November 2008 - Boost - lists.preview.boost.org

Re: [boost] [unicode] Interest Check / Proof of Concept
by Kasra 20 Nov '08

20 Nov '08

Hi guys, I have been reading this thread with enthusiasm. I like the idea of unicode string class. However, since unicode is rather different from the old fashioned ANSI string literals some interfaces might need to be changed. Maybe what I have to say could help you in one form or the other: \+ typedef std::basic_string<uint32_t> unicode_string; this string class will have all of the normal std::string interface as all of the unicode encoding is mapped within a 32-bit integral type. we could then have a class called encoding; such that it allows for static conversion from unicode_string type to utf8, utf16, etc. for example: \+ unicode_string s = U"Hello Wolrd!"; \+ utf8_string utf8 = encoding::utf32::to_utf8(s); // also we could have a implicit conversion such that utf8 could // perform the conversion from utf32 string I personally like this approach a lot. Since when you think about it the interface does not changed we could still have efficient iterators and insertion, at the same time whenever there is a need for memory compact encoding we convert to the desired encoding as required. So in a sense what I am saying is keep the encoding using utf32 for performance on memory however, when we want to export the encoding (to screen, file and etc) convert to the desired encoding. If there is any interest for this approach drop me a line and I will send you the interface which I have for this. With best regards Kasra

1 0

[function] minor doc issues
by Thorsten Ottosen 20 Nov '08

20 Nov '08

Hi, Reading http://www.boost.org/doc/libs/1_36_0/doc/html/function/reference.html it says: A function object f is compatible if for the given set of argument types Arg1, Arg2, ..., ArgN and a return type ResultType, the appropriate following function is well-formed: // if ResultType is not void ResultType foo(Arg1 arg1, Arg2 arg2, ..., ArgN argN) { return f(arg1, arg2, ..., argN); } // if ResultType is void ResultType foo(Arg1 arg1, Arg2 arg2, ..., ArgN argN) { f(arg1, arg2, ..., argN); } I believe the "If ResultType is void" is not needed. We can return void in C++. -Thorsten

2 5

[Serialization] (and [TypeTraits]?) Overloaded new operator are not called from serialization
by Runar Undheim 20 Nov '08

20 Nov '08

Could someone help to update the serialization code in boost to call the overloaded new operator, please? There already exist some code, but the code is not active because of a #if 0 statement. And if I set the code to #if 1 the code is not compiling with Visual C++ 2005. So I have updated the heap_allocator implementation in archive\detail\iserializer.hpp to support overload of new operator, but I'm not sure if it will work on all compilers. I have just tested the code with Visual C++ 2005. First the code that I used to check if the new operator is overloaded. If there are compilers that don't support the code, the implementation could be updated using defines and only return false. This way we take advantage of the information on supported compilers, an the code will work as before on unsupported compilers. Should the code that check if it has a new operator be part of the Type Traits library? typedef char True; // sizeof(True) == 1 typedef struct { char a[2]; } False; // sizeof(False) > 1 template <class U, U x> struct test; template <typename C> const True *hasNewOperator(C*, test<void *(*)(size_t),&C::operator new>* = NULL) { // Function must be implemented on Visual C++ 2005 to prevent linking error return NULL; } template <typename R> const False *hasNewOperator(...) { // Function must be implemented on Visual C++ 2005 to prevent linking error return NULL; } //#define hasNewOperator(f) (sizeof(*hasNewOperator<f>(static_cast<f*>(NULL))) == sizeof(True)) Then the heap_allocator is updated to use the hasNewOperator: template<class T> struct heap_allocator { template<class U> static T* new_operator(U); static T* new_operator(const True*) { return static_cast<T *>((T::operator new)(sizeof(T))); } static T* new_operator(const False*) { return static_cast<T *>(operator new(sizeof(T))); } static T * invoke(){ return new_operator(hasNewOperator<T>(static_cast<T *>(NULL))); } }; regards, Runar

2 2

Boost and Borland C++ Compiler 5.5
by Marcin Kurek 20 Nov '08

20 Nov '08

Hello everybody, I faced a problem when upgrading a piece of software originally written using Borland C++ Builder 5. Because there are some configuration files I need to parse I thought it might be a good idea to use Boost Regex. However, I've encountered many errors when I tried to build it. I've downloaded 1.37.0 release and followed instructions in "Getting started on Windows": E:\boost_1_37_0\bjam.exe --build-dir E:\boost-build --toolset=borland As result I get multiple errors: "bcc32.exe" -j5 -g255 -q -c -P -Ve -Vx -a8 -b- -O2 -tWC -tWR -tWC -WM- -DBOOST_ALL_NO_LIB=1 -DNDEBUG -I"." -I"C:\Program Files\Borland\CBuilder5/include/" build\borland\release\link-static\copysignf.obj" "libs\math\build\..\src\tr1\copysignf.cpp" ...failed borland.compile.c++ e:\boost-build\boost\bin.v2\libs\math\build\borland\release\link-static\copysignf.obj... borland.compile.c++ e:\boost-build\boost\bin.v2\libs\math\build\borland\release\link-static\erfcf.obj libs\math\build\..\src\tr1\erfcf.cpp: Error E2176 .\boost/math/tr1.hpp 179: Too many types in declaration Error E2176 .\boost/math/tr1.hpp 180: Too many types in declaration Error E2176 .\boost/math/tr1.hpp 181: Too many types in declaration I've downloaded BCC 5.5.1 from Borland, but the effect was the same. I also was trying some older releases of Boost, including 1.32.0 and 1.33.1. E:\boost_1_33_1\bjam.exe -sTOOLS=borland install "bcc32" -j5 -g255 -q -c -P -w -Ve -Vx -a8 -b- -DNDEBUG -DNDEBUG -DBOOST_IOSTREAMS_DYN_LINK=1 -O2 -vi -w-inl -tWR -tWD -WM- -w-8001 -I"bin\boost\libs\iostreams\build" -I"E:\boost_1_33_1" -I"E:\bo ost_1_33_1" -o"bin\boost\libs\iostreams\build\boost_iostreams.dll\borland\release\mapped_file.obj" "E:\boost_1_33_1\libs\iostreams\build\..\src\mapped_file.cpp" (...) borland-C++-action bin\boost\libs\iostreams\build\boost_iostreams.dll\borland\release\threading-multi\mapped_file.obj E:\boost_1_33_1\libs\iostreams\build\..\src\mapped_file.cpp: Error E2401 E:\boost_1_33_1\boost/iostreams/detail/is_iterator_range.hpp 20: Invalid template argument list Error E2303 E:\boost_1_33_1\boost/iostreams/detail/is_iterator_range.hpp 20: Type name expected Error E2238 E:\boost_1_33_1\boost/range/iterator.hpp 102: Multiple declaration for 'range_iterator<char *>' Error E2344 E:\boost_1_33_1\boost/range/iterator.hpp 77: Earlier declaration of 'range_iterator<char *>' Error E2238 E:\boost_1_33_1\boost/range/iterator.hpp 108: Multiple declaration for 'range_iterator<wchar_t *>' Error E2228 E:\boost_1_33_1\boost/range/iterator.hpp 108: Too many error or warning messages I don't have an idea what's wrong. I was googling a lot, but after 6 hours I guess I'll give up for today. Maybe there's something obvious I don't see, could anybody point me in the right direction, please? Thanks a lot, Marcin

2 1

limit symbol names of fusion::vector
by Christian Holmquist 19 Nov '08

19 Nov '08

Hi, I'm using fusion::result_of::as_vector<some_mpl_transform_view>::type quite frequently, but having troubles with extremely long type names (our vc8.pdb are close to 0.5 Gb). Is there any way in fusion like in mpl, where one can do mpl::vector<a, b, c>::type ^^ which gives the shorter name mpl::vector3<a, b, c> instead of mpl::vector<a, b, c, na, na, na, ...> In essence, can I have fusion::vector2<int, float> from the following: fusion::result_of::as_vector<mpl::vector<int, float>::type >::type Thanks for a great lib! / Christian

3 6

BOOST_FOREACH slow?
by Hansi 19 Nov '08

19 Nov '08

Hello, I have made a few tests with the performance of BOOST_FOREACH. The results are for me unexpected slow...may be one has an idea if I make something wrong... vector[int]: index:0.00167619 vector[int]: iterator:0.00167619 vector[int]: BOOST_FOREACH:9.40036 list[int]: iterator:239.669 list[int]: BOOST_FOREACH:233.219 vector[TestData]: index:0.00167619 vector[TestData]: iterator:0.00167619 vector[TestData]: BOOST_FOREACH:9.40008 list[TestData]: iterator:233.827 list[TestData]: BOOST_FOREACH:238.513 testArray[int]: index:0.00167619 testArray[int]: BOOST_FOREACH:9.40008 Best regards Hansjörg the test program was: // ForEach.cpp : Defines the entry point for the console application. // #include "stdafx.h" #include "..\include\PerformanceCounter.h" #include <vector> #include <boost\foreach.hpp> #include <iostream> #include <list> using namespace std; using namespace Microtec; class TestData { public: TestData(int val1, int val2, int val3):var1(val1),var2(val2),var3(val3){} int var1; int var2; int var3; }; int testArray[10000000]; int _tmain(int argc, _TCHAR* argv[]) { vector<int> data; list<int> lstData; vector<TestData> vectorTestData; list<TestData> lstTestData; for(int i = 0; i < 10000000;i++) { data.push_back(i); lstData.push_back(i); vectorTestData.push_back(TestData(i,i,i)); lstTestData.push_back(TestData(i,i,i)); } int res = 0; const int actSize = (int)data.size(); for(int i = 0; i < actSize; i++) { res += data[i]; } //takes 0.0017 ms res = 0; for(vector<int>::iterator it = data.begin(); it != data.end(); it++) { res += *it; } //takes 0.0017 ms res = 0; BOOST_FOREACH(int val,data) { res += val; } //takes 9.400 ms //list res = 0; for(list<int>::iterator it = lstData.begin(); it != lstData.end(); it++) { res += *it; } //takes 239.669 ms res = 0; BOOST_FOREACH(int val,lstData) { res += val; } //takes 233.219 ms //vector TestData for(int i = 0; i < actSize; i++) { res += vectorTestData[i].var1; } //takes 0.0017 ms res = 0; for(vector<TestData>::iterator it = vectorTestData.begin(); it != vectorTestData.end(); it++) { res += it->var1; } //takes 0.0017 ms res = 0; BOOST_FOREACH(TestData& val,vectorTestData) { res += val.var1; } //takes 9.400 ms //list res = 0; for(list<TestData>::iterator it = lstTestData.begin(); it != lstTestData.end(); it++) { res += it->var1; } //takes 233.827 ms res = 0; BOOST_FOREACH(TestData& val,lstTestData) { res += val.var1; } //takes 238.827 ms for(int i = 0; i < actSize; i++) { res += testArray[i]; } //takes 0.0017 ms res = 0; BOOST_FOREACH(int val,testArray) { res += val; } //takes 9.400 ms return 0; }

9 25

[asio] Removing gcc warning
by Richard Smith 19 Nov '08

19 Nov '08

The function boost::asio::detail::pipe_select_interrupter::interupt() in <boost/asio/detail/pipe_select_interrupter.hpp> produces the following warning in gcc 4.3.2 / glibc 2.8 when compiled with -O1 or higher. The warning occurs even without any -W options. ./boost/asio/detail/pipe_select_interrupter.hpp:75: warning: ignoring return value of ‘ssize_t write(int, const void*, size_t)’, declared with attribute warn_unused_result As the message implies, gcc issues this warning whenever a function declared with __attribute__((warn_unused_result)); it appears that gcc implicitly places this attribute on ::write as the glibc header doesn't contain it. A short test case is: #include <unistd.h> int main() { ::write(1, 0, 0); } To suppress the warning, the function interupt() function can be altered as follows: #include <boost/concept_check.hpp> void interrupt() { char byte = 0; ignore_unused_variable_warning( ::write(write_descriptor_, &byte, 1) ); } Is there any chance of making a change along these lines? Apologies if this is a know issue: a quick search of the mailing list didn't turn anything up. -- Richard Smith

1 0

Boost.Build causing failures on the release branch (and Trunk too)
by John Maddock 19 Nov '08

19 Nov '08

There seem to be a lot of failures on the release branch caused by some libraries not being found, so for example on acc: http://tinyurl.com/6r76o5 or on Darwin: http://tinyurl.com/6dyezn The library not being found is part of the project requirements BTW in case that helps, Regards, John.

1 1

[signals2][review] The review of the signals2 library (formerly thread_safe_signals) begins today, Nov 1st
by Stjepan Rajko 19 Nov '08

19 Nov '08

Hello all, The review for the Signals2 library (formerly known as thread_safe_signals) submitted by Frank Mori Hess begins today (Nov 1st) and is scheduled to end on Nov 10th. I would like to thank Franz Alt, Terry Golubiewski, Doug Gregor, Ravikiran Rajagopal and Andrew Webber for making this review possible by committing to reviewing the library. How to submit a review: -------- As usual, EVERYONE is welcome to participate in the review discussions and to submit a review. I strongly encourage participation from reviewers that would examine the library from a purely user standpoint (commenting on the interface and / or the documentation), as well as reviewers that would be willing to look into the details of the implementation (i.e., you don't have to focus on both). Here are some questions you might want to answer in your review (feel free to skip those that don't apply to your analysis): * What is your evaluation of the design? * What is your evaluation of the implementation? * What is your evaluation of the documentation? * What is your evaluation of the potential usefulness of the library? * Did you try to use the library? With what compiler? Did you have any problems? * How much effort did you put into your evaluation? A glance? A quick reading? In-depth study? * Are you knowledgeable about the problem domain? And finally, every review should answer this question: * Do you think the library should be accepted as a Boost library? Be sure to say this explicitly so that your other comments don't obscure your overall opinion. Please submit your review either to the list or privately to me by the end of the review period. If you send me a review privately I will forward it to the list so the review can be discussed. If you would prefer me to forward your review anonymously (with your name removed) please indicate that in your e-mail. If you are a first time reviewer, here is some more information to get you started: * http://www.boost.org/community/reviews.html has more information on the review process * the review officially takes place on both the boost dev and boost-user lists, but typically more of the discussion happens on the boost dev list * if you want to get a feel for past reviews, you can find them in the archives (see http://www.boost.org/community/groups.html) - the past review dates are on the review schedule (http://www.boost.org/community/review_schedule.html) Typically, there is a period of discussion between the reviewers and the author, with official reviews coming in towards the tail end of the review period. About the library: -------- * The library can be downloaded from: http://www.boostpro.com/vault/index.php?&directory=thread_safe_signals (latest version is signals2-2008-10-08.zip) * Documentation is also available online: http://www.comedi.org/projects/signals2/libs/signals2/doc/html/index.html * A synopsis: The Boost.Signals2 library (formerly known as thread_safe_signals) is an implementation of a managed signals and slots system. Signals represent callbacks with multiple targets, and are also called publishers or events in similar systems. Signals are connected to some set of slots, which are callback receivers (also called event targets or subscribers), which are called when the signal is "emitted." Signals and slots are managed, in that signals and slots (or, more properly, objects that occur as part of the slots) can track connections and are capable of automatically disconnecting signal/slot connections when either is destroyed. This enables the user to make signal/slot connections without expending a great effort to manage the lifetimes of those connections with regard to the lifetimes of all objects involved. When signals are connected to multiple slots, there is a question regarding the relationship between the return values of the slots and the return value of the signals. Boost.Signals2 allows the user to specify the manner in which multiple return values are combined. * Relationship to Boost.Signals: This is a thread-safe variant of the original Boost.Signals library. There have been some changes to the interface to support thread-safety, mostly with respect to automatic connection management. The following thread offers some more details on the differences between the two implementations, as well as a plan of a phased replacement of Boost.Signals should Signals2 be accepted: http://tinyurl.com/4sqau3 [nabble] The Signals2 FAQ also covers some differences: http://tinyurl.com/576gl6 Kind regards, Stjepan (review manager)

7 15

task_array<n> and variadic templates
by vicente.botet 19 Nov '08

19 Nov '08

Hi, I want to create an static task_array<n> which must be initiaized with n functors. In pseudo code: template <std::size N> struct task_array { template <typename F_1, ..., typename F_N> task_array(F_1 f1, ..., F_N fN); }; My current implementation with N=1..3 is like that template <std::size N> struct task_array; template <> struct task_array<1>; { template <typename F_1,> task_array(F_1 f1); }; template <> struct task_array<2>; { template <typename F_1, typename F_2> task_array(F_1 f1, F_2 f2); }; template <> struct task_array<3>; { template <typename F_1, typename F_2, typename F_3> task_array(F_1 f1, F_2 f2, F_3 f3); }; I can use the Boost.Preprocessor library to generate this up to a maximum know at compile time, but can I define this class with the help of variadic templates? If yes how? Thanks in advance, Vicente

3 6