January 2009 - Boost-users - lists.preview.boost.org

Boost.MPL: Is it possible to apply a variable substitution without evaluating the lambda-expression?
by Luke Simon 27 Jan '09

27 Jan '09

I am using BOOST_MPL_ASSERT to assert a lambda-expression applied to a sequence of types by folding the lambda-expression onto a sequence of types (see code and continue discussion below). template<class predicate, class argument> struct mpl_assert { typedef typename apply<predicate, argument>::type type; BOOST_MPL_ASSERT((type)); }; template<class pred, class xs> struct assert_for_all : foldr< mpl_assert<typename lambda<pred>::type, _>, true_, xs > {}; The foldr metafunction is a strict version of Haskell's foldr, i.e., MPL's reverse_fold with parameters passed in the order that Haskell's foldr passes parameters. The distinction isn't important for this discussion, so don't dwell on that detail. My intent is that if the lambda-expression is false for one of the types in the sequence, then the lambda-expression applied to the "bad" argument is printed in a compile-time error. However, what gets printed is the result of the lambda-expression applied to the bad argument: integral_constant<bool, false>, which is very uninformative. If the definition of mpl_assert is changed to the following: template<class predicate, class argument> struct mpl_assert { typedef typename apply<predicate, argument>::type type; BOOST_MPL_ASSERT((apply<predicate, argument>)); }; ...then what gets printed is the unevaluated application of a lambda-expression to the bad argument. While this is more informative, it is still too messy because the type is a lambda-expression containing placeholders. The most desirable type to have printed would be the lambda-expression after placeholder substitution, but before further evaluation. For example, imagine: typedef assert_for_all<is_same<foo, _>, boost::mpl::vector<foo, bar> >::type type; ...the first definition of mpl_assert prints the following uninformative type in the displayed error: boost::integral_constant<bool, false> ...the second definition of mpl_assert prints the more descriptive but overly messy type: boost::mpl::apply<boost::mpl::protect<boost::mpl::bind2<boost::mpl::quote2<boost::is_same, mpl_::void_>, foo, mpl_::arg<-0x000000001> >, 0>, bar, mpl_::na, mpl_::na, mpl_::na, mpl_::na> ...but what I want to have printed is the type: boost::is_same<foo, bar> Boost.MPL is already calculating such a type, because apply<is_same<foo, _>, bar> returns is_same<foo, bar>::type, so the needed type is there... but does Boost.MPL provide a means of accessing it as opposed to accessing its ::type field?

2 1

[boost.bind] How to retrieve the function type of a boost::bind result
by Olaf Krzikalla 27 Jan '09

27 Jan '09

Hi, The following snippet won't compile since the compiler can't deduce the template argument: template<typename T> void foo(boost::function<T> func) { /*...*/ } void bar() {} void baz() { foo(boost::bind(&bar)); // error foo<void()>(boost::bind(&bar)); // OK foo(boost::function<void()>(boost::bind(&bar))); // OK } So how do I get the first line of baz to work? That is, is there a generic way to compute the boost::function type from the bind result? Background code follows here: template<typename T> T getArg(int i) { /*...*/ } template<typename T> struct arity_dispatch; template<> struct arity_dispatch< boost::function<void()> > { static void call(const boost::function<void()>& f) { f(); } }; template<class T> struct arity_dispatch< boost::function<void(T)> > { static void call(const boost::function<void(T)>& f) { f(getArg<T>(0)); } }; template<class T, class U> struct arity_dispatch< boost::function<void(T, U)> > { static void call(const boost::function<void(T, U)>& f) { f(getArg<T>(0), getArg<T>(1)); } }; //aso. template<typename T> void generic_call(boost::function<T> func) { arity_dispatch< boost::function<T> >::call(func); } Neither is it possible for me to call "generic_call" with a boost::bind result nor is it possible to (partial) specialize for the _bi::_bind_t types as these types are internal ones. Also it seems that I can't even statically compute the arity of a boost::bind result (which would indeed help a lot). Any suggestions? If not I propose the introduction of a type function_type in the return type of a bind expression (I already needed it once before). Thanks in advance Olaf Krzikalla

3 5

Maximum Size of Matrix in Boost
by Gundala Viswanath 27 Jan '09

27 Jan '09

Hi, I have a matrix withn 10^7 * 10^7 rows, each with double value like 0.333. All this printed in text file. I want to slurp all of them in Boost Matrix, with 6GB RAM, can it handle such matrix? - Gundala Viswanath Jakarta - Indonesia

5 5

[boost.build] no best alternative error
by Sandeep Gupta 27 Jan '09

27 Jan '09

Hi, I am missing something obvious but i am at lost extending bjam. I want to chain two rules Rule 1 : generate t.gp --> x.mp Rule 2 : x.mp --> x.mps My original attempts failed (discussed here http://tinyurl.com/dg7hle)*. * I would greatly appreciate any suggestions. thanks sandeep

1 0

Re: [Boost-users] Looking for a book on graph algorithms
by Roger House 27 Jan '09

27 Jan '09

Thanks to everyone who replied to my request for pointers to books, etc., on graph algorithms. I now have several books on order. Roger House

3 2

Using Boost.Test on 64-bit OSX
by Michael Dickey 27 Jan '09

27 Jan '09

I can't seem to get Boost's unit test framework to work on OS X Leopard (10.5) when using "-arch x86_64" for 64-bit builds (note: the programs I build using boost seem to work fine, but the unit tests blow up). I have Xcode 3.1.2 installed, and the results are the same if I'm using gcc 4.0 or 4.2. I used the 32_64.diff patch from trac to make 4-way Universal boost libraries (https://svn.boost.org/trac/boost/ticket/1679 ). Any ideas would be greatly appreciated =) Thanks, -Mike Without "-arch x86_64" (defaults to 32-bit intel): Mikes-Mac-Pro:~ mdickey$ g++ -DBOOST_TEST_DYN_LINK -I/usr/local/ include/boost-1_37 -L/usr/local/lib -lboost_unit_Test_framework-xgcc42- mt -o test test.cpp Mikes-Mac-Pro:~ mdickey$ ./test Running 5 test cases... *** No errors detected With "-arch x86_64": Mikes-Mac-Pro:~ mdickey$ g++ -arch x86_64 -DBOOST_TEST_DYN_LINK -I/usr/ local/include/boost-1_37 -L/usr/local/lib -lboost_unit_Test_framework- xgcc42-mt -o test test.cpp Mikes-Mac-Pro:~ mdickey$ ./test --log_level=all Running 5 test cases... Entering test suite "simple-unit-test" Entering test suite "SimpleTests_S" Entering test case "test1" Test case test1 doesn't include any assertions Leaving test case "test1"; testing time: 22mks Entering test case "test2" Test case test2 doesn't include any assertions Leaving test case "test2"; testing time: 15mks Entering test case "test3" Test case test3 doesn't include any assertions Leaving test case "test3"; testing time: 16mks Entering test case "test4" Test case test4 doesn't include any assertions Leaving test case "test4"; testing time: 15mks Entering test case "test5" Test case test5 doesn't include any assertions Leaving test case "test5"; testing time: 14mks Leaving test suite "SimpleTests_S" Leaving test suite "simple-unit-test" *** No errors detected test(93789) malloc: *** error for object 0x30001001004d0: Non-aligned pointer being freed (2) *** set a breakpoint in malloc_error_break to debug test(93789) malloc: *** error for object 0x30001001005d0: Non-aligned pointer being freed (2) *** set a breakpoint in malloc_error_break to debug test(93789) malloc: *** error for object 0x30001001006c0: Non-aligned pointer being freed (2) *** set a breakpoint in malloc_error_break to debug test(93789) malloc: *** error for object 0x30001001007c0: Non-aligned pointer being freed (2) *** set a breakpoint in malloc_error_break to debug gdb backtrace to malloc_error_break (note above that it's happening after completion of all tests): #0 0x00007fff810a9a11 in malloc_error_break () #1 0x00007fff810a4a90 in szone_error () #2 0x00000001000156ec in ~framework_impl [inlined] () at impl/ framework.ipp:133 #3 0x00000001000156ec in __tcf_1 () at impl/framework.ipp:222 #4 0x00007fff8100131c in __cxa_finalize () #5 0x00007fff81001232 in exit () #6 0x0000000100001057 in start () Confirm test program is a 64-bit executable: Mikes-Mac-Pro:~ mdickey$ file test test: Mach-O 64-bit executable x86_64 Dynamic linkage of test program: Mikes-Mac-Pro:~ mdickey$ otool -L test test: libboost_unit_test_framework-xgcc42-mt-1_37.dylib (compatibility version 0.0.0, current version 0.0.0) /usr/lib/libstdc++.6.dylib (compatibility version 7.0.0, current version 7.4.0) /usr/lib/libgcc_s.1.dylib (compatibility version 1.0.0, current version 1.0.0) /usr/lib/libSystem.B.dylib (compatibility version 1.0.0, current version 111.1.3) Confirm Universal/FAT build of Boost libraries: Mikes-Mac-Pro:~ mdickey$ file /usr/local/lib/ libboost_unit_test_framework-xgcc42-mt.dylib /usr/local/lib/libboost_unit_test_framework-xgcc42-mt.dylib: Mach-O universal binary with 4 architectures /usr/local/lib/libboost_unit_test_framework-xgcc42-mt.dylib (for architecture i386): Mach-O dynamically linked shared library i386 /usr/local/lib/libboost_unit_test_framework-xgcc42-mt.dylib (for architecture ppc7400): Mach-O dynamically linked shared library ppc /usr/local/lib/libboost_unit_test_framework-xgcc42-mt.dylib (for architecture x86_64): Mach-O 64-bit dynamically linked shared library x86_64 /usr/local/lib/libboost_unit_test_framework-xgcc42-mt.dylib (for architecture ppc64): Mach-O 64-bit dynamically linked shared library ppc64 Source code for test.cpp (note that if I have just one or two test cases it doesn't crash.. having several IS significant): #define BOOST_TEST_MODULE simple-unit-test #include <boost/test/unit_test.hpp> class SimpleFixture { public: SimpleFixture() {} virtual ~SimpleFixture() {} }; BOOST_FIXTURE_TEST_SUITE(SimpleTests_S, SimpleFixture) BOOST_AUTO_TEST_CASE(test1) {} BOOST_AUTO_TEST_CASE(test2) {} BOOST_AUTO_TEST_CASE(test3) {} BOOST_AUTO_TEST_CASE(test4) {} BOOST_AUTO_TEST_CASE(test5) {} BOOST_AUTO_TEST_SUITE_END()

2 1

[test] test_case_filter not building on Solaris
by David Graham 27 Jan '09

27 Jan '09

Hello, We are having trouble with boost::unit_test::test_case_filter while using boost 1.37 on Solaris 10 (x86) with Sun's own C++ compiler (5.9). We're using Sun's own STL implementation, *not* stlport. $ CC -V CC: Sun C++ 5.9 SunOS_i386 Patch 124864-09 2008/12/16 The problem occurs in the constructor of test_case_filter, specifically at this line: ------- m_filters.push_back( std::vector<single_filter>( string_token_iterator( *tit, (dropped_delimeters = ",", kept_delimeters = dt_none) ), string_token_iterator() ) ); ------- Here is the compiler's complaint: ------- "[...]boost-1.37.0/include/boost/test/impl/unit_test_main.ipp", line 116: Error: Could not find a match for std::vector<boost::unit_test::test_case_filter::single_filter>::vector(boost::unit_test::basic_string_token_iterator<char, boost::unit_test::ut_detail::default_char_compare<char>>, boost::unit_test::basic_string_token_iterator<char, boost::unit_test::ut_detail::default_char_compare<char>>) needed in boost::unit_test::test_case_filter::test_case_filter(boost::unit_test::basic_cstring<const char>). "[...]boost-1.37.0/include/boost/test/impl/unit_test_main.ipp", line 116: Error: Formal argument x of type const std::vector<boost::unit_test::test_case_filter::single_filter>& in call to std::vector<std::vector<boost::unit_test::test_case_filter::single_filter> >::push_back(const std::vector<boost::unit_test::test_case_filter::single_filter>&) is being passed int. ------- It would appear that, even though Sun's definition of std::vector does include a (template) constructor that takes two iterators (see /opt/sunstudio12/SUNWspro/prod/include/CC/Cstd/vector), the compiler is unable to match the boost code to this constructor. In fact, it seems unable to infer a match for this constructor even with some simple test code: ------- std::list<int> l; std::vector<int> v(l.begin(), l.end()); ------- ...which results in this error: ------- Error: Could not find a match for std::vector<int>::vector(std::list<int>::iterator, std::list<int>::iterator) ------- Apologies for the length of this post so far. I have two ensuing questions: 1. Is using Sun's C++ compiler and STL implementation with boost a bad idea? (unfortunately, due to other library dependencies, we cannot use gcc and/or stlport) 2. If we provide some workaround code for this particular problem, is it likely that it would be accepted as an addition to boost? Naturally we don't want to maintain our own local boost patches, so if our workaround would be accepted then it would hopefully benefit others as well as saving us from having to maintain local patches. many thanks in advance for any responses, David Graham Senior Software Engineer CR2 Ltd - http://www.cr2.com/ CR2

3 2

[BGL] - Undocumented Predicates for filtered_graph
by Michael Fawcett 27 Jan '09

27 Jan '09

In filtered_graph.hpp there appear to be a number of predefined predicates for general use (they are no in the detail namespace). One of them that I'd like to use is is_in_subset, but it appears that set_contains is not defined anywhere, not even for the most common case (std::set). Shouldn't Boost provide something like template <typename T, typename L, typename A, typename E> bool set_contains(const std::set<T, L, A> &s, const E &e) { return s.find(e) != s.end(); } and vice-versa for is_not_in_subset so that users don't have to open up the boost namespace and define it themselves? Thanks, --Michael Fawcett

2 2

Inner Product and Sparse Matrix in Boost
by Gundala Viswanath 27 Jan '09

27 Jan '09

Hi all, Is there a boost library that allow inner product between sparse matrix and vector? - Gundala Viswanath Jakarta - Indonesia

2 1

boost filesystem path questions
by Sean Farrow 27 Jan '09

27 Jan '09

Hi: I have a few questions re boost::filesystem paths: 1. if I have a path: device1\windows\system32, can I ranslate this using the boost filesystem library to c:\windows\system32"? 2. Does the boost filesystem library accept/use uincode i.e LPWSTR and std::wstring? 3. is the a method for just obtaining the filename from a full path that has been supplied. Any help apreciated. Sean.

2 2