July 2004 - Boost-users - lists.preview.boost.org

RE: [Boost-users] Re: Running boost::thread compiled on w2k on XP
by Paul Rose 28 Jul '04

28 Jul '04

This also isn't exactly what you are looking for, but I took built your test.cpp on vc7.1 running on windows XP using boost 1.31 (also built with vc7.1 running on windows XP). Both the debug (using boost_thread-vc71-mt-gd-1_31.dll) and the release (using boost_thread-vc71-mt-1_31.dll) builds of test.cpp worked fine. My boost dlls were built with all default settings (except I used the alternate namespace hack for boost::signals since we also use Qt). I don't have a boost 1.31 built on win2k to try. > >Perhaps I should add that I am using version 1.31 of boost. I have also >discovered that it seems (it might be an illusion, you never know) to work >when I compile against boost_thread-vc71-mt-gd-1_31.lib, but not against >the release lib boost_thread-vc71-mt-1_31.lib. >/Ragnar Ouchterlony _________________________________________________________________ Overwhelmed by debt? Find out how to Dig Yourself Out of Debt from MSN Money. http://special.msn.com/money/0407debt.armx

1 0

[boost.build] Patch for python.jam
by Markus Schöpflin 28 Jul '04

28 Jul '04

Hello, attached is a patch for pyhton.jam which allows to use python installation on Unix which have been configured with different prefix and exec-prefix. Currently this is not supported by boost.build. The patch simply adds the missing include directory to the compiler command line as recommended in http://www.python.org/doc/2.3.4/api/includes.html . After applying the patch you can either use bjam -sPYTHON_ROOT=<prefix> or bjam -sPYTHON_ROOT=<prefix> -sPYTHON_EXEC_ROOT=<exec prefix> Markus --- net/src/orig/boost_1_31_0/tools/build/v1/python.jam 2004-01-28 23:52:25.000000000 +0100 +++ local/src/boost_1_31_0/tools/build/v1/python.jam 2004-07-28 16:15:37.000000000 +0200 @@ -118,7 +118,7 @@ { PYTHON_ROOT ?= /usr/local ; PYTHON_ROOT = $(PYTHON_ROOT:J=" ") ; - PYTHON_INCLUDES ?= $(PYTHON_ROOT)/include/python$(PYTHON_VERSION) ; + PYTHON_INCLUDES ?= $(PYTHON_ROOT)/include/python$(PYTHON_VERSION) $(PYTHON_EXEC_ROOT)/include/python$(PYTHON_VERSION) ; PYTHON_LIB_PATH ?= $(PYTHON_ROOT)/lib/python$(PYTHON_VERSION)/config ; PYTHON_PROPERTIES ?=

1 0

RE: [Boost-users] Re: Bjam- racking my brain
by Brian Braatz 28 Jul '04

28 Jul '04

> > Brian Braatz wrote: > > > Any help would be appreciated, tomorrow afternoon I am going to go back > > at it, and this time I might "get all medieval" about things :) > Rene Rivera Wrote: > build/v2 is not used for building Boost. It's the experimental, and > under development, next version of Boost.Build. If you plan on using > BBv2 read its separate documentation at > boost-root/tools/build/v2/index_v2.html I am sorry - I am confused and am trying to find my way out: *- What is listed for download on sourceforge does not reference a version of build v1 or bjam v1 or what have you. ( I am confused as to how to get the older version) * when I build the code in jam_src and run it- it reports: Boost.Jam Version 3.1.8. OS=NT. This is the same version I used to build boost with (successfully) with vc7.1. I am using the source for bjam included in the 1.31 release of boost * when I try to build boost with a simple bjam stage at the root dir of boost, it succeeds * when I double check the BOOST_ROOT env var with a dir command- it shows me the right dir * my TOOLS envvar is set to TOOLS=vc7.1 This was necessary before being able to build boost * when I CD into the HELLO example for bjam and try a Bjam I GET warning: no toolsets are configured. warning: you won't be able to build C++ programs. warning: please consult the documentation. C:/Data/Rmtg/Root/_Common/_Lib/Boost/V_1_31/tools/build/v2/build\targets .jam:1073: in construct from module object(exe-t arget-class)@1 error: unable to construct ./hello C:/Data/Rmtg/Root/_Common/_Lib/Boost/V_1_31/tools/build/v2/build\targets .jam:914: in object(exe-target-class)(a)1.generate from module object(exe-target-class)@1 C:/Data/Rmtg/Root/_Common/_Lib/Boost/V_1_31/tools/build/v2/build\targets .jam:510: in generate-really from module object( main-target)@1 C:/Data/Rmtg/Root/_Common/_Lib/Boost/V_1_31/tools/build/v2/build\targets .jam:483: in object(main-target)(a)1.generate from module object(main-target)@1 C:/Data/Rmtg/Root/_Common/_Lib/Boost/V_1_31/tools/build/v2/build\targets .jam:179: in object(project-target)(a)1.generate f rom module object(project-target)@1 C:/Data/Rmtg/Root/_Common/_Lib/Boost/V_1_31/tools/build/v2\build-system. jam:160: in load from module build-system C:\Data\Rmtg\Root\_Common\_Lib\Boost\V_1_31\tools\build\v2\example\..\ke rnel\modules.jam:259: in import from module modu les C:\Data\Rmtg\Root\_Common\_Lib\Boost\V_1_31\tools\build\v2\example\..\ke rnel\bootstrap.jam:120: in boost-build from modu le C:\Data\Rmtg\Root\_Common\_Lib\Boost\V_1_31\tools\build\v2\example\boost -build.jam:2: in module scope from module When I TRY: bjam -sTOOLS=vc7.1 bjam "-sTOOLS=vc7.1" bjam debug bjam release bjam "-sTOOLS=vc7.1" debug bjam "-sTOOLS=vc7.1" release I always get the above errors If I COPY the files in \tools\build\v2\example\hello to the ROOT of boost (and overwrite the jam file there)- IT DOES BUILD PROPERLY I have build bjam and stepped through it in the IDE and was not able to make much of what the source was doing (other than it looks like a lot of recursive parsing of files) (spent most of my evening yesterday doing that. MY QUESTIONS: * Which is the right version of boost to be running here? * does the \tools\build\v2\example\hello sample bjam compile for anyone else? (i.e. is it broken?) * is there a good sample of a simple working example for bjam somewhere else? * Any ideas on why I can BUILD BOOST but not any of the subprojects (everything else I have tried has the same problems) * How do you folks debug this thing- ****any**** techniques would be helpful * is the toolsets it is complaining about the COMPILER (i.e. TOOLS envvar?) or is it something else entirely? * HELP ! :) I am doing this pain because I am trying to get boost defined as the standard in my company (It is within my power to do that). I would also like to use the bjam tool as our standard build tool, but I FIRST need to figure out how to build projects with it (grin). Thanks to anyone for thoughts or things to try. Thanks Brian C Braatz

2 1

Difference in adjacency_list
by Daniele Carlucci 27 Jul '04

27 Jul '04

Hi, I'm a newby. Anyone can explain me the difference between: typedef adjacency_list<listS, vecS, undirectedS... and typedef adjacency_list<vecS, vecS, undirectedS... Thanks Daniele

3 4

Bjam- racking my brain
by Brian Braatz 27 Jul '04

27 Jul '04

For the last several days (a few hours each day) I have been trying to get the \tools\build\v2\example\hello sample to build with bjam- (windows XP, vc7.1) I CAN build boost from the command line, but when I move into the \tools\build\v2\example\hello dir and try various combinations of bjam (with and without -sTools=vc7.1 etc..) I get the same error no matter what I do. I took the bjam sources and compiled a debug version in the Visual Studio IDE, and have spent the evening stepping through the code and reading the docs. This is one of these things I think that I just have to completely internalize how this thing works before I will crack this. What I am asking for from this list: * Any basic advice on debugging what bjam is doing * any basic advice or good sample that is a bare and working bjam project to look at * This is weird to me, as the ENV var for the toolset DOES work for building boost itself * I feel like I am missing something obvious and stupid, anyone who wishes to point out what dumb thing I missed, please go at it :) * any help in looking at this output: C:\Data\Rmtg\Root\_Common\_Lib\Boost\V_1_31\tools\build\v2\example\hello >bjam warning: no toolsets are configured. warning: you won't be able to build C++ programs. warning: please consult the documentation. C:/Data/Rmtg/Root/_Common/_Lib/Boost/V_1_31/tools/build/v2/build\targets .jam:1073: in construct from module object(exe-t arget-class)@1 error: unable to construct ./hello C:/Data/Rmtg/Root/_Common/_Lib/Boost/V_1_31/tools/build/v2/build\targets .jam:914: in object(exe-target-class)(a)1.generate from module object(exe-target-class)@1 C:/Data/Rmtg/Root/_Common/_Lib/Boost/V_1_31/tools/build/v2/build\targets .jam:510: in generate-really from module object( main-target)@1 C:/Data/Rmtg/Root/_Common/_Lib/Boost/V_1_31/tools/build/v2/build\targets .jam:483: in object(main-target)(a)1.generate from module object(main-target)@1 C:/Data/Rmtg/Root/_Common/_Lib/Boost/V_1_31/tools/build/v2/build\targets .jam:179: in object(project-target)(a)1.generate f rom module object(project-target)@1 C:/Data/Rmtg/Root/_Common/_Lib/Boost/V_1_31/tools/build/v2\build-system. jam:160: in load from module build-system C:\Data\Rmtg\Root\_Common\_Lib\Boost\V_1_31\tools\build\v2\example\..\ke rnel\modules.jam:259: in import from module modu les C:\Data\Rmtg\Root\_Common\_Lib\Boost\V_1_31\tools\build\v2\example\..\ke rnel\bootstrap.jam:120: in boost-build from modu le C:\Data\Rmtg\Root\_Common\_Lib\Boost\V_1_31\tools\build\v2\example\boost -build.jam:2: in module scope from module C:\Data\Rmtg\Root\_Common\_Lib\Boost\V_1_31\tools\build\v2\example\hello >set TOOLs TOOLS=vc7.1 Any help would be appreciated, tomorrow afternoon I am going to go back at it, and this time I might "get all medieval" about things :)

2 1

long long support with GCC
by Aaron W. LaFramboise 27 Jul '04

27 Jul '04

Hi, As I'm sure you know, long long and unsigned long long are supported by GCC in C++. However, -std=c++98 (not quite sure why, to be honest) and -pedantic cause lots of warnings and errors to be generated in the Boost headers when long long is used. In my project, it is desirable to use these flags for other reasons, but the effect these flags have on the Boost headers is decidedly undesirable. So far, I've been wrapping the Boost headers in forwarding headers that mark the Boost headers as system headers so that they compile without warning or error. Is there any disadvantage to doing this? GCC supports long long even with the above options, but only if __extension__ is used. I notice that there a Boost-standard 64bit type, boost::int64_t. Is there some reason that this can't be typedef'ed using the GCC extension, and then this identifier used in all places where 'long long' is being used? Are there type system rules that prevent this, or some other reason? Thank you for your consideration, Aaron W. LaFramboise

4 6

building boost on vc7.1 with stlport
by tweety 26 Jul '04

26 Jul '04

i first compiled (succesfully) stlport 4.6.2. then i ran "bjam "-sTOOLS=vc7.1-stlport" install" and it failed compiling the regex library. then i manually went to the libs\regex\build folder and ran "nmake /f vc7.1-stlport clean install" and it compiled the first thing ok and the 2nd time around threw a LOT of linker errors (well, only ~49, but they're stl-style-long :D). I set correctly the STLPORT_VERSION and the STLPORT_PATH env vars, the tools were loaded ok, the INCLUDE env var had stlport's include dir first... AND IT COMPILED OK the first one... here's a screenshot of the last page of errors. please help me... winstances.obj : error LNK2019: unresolved external symbol "__declspec(dllimport ) public: __thiscall _STL::basic_string<wchar_t,class _STL::char_traits<wchar_t> ,class _STL::allocator<wchar_t> >::basic_string<wchar_t,class _STL::char_traits< wchar_t>,class _STL::allocator<wchar_t> >(wchar_t const *,unsigned int,class _ST L::allocator<wchar_t> const &)" (__imp_??0?$basic_string@_WV?$char_traits@_W@_ST L@@V?$allocator@_W@2@@_STL@@QAE@PB_WIABV?$allocator@_W@1@@Z) referenced in funct ion "public: class _STL::basic_string<wchar_t,class _STL::char_traits<wchar_t>,c lass _STL::allocator<wchar_t> > __fastcall boost::reg_expression<wchar_t,class b oost::regex_traits<wchar_t>,class _STL::allocator<wchar_t> >::str(void)const " ( ?str@?$reg_expression@_WV?$regex_traits@_W@boost@@V?$allocator@_W@_STL@@@boo st@@ QBI?AV?$basic_string@_WV?$char_traits@_W@_STL@@V?$allocator@_W@2@@_STL@@XZ) winstances.obj : error LNK2019: unresolved external symbol "__declspec(dllimport ) public: int __thiscall _STL::basic_string<wchar_t,class _STL::char_traits<wcha r_t>,class _STL::allocator<wchar_t> >::compare(class _STL::basic_string<wchar_t, class _STL::char_traits<wchar_t>,class _STL::allocator<wchar_t> > const &)const " (__imp_?compare@?$basic_string@_WV?$char_traits@_W@_STL@@V?$allocator@_W@2@@ _S TL@@QBEHABV12@@Z) referenced in function "public: int __fastcall boost::reg_expr ession<wchar_t,class boost::regex_traits<wchar_t>,class _STL::allocator<wchar_t> >::compare(class boost::reg_expression<wchar_t,class boost::regex_traits<wchar_ t>,class _STL::allocator<wchar_t> > const &)const " (?compare@?$reg_expression@_ WV?$regex_traits@_W@boost@@V?$allocator@_W@_STL@@@boost@@QBIHABV12@@Z) winstances.obj : error LNK2019: unresolved external symbol "__declspec(dllimport ) public: class _STL::basic_string<wchar_t,class _STL::char_traits<wchar_t>,clas s _STL::allocator<wchar_t> > & __thiscall _STL::basic_string<wchar_t,class _STL: :char_traits<wchar_t>,class _STL::allocator<wchar_t> >::assign(unsigned int,wcha r_t)" (__imp_?assign@?$basic_string@_WV?$char_traits@_W@_STL@@V?$allocator@_W@2@ @_STL@@QAEAAV12@I_W@Z) referenced in function "private: struct boost::re_detail: :re_syntax_base * __fastcall boost::reg_expression<wchar_t,class boost::regex_tr aits<wchar_t>,class _STL::allocator<wchar_t> >::compile_set_aux(class boost::re_ detail::jstack<class _STL::basic_string<wchar_t,class _STL::char_traits<wchar_t> ,class _STL::allocator<wchar_t> >,class _STL::allocator<wchar_t> > &,class boost ::re_detail::jstack<class _STL::basic_string<wchar_t,class _STL::char_traits<wch ar_t>,class _STL::allocator<wchar_t> >,class _STL::allocator<wchar_t> > &,class boost::re_detail::jstack<unsigned long,class _STL::allocator<wchar_t> > &,class boost::re_detail::jstack<class _STL::basic_string<wchar_t,class _STL::char_trait s<wchar_t>,class _STL::allocator<wchar_t> >,class _STL::allocator<wchar_t> > &,b ool,struct boost::re_detail::_narrow_type const &)" (?compile_set_aux@?$reg_expr ession@_WV?$regex_traits@_W@boost@@V?$allocator@_W@_STL@@@boost@@AAIPAUre_sy ntax _base@re_detail@2@AAV?$jstack@V?$basic_string@_WV?$char_traits@_W@_STL@@V?$a lloc ator@_W@2@@_STL@@V?$allocator@_W@2@@42@0AAV?$jstack@KV?$allocator@_W@_STL@@@ 42@0 _NABU_narrow_type@42@@Z) vc71-stlport/boost_regex-vc71-mt-p-1_31.dll : fatal error LNK1120: 49 unresolved externals NMAKE : fatal error U1077: 'link' : return code '0x460' Stop. ---------------------------------- Peace and love, Tweety mitea(a)sympatico.ca - tweety_04_01(a)users.sourceforge.net YahooID: tweety_04_01

2 18

RE: [Boost-users] using boost::shared_ptr with boost::signal
by Ben Hutchings 26 Jul '04

26 Jul '04

Oliver Kowalke <oliver.kowalke(a)t-online.de> wrote: > Hi, > > I want to use boost::shared_ptr with boost::signal. > > class A > { > private: > int _i; > public: > A() : _i( 0) {} > > void operator()( int, std::string & s) > { i+= 1; } > > void operator()( std::string & s, int) > { i+= 2; } > }; > > boost::shared_ptr< A > a( new A() ); > boost::signal< void( int, std::string &) > rsig; > boost::singal< void( std::string &, int) > lsig; > > rsig.connect( a); > lsig.connect( a); > > rsig(0,"called from right side"); > lsig("called from left side",0); > > Because a is a smart_ptr this code would not compile because > operator() is not a member function. If I dereference the smart > pointer when I connect to signal, changes of a are never know > by the other signal instance. > Maybe I can use a proxy class which aggregates the smart pointer > but maybe there is a better solution?! As a general rule, functors are expected to be copiable and not to have mutable state of their own. So how about putting the shared_ptr inside A instead: class A { private: boost::shared_ptr<int> _i; public: A() : _i(new int(0)) {} void operator()(int, std::string & s) { *_i += 1; } void operator()(std::string & s, int) { *_i += 2; } }; A a;

1 0

regex bug (regex_search throws bad_expression)
by Adam Molnar 26 Jul '04

26 Jul '04

hi! using the latest regex patch and vc7.1 i've accidentally encountered the following: ... regex re("([^\n]*\\n+\\s+)+NEEDEDSUBITEM2:[^\\s]"); bool matched = regex_search(text, re); // bad_expression ... i figured out those \\s -s may cause the problem. sample text attached. and one question: having "DATA.*?ITEM1(ITEM2)?" and an input like "DATA ITEM1 ITEM1ITEM2" should ITEM2 be extracted? i think it would be good to make a note on this case in the doc. yours, adam begin 666 match_test.txt M#0I)5$5-5%E014$Z#0H@(" @4U5"251%33$Z('AX> T*#0H@(" @3D5%1$5$ M4U5"251%33$Z(" @('AX>'@-"B @("!.145$141354))5$5-.B @("!X>'AX M#0H-"DE414U465!%0CH-"B @("!354))5$5-,3H@>'AX#0H-"B @("!.145$ M141354))5$5-,3H@(" @>'AX> T*(" @($Y%141%1%-50DE414TR.B @("!X '>'AX#0H-"@`` ` end

3 4

Building boost::fs to a dll
by boost＠dalvander.com 26 Jul '04

26 Jul '04

Is it possible to build boost::filesystem to a dll instead of creating a static linked library? // Anders Dalvander

2 1