL_MACRO(n) \ > void foo( BOOST_PP_SEQ_ELEM(n, ARGS) ); \ > /**/ > #define BOOST_PP_LOCAL_LIMITS \ > (0, BOOST_PP_SEQ_SIZE(ARGS) - 1) \ > /**/ > #include BOOST_PP_LOCAL_ITERATE() > > #undef ARGS > > // ----- // > > Regards, > Paul Mensonides [Non-text portions of this message have been removed]

3 8

RE: [Boost-Users] bug in pthreads implementation of recursive_mut ex
by Feldman, Yulik 06 May '03

06 May '03

Thanks Alexander for the pointer to this long and interesting discussion. I'll check my design whether I could get rid of the recursive mutexes in order to reduce the probability of not holding invariants when the wait() is called, especially since the code still doesn't behave correctly, even after the Boost bug is fixed :) William, unfortunately I don't have a small testcase for the bug and it seems that it will be too time-consuming for me to try to prepare a small "working" testcase. Thanks for confirming the bug report (it makes me feel better that the debugging time was not wasted). --Yulik. -----Original Message----- From: William E. Kempf [mailto:wekempf@cox.net] Sent: Monday, May 05, 2003 6:44 PM To: Boost-Users(a)yahoogroups.com Subject: Re: [Boost-Users] bug in pthreads implementation of recursive_mutex Feldman, Yulik said: > Hi, > > > > It looks that there is a bug in > .../boost_1_30_0/libs/thread/src/recursive_mutex.cpp. The bug shows in > the pthreads implementation of the recursive_mutex when the mutex is > used with the "condition" synchronization primitive. Here is the > description of the bug: You're evaluation of the bug is correct. I'll fix this in CVS shortly. Do you maybe have a test case for this that I could add to the regression tests? BTW, Mr. Terekhov does provide a good resource in his posting on this subject. I disagree with some points in the thread he posted, but it is quite true that you absolutely must gaurantee that invariants hold at the point cond.wait() is called, and that doing so is complicated by the use of recursive locks. Either avoid recursive locks entirely, avoid condition variables with recursive mutexes, or be very sure of the design and the state of your invariants. -- William E. Kempf Info: <http://www.boost.org> Wiki: <http://www.crystalclearsoftware.com/cgi-bin/boost_wiki/wiki.pl> Unsubscribe: <mailto:boost-users-unsubscribe@yahoogroups.com> Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/

2 1

Problems with mpl::push_back
by joaquintides 06 May '03

06 May '03

#include <boost/mpl/list.hpp> #include <boost/mpl/push_back.hpp> int main() { typedef boost::mpl::list<> type0; typedef boost::mpl::push_back<type0,int>::type type1; typedef boost::mpl::push_back<type1,char>::type type2; return 0; } This fails in MSVC 6.5 and gcc 3.2. The error seems to indicate that push_back does not provide a type named 'type'. If I replace push_back with push_front, eveything works wonderfully. Am I doing any gross error, is the documentation wrong, ideas, clues? Thanx in advance. Joaquín M López Muñoz Telefónica, Investigación y Desarrollo

2 1

boost::shared_ptr templated constructor?
by Lehrer, Joshua 06 May '03

06 May '03

WRT boost::shared_ptr<>, why not template the explicit constructor that takes a T*? This way, it could select the correct destructor at construction time: struct base { ~base(); }; struct sub : public base { ~sub() }; boost::shared_ptr<base>(new sub); The above code currently only calls ~base(); Templating the explicit single argument constructor would allow the above code to call ~sub(). The constructor now takes only a T*, and therefore populates the pointer to the destructor function with a call to ~T(), which is ~base() in the above example. If, however, the constructor took an X*, the above example would pass a sub*, and it could populate the default destructor function with a pointer to ~X(). Comments? Joshua Lehrer FactSet Research Systems NYSE:FDS

3 2