February 2006 - Boost-users

Crossbuild, on linux using mingw32 for win32
by Rafał Maj Raf256 05 Feb '06

05 Feb '06

What is the best way to crossbuild on linux, using mingw32, to win32? I dont see no description about such configuration on http://boost.org/more/getting_started.html mingw executable is: i586-mingw32msvc-g++

2 3

[Review] Fixed strings library
by Hartmut Kaiser 05 Feb '06

05 Feb '06

Hi all, the formal review period of the Fixed strings library written by Reece Dunn ends today. Since there was complete consent with regard to this library among all submitted reviews, it is straight forward for me to make the final decision. The library is rejected and will not be included into Boost at this point in time. The main points listed during review were: - Do we need a fixed_string library at all, is it worth the overhead? - Insufficient/confusing documentation . insufficient description of general design, design rationales and motivations . it says more about implementation details rather then the public interface . missing or insufficient documentation of parts of the library - The overall architecture of the implementation is too complicated and needs to be reworked - Missing interface for construction/interaction with std::string's - Unclear benefit to have the strings truncated to avoid overflows - Missing performance comparison with std::string and other alternatives - Compilation problems with the library as submitted for review Reece already expressed his interest to resubmit the library for review again, after he has been able to address the mention issues. I'ld like to thank Reece for submitting this library and all reviewers for their time they invested in preparing the reviews. Regards Hartmut Review Manager

1 0

Re: [Boost-users] boost::pp application
by Marco Correia 05 Feb '06

05 Feb '06

Thanks! Today I managed to get a little bit further, let me tell you where I am. First I aliased every symbol of the alphabet as an integer. #define A 1 #define B 2 ... Then I guess there are two options: 1) #define NOT_EQ(s,E1,E2) \ BOOST_PP_NOT_EQUAL(E1,E2) #define REMOVE_ALL(Elem,Seq) \ BOOST_PP_SEQ_FILTER(NOT_EQ, Elem, Seq) #define REMOVE_ELEM(s,Seq,Elem) REMOVE_ALL(Elem,Seq)(Elem) #define REMOVE_DUPLICATES(Seq) \ BOOST_PP_SEQ_FOLD_LEFT(REMOVE_ELEM, Seq, Seq) This starts with the full sequence, and for each element, removes all occurences and add it in the end. 2) The other option is similar but instead of starting with the full sequence and remove each element, it starts with an empty sequence and adds each element if it is not already there (kind of insertion sort but without sorting). I think the solution 1) is faster, but I confess it's only an intuition :) Do you see any other way, more efficient than this? Thanks a lot for trying to help! Marco On Friday 03 February 2006 22:48, you wrote: > Hi Marco, > > I just say this post now, so I'll get back to it after I get home from > work. I the pp-lib can help you with this. > > Regards, > Paul Mensonides > > > -----Original Message----- > > From: boost-users-bounces(a)lists.boost.org > > [mailto:boost-users-bounces@lists.boost.org] On Behalf Of > > Marco Correia > > > > I would like to ask if there is a way boost::preprocessor may > > be used to solve the following problem: > > > > The problem is to define a macro that expands to the > > concatenation of their arguments, without duplicates. The > > arguments are symbols belonging to a known alphabet. -- Marco Correia <mvc(a)netcabo.pt>

2 2

<cassert> in the boost libraries
by Joerg Ruedenauer 05 Feb '06

05 Feb '06

Hello, we use our own assertion handler and assert macro in our code, defined in a header which gets included at the end of our precompiled header file. Unfortunately, many boost libraries undermine this because they include the cassert header, which redefines the assert macro. Currently we work around this by putting some boost libraries in the precompiled header and including our own assert header again in each source file which includes boost headers, but this is a clumsy and error-prone solution. Boost already nominately provides a solution through BOOST_ENABLE_ASSERT_HANDLER, but it's just not used consequently throughout the libraries. Could the authors of the libraries (or the reviewers) please grep through the code and change the cassert (or assert.h) includes to boost/assert.hpp, and consequently, change assert(...) to BOOST_ASSERT(...)? I've found the inclusion particularly often in the serialization, graph, iostream and random libraries, but also in regex, signals, ptr_container and many other headers. Regards, Joerg Ruedenauer

3 4

Re: [Boost-users] Boost-users Digest, Vol 807, Issue 1
by Drumheller, Michael 05 Feb '06

05 Feb '06

>>> ------------------------------ >>> >>> Message: 9 >>> Date: Sat, 04 Feb 2006 10:30:25 +0000 >>> From: Jim Douglas <jim(a)dramatec.co.uk> >>> Subject: Re: [Boost-users] Getting started with Boost: what did you >>> need to know? >>> To: boost-users(a)lists.boost.org >>> Message-ID: <ds1vo5$ue6$1(a)sea.gmane.org> >>> Content-Type: text/plain; charset=ISO-8859-1; format=flowed >>> >>> Drumheller, Michael wrote: >>> > Main stumbling point for me was bjam. It's not so much >>> of a problem for <..snip...> >>> Perhaps people don't complain because most only use bjam it >>> to build the >>> libraries (archives and shared objects) and then (like me) they use >>> 'make' to build the projects that incorporate Boost. I >>> would guess that you are in the minority if you write in-house extensions to >>> the Boost library. Sorry about that, I was being very unclear. I did not mean in-house extensions to the Boost library, I meant C++ extensions to Python, using Boost.Python. The Boost.Python tutorial contains an extremely ominous warning, in a big yellow box, here: <<http://www.boost.org/libs/python/doc/tutorial/doc/html/python/hello.ht ml>>, It says, in so many words, that you're a masochist if you try to build your Boost.Python extensions with anything but bjam. >>> Having said that, I can appreciate that bjam is a *very* >>> powerful tool and has many advantages over make, but I just can't find >>> the time to study the docs enough to make a trial switch. >>> >>> Jim Don't get me wrong, I can't stand the error-prone, undebuggable dinosaur that is make, and I would never advocate it for a new project. I too can appreciate that bjam seems very powerful. And elegant, in its own way. It also seems very fast. But with a monster learning curve. That's not necessarily a downer--you don't get something for nothing, and I actually *like* learning new tools, so I was happy to see the "Abandon hope, all ye who <would use a tool other than bjam>" warning on the gates of Boost.Python (yellow box above), because I thought, well, there's my motivation to learn bjam--maybe bjam will be my new default build tool(!) I read all the documents I could find, but man, it's like learning Turkish or Vietnamese or something. You need a teacher, and there are just zero teachers in my org. Alas, the bjam files in Boost themselves (those that are used to build the Boost libraries) are (or at least seem) *extremely* sohpisticated, and are woefully uncommented--in fact, absolutely *un*commented if you define a comment as something with tutorial value, e.g., "Here is an excellent example of a rule to do such-and-such." Personally, I was unable to get past the knee of the curve given my other commitments. Again, maybe it's just me. I'm curious: How many Boost users out there use bjam as the default build tool for their day-to-day programming? I recall hearing at one time that there was a movement afoot, at one time at least, to switch the Boost build process over to using scons. Any truth to that? Thanks, MD

2 1

Re: [Boost-users] <cassert> in the boost libraries
by Merrill Cornish 05 Feb '06

05 Feb '06

Include files such as <cassert> usually have a guard involving the #define of a specific symbol to prevent then from being included twice. Once you have defined your own version of assert(), perhaps you could #define the <cassert> guard symbol as a way of preventing the local <cassert> from being included. This would work, of course, only if your version of <cassert> would work for Boost also. Merrill

2 1

Problems Compiling in VS 2005
by Sergey Tiraspolsky 05 Feb '06

05 Feb '06

I'm attempting to upgrade my solution to 2005 from 2003 and it compiles fine until it gets to boost then I get errors. Here's the error: Error 889 error C2665: 'boost::python::detail::make_getter' : none of the 3 overloads could convert all the argument types ...\boost-1.32.0\include\boost\python\data_members.hpp 277 I thought it may be related to having an older version of Boost so I upgraded to 1.33.1 and am still having the same error: boost-1.33.1\include\boost\python\data_members.hpp(277) : error C2665: 'boost::python::detail::make_getter' : none of the 3 overloads could convert all the argument types Any ideas? --------------------------------- Brings words and photos together (easily) with PhotoMail - it's free and works with Yahoo! Mail.

3 2

please help with cross building linux->win32 using mingw
by Rafał Maj Raf256 04 Feb '06

04 Feb '06

Hi, I have had a terrible time when trying to crossbuild. Is it really not possible with boost? I want to build, on linux, boost for win32 programs (to crossbuild programs using boost), using mingw g++ compiler. It worked fine with other libraries, usually just CXX=i586-mingw-msvc-g++ && ./configure --prefix=/usr/local/win32/ && make install but it do not work with boost / bjam :-/ I even got wine installed and installed MinGW 5 and MSys there, the MSys didnt worked (probably wine windows emulator fault), as for MinGW bjam have problems with some shell commands like it tries to call MkDir1, raf256@lore:~/.wine/drive_c/boost_1_33_1$ wine ../bjam/bjam.exe Building Boost.Regex with the optional Unicode/ICU support disabled. Please refer to the Boost.Regex documentation for more information (and if you don't know what ICU is then you probably don't need it). ****************************************************** Building Boost.Iostreams with bzip2 support disabled. To enable bzip2, consult the Boost.Iostreams documentation ****************************************************** ****************************************************** Building Boost.Iostreams with zlib and gzip support disabled. To enable zlib and gzip, consult the Boost.Iostreams documentation ****************************************************** --------------------------------------------------------------------- *** If you don't need Boost.Python, you can ignore this section *** *** pass --without-python to suppress this message in the future *** skipping Boost.Python library build due to missing or incorrect configuration couldn't find Python.h in "c:/Python24/include" You can configure the location of your python installation by setting: PYTHON_VERSION - The 2-part python Major.Minor version number (e.g. "2.2", NOT "2.2.1") - currently "2.4" PYTHON_ROOT - automatically configured from PYTHON_VERSION if not otherwise set ; currently "c:/Python24" The following are automatically configured from PYTHON_ROOT if not otherwise set: PYTHON_LIB_PATH - path to Python library object; currently "c:/Python24/libs" --------------------------------------------------------------------- ...patience... ...patience... ...found 9007 targets... ...updating 1363 targets... MkDir1 bin spawn: No such file or directory bjam with -d+4 ends with: PYTHON_LIB_PATH - path to Python library object; currently "c:/Python24/libs" --------------------------------------------------------------------- Executing through .bat file Executing command [cmd.exe] [/Q/C] [c:\windows\temp\jam8-00.bat] spawn: No such file or directory and his bat contains only mkdir "bin" ... $ cat ../windows/temp/jam8-00.bat mkdir "bin" btw. the moderation system of NNTP access is horrible, it took 3 days to send this meesage, after several email (register, verify, send, re-verify, and so on) it really is not too wellcomming for new members of community, IMHO, perhaps such strong anti-spam security is not needed.

1 0

[regex] - bk_parens and bk_braces missing from 1.33.x
by Keith MacDonald 04 Feb '06

04 Feb '06

It states in history.html for regex 1.33.0 that "some of the more esoteric options have now been removed, so there is the possibility that existing code may fail to compile: however equivalent functionality should still be available". I suppose bk_parens and bk_braces could be considered esoteric, but I cannot find any equivalent functionality. Is that intentional? Thanks, Keith MacDonald

1 0

[date_time] local_time_facet to format strings
by Christian Henning 04 Feb '06

04 Feb '06

Hi there, I'm trying to format a time string based on a local_time_facet. For some reason it isn't working as expected. Here is the code (I basically copied from the Date Time IO Tutorial): ptime t( microsec_clock::local_time() ); std::stringstream ss; local_time_facet* output_facet = new local_time_facet(); ss.imbue( std::locale( std::locale::classic(), output_facet)); output_facet->format("%a %b %d, %H:%M %z"); ss.str(""); ss << t; std::cout << ss.str() << std::endl; The output isn't formated in the way it's specified. Cout will basically spit out the same result as to_simple_string(). 2006-Feb-03 16:38:57.046875 The expected result would something like this: Fri Feb 3, 16:38 EDT I'm using MSVC 7.1 and Boost 1.33.1. Thanks in advance, Christian

2 2