October 2004 - Boost - lists.preview.boost.org

bug in boost/date/time/date_parsing.hpp - wrong wchar_t -> char conversion
by Vyacheslav E. Andrejev 14 Oct '04

14 Oct '04

Hi there, I've found wrong usage of std::stringstream::narrow in date_parsing.hpp template<class date_type, class iterator_type> inline date_type from_stream_type(iterator_type& beg, iterator_type& end, wchar_t) { std::stringstream ss(""); while(beg != end) { ss << ss.narrow(*beg++, 'X'); // 'X' will cause exception to be thrown } return parse_date<date_type>(ss.str()); } In the code above ss.narrow does nothing, because the argument of narrow function already has type char, not wchar_t at all. I think, that the right code should be like this: template<class date_type, class iterator_type> inline date_type from_stream_type(iterator_type& beg, iterator_type& end, wchar_t) { std::stringstream ss(""); while(beg != end) { ss << std::use_facet<std::ctype<wchar_t> >(std::locale()).narrow(*beg++, 'X'); } return parse_date<date_type>(ss.str()); } -- ______________________________ Vyacheslav E. Andrejev System Architect, Excelsior, LLC E-mail: vandrejev(a)excelsior-usa.com ICQ: 7152604

1 0

bug in boost/date_time/time_system_split.hpp - wrong static_cast
by Vyacheslav E. Andrejev 14 Oct '04

14 Oct '04

Hi there I've found a minor bug in time_system_split.hpp. In the following code date_duration_type day_overflow(static_cast<int_type>(doff)); int_type is __int64 if BOOST_DATE_TIME_POSIX_TIME_STD_CONFIG is on. But date_duration_type contructor expects int_adaptor<long> instead. So static_cast is useless here. I think, that right code should look like this: date_duration_type day_overflow(static_cast<date_duration_type::duration_rep_type>(doff)); Proposed patch's attached to this post. -- ______________________________ Vyacheslav E. Andrejev System Architect Excelsior, LLC E-mail: vandrejev(a)excelsior-usa.com ICQ: 7152604

1 0

Redmond Meeting
by Jonathan Turkanis 14 Oct '04

14 Oct '04

Hi All, I'd post this to comp.std.c++ but I'll be traveling soon and hope to get an answer faster here. I'll be in the Seattle area next week and hope to attend the standard committee meeting as a 'technical expert' (no giggling please). I don't see anything on the boost site (http://www.boost.org/more/cpp_committee_meetings.html) or the C++ Standard Committee site about registering advance for meetings. May I assume that it is not necessary? Jonathan

2 1

mpl apply_if.hpp question
by Robert Ramey 14 Oct '04

14 Oct '04

Would someone have a fit if something like this were added to boost/mpl ? Robert Ramey //-------------------------------------------------------------------------- --- // boost/mpl/apply_if.hpp header file // See http://www.boost.org for updates, documentation, and revision history. //-------------------------------------------------------------------------- --- // // Copyright (c) 2004 // Robert Ramey // // Permission to use, copy, modify, distribute and sell this software // and its documentation for any purpose is hereby granted without fee, // provided that the above copyright notice appears in all copies and // that both the copyright notice and this permission notice appear in // supporting documentation. No representations are made about the // suitability of this software for any purpose. It is provided "as is" // without express or implied warranty. // apply_if is deprecated. Use eval_if instead. This is only included // to support older code like spirit 1.6 #ifndef BOOST_MPL_APPLY_IF_HPP_INCLUDED #define BOOST_MPL_APPLY_IF_HPP_INCLUDED #include "boost/mpl/eval_if.hpp" namespace boost { namespace mpl { template< typename C , typename F1 , typename F2 > struct apply_if { typedef BOOST_DEDUCED_TYPENAME eval_if<C,F1,F2>::type type; }; // (almost) copy & paste in order to save one more // recursively nested template instantiation to user template< bool C , typename F1 , typename F2 > struct apply_if_c { typedef BOOST_DEDUCED_TYPENAME eval_if_c<C,F1,F2>::type type; }; } // namespace mpl } // namespace boost #endif // BOOST_MPL_APPLY_IF_HPP_INCLUDED

6 17

Proposed specialization of algorithm/minmax_element
by Robert Minsk 13 Oct '04

13 Oct '04

I would like to propose to specialize algorithm/minmax_element for random access iterators. The inner loop for a random access iterator would have one less compare and does not have the "potential_min" special case. I have included a sample implementation without all the support functions. template <typename RandomAccessIter, typename Compare> std::pair<RandomAccessIter, RandomAccessIter> basic_minmax_element(RandomAccessIter first, RandomAccessIter last, Compare comp, std::random_access_iterator_tag) { if (first == last) return std::make_pair(first, first); RandomAccessIter min(first), max(first), second(first + 1); // Even or odd number of elements if (std::distance(first, last) & 0x01) { // Odd - swap the first and second so first will equal last somewhere down // the line std::swap(first, second); } else { // Even - so treat first pair separately (only one comparison for first two // elements) if (comp(first, second)) max = second; else min = second; first += 2; } // Each element by pairs, with at most 3 comparison per pair while (first != last) { second += 2; if (comp(first, second)) { if (comp(first, min)) min = first; if (comp(max, second)) max = second; } else { if (comp(second, min)) min = second; if (comp(max, first)) max = first; } first += 2; } return std::make_pair(min, max); } template <typename Iter> std::pair<Iter, Iter> minmax_element(Iter first, Iter last) { typedef typename std::iterator_traits<Iter>::iterator_category IterType; return detail::basic_minmax_element(first, last, detail::less_over_iter<Iter>(), IterType()); } template <typename Iter, class _BinaryPred> std::pair<Iter, Iter> minmax_element(Iter first, Iter last, _BinaryPred comp) { typedef typename std::iterator_traits<Iter>::iterator_category IterType; return detail::basic_minmax_element(first, last, detail::binary_pred_over_iter<Iter, _BinaryPred>(comp), IterType()); }

1 0

Mpl Set
by Jaap Suter 13 Oct '04

13 Oct '04

Hi, I noticed that there is a boost/mpl/set directory with some files in them (CVS from circa one week ago). Is this work-in-progress or can I actually start using it in some way? Do you need any help in writing these? I'm in dire need for a compile time set. Regards, Jaap Suter

3 3

New Library Proposals
by Lemings, Eric B. 13 Oct '04

13 Oct '04

Hi, I'd like to solicit interest in the following brief proposals for inclusion in the Boost library. cla (Command-Line Arguments) - A generic C++ framework for parsing command-line arguments. The library design would be based on the popt (http://www.rpm.org/dark_corners/popt/) library. uom (Units Of Measure) - A generic C++ framework for arithmetic calculations and conversions between different units of measure. The library design would be based on The Units Of Measure Library (http://tuoml.sourceforge.net/) Questions and comments welcomed. Thanks, Eric.

7 6

[array] bug, interface change and more
by Joao Abecasis 13 Oct '04

13 Oct '04

Hi all! 1) I want to report that array::at throws range_error when it should throw out_of_range to comply to Sequence requirements. After I decided to report this issue I learned that Nicolai Josuttis maintains a parallel version boost/array.hpp on his website that corrects this and has a few other changes. Namely, an overload for 0-sized arrays was added and c_array has been renamed to data. I believe these changes were made after discussion for proposed inclusion in TR1. 2) Now I'm wondering if c_array should be renamed to data before the next release, to avoid having to change the interface later. 3) Could get<Index> functions be added to array, in order to mimick tuple? There are probably other things that can be done to make boost::array behave like a tuple. I think array could also be considered for inclusion in fusion. 4) On a side note, I wrote an initialized_array class that is very much like array but provides the constructors (and thus cannot provide aggregate initialization syntax): * default element construction * default array construction (boost::array provides this) * initialized_array<T, N>(T const & val) // elements initialized to val * initialized_array<T, N>(T const (&vals)[ N2 ]) // elements initialized with array. If N2 < N remaining elements are default constructed * initialized_array<T, N>(InputIterator start, InputIterator finish) // similar to previous but with iterators Exception behaviour is the same as for arrays (accomplished with a linear hierarchy of base classes) and each element is initialized exactly once (except with default array construction). Is there any interest in this class? br, João Abecasis

2 1

signals: vc7.1 warning in named_slot_map.hpp
by Eric Niebler 13 Oct '04

13 Oct '04

This is actually an error for me since I have "warnings as errors" turned on. For the Dinkumware lib, type_info::operator== and operator!= return int instead of bool. This causes problems with the any_bridge_compare class in named_slot_map.hpp template<typename Compare, typename Key> class any_bridge_compare { ... bool operator()(const any& k1, const any& k2) const { if (k1.type() == typeid(front_type)) return k2.type() != typeid(front_type); -------------^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ This results in warning C4800: 'int' : forcing value to bool 'true' or 'false' (performance warning) I recommend changing the line in question to: return !(k2.type() == typeid(front_type)); -- Eric Niebler Boost Consulting www.boost-consulting.com

2 1

Need more Boost.Build files for CodeWarrior?
by Daryle Walker 13 Oct '04

13 Oct '04

I noticed that there are *.jam files for CodeWarrior 8.3, 9.2, and 9.3. I also noticed that there are separate *.jam files for the Linux and Windows-32 versions of Intel's compiler. But CodeWarrior also comes for multiple platforms (Windows and Mac). Shouldn't the CodeWarrior *.jam files be split up as well? (Unless we're certain that CW reacts exactly the same way for the two platforms.) -- Daryle Walker Mac, Internet, and Video Game Junkie darylew AT hotmail DOT com

2 1

Re: [boost] Re: [Poll] New look-n-feel for Boost documentation
by brangdon＠cix.compulink.co.uk 13 Oct '04

13 Oct '04

In-Reply-To: <415CBB84.1050608(a)boost-consulting.com> joel(a)boost-consulting.com (Joel) wrote (abridged): > <http://tinyurl.com/5emyz> > > Serif or sans serif? > For continuous reading in print, serif fonts are generally more > readable than sans serif. However, on the web, the opposite is true. > Serifs are tiny, subtle strokes which, on screen in a small size, > become a rather crude series of little square bitmaps. Their absence > makes the font more readable. On my 1280x1024 desktop machine, the serif on the middle bar of my default Times New Roman 'F' is drawn with 12 pixels and 4 different colours. Maybe the studies you cite were done before anti-aliasing was common? Regardless, I think it should be up to the user. -- Dave Harris, Nottingham, UK

3 2

[scoped_ptr] Invalid example in the documentation.
by Jurko Gospodneti� 13 Oct '04

13 Oct '04

Hi all. The example at the bottom of http://www.boost.org/libs/smart_ptr/scoped_ptr.htm specifies void as the return type for main() which is not valid C++. Could someone please fix this? Thanks, Jurko Gospodnetiæ

1 0

[filesystem] Failures on FAT32
by Stefan Slapeta 12 Oct '04

12 Oct '04

Hi all, there are two failures reported for the intel-win32-8.1 toolset (both in operations_test): http://tinyurl.com/5dbla After some investigation I've found out now that the tests fail only because they are run on a FAT32 disk! It was just coincidence that the failures popped up in the Intel 8.1 tests (...and it's quite amusing to say that we would need another toolset dimension for the filesystem :) ) Anyway, here some analysis (the second failure seems to be a general design problem!): First failure: operations_test.cpp(287): test time_diff >= 0.0 && time_diff < 60.0 failed --> the actual difference is -2! Second failure: std::exception: boost::filesystem::last_write_time ("wrong parameter") last_write_time contains this code: ::utimbuf buf; buf.actime = std::time_t(); ^^^^^^^^^^^^^ buf.modtime = new_time; This means that the last access (!) time is set to 1970-01-01-... when the last write time is changed by this call! IMO that's absolutely incorrect because it's a side effect which I'm sure the user wouldn't expect to happen! Furthermore - a nice coincidence again - it seems not to be allowed to set this early access time for FAT32. I didn't investigate too much, but what I can say for sure is that it works to set an access time of this year, but not for 1970 (-> results in an error "wrong parameter"). Stefan

3 5

as_sequence<R(T0,T1)> ?
by Alexander Nasonov 12 Oct '04

12 Oct '04

I have one addition to '[type_traits] "member_function_pointer_traits" ?' thread started long time ago (29 Jul 2004) by Tobias Schwinger. Is it makes sense to modify as_sequence to convert function type into a sequence? Eg: as_sequence<R (T0,T1)> ---> vector<R,T0,T1> as_sequence<R(*)(T0,T1)> ---> vector<R,T0,T1> as_sequence<R(&)(T0,T1)> ---> vector<R,T0,T1> as_sequence<R(C::*)(T0,T1)> ---> vector<R,C,T0,T1> -- Alexander Nasonov

5 16

Boost.Build V2 link on top-level page?
by Vladimir Prus 12 Oct '04

12 Oct '04

Are there any objection to my adding Boost.Build V2 link (http://boost.org/boost-build2) to the top-level "index.htm", under "Official branch sites", using the patch below? - Volodya

2 3

Review of Daryle Walker's "More IO" library begins today, August 21, 2004.
by tom brinkman 12 Oct '04

12 Oct '04

The review of Daryle Walker's "More IO" library begins today, August 21, 2004. Download: http://groups.yahoo.com/group/boost/files/more_io_4.zip [or] http://boost-sandbox.sourceforge.net/more_io_4.zip Description: Provides a multitude of I/O-oriented extensions. Some are extra manipulators; variants of "std::endl", a resetting manipulator, and a configurable manipulator. The library includes a set of base class templates for wrapping a new stream-buffer type around a stream class. Changes from the last interim version (called "more_io_3b.zip"): * split the manipulator header into several headers * refined the license statements in some already-affected files Changes from the first formal review (called "more_io.zip"): * refined the interface and implementation of array-streams and stream-buffer wrappers (including a wrapping example) * added the null-stream, value-stream, and pointer-stream headers * added generalizations to the new-line and skip-line manipulators * added the non-oriented 'reset' and 'form' manipulators * split the manipulator header into several headers * changed the license statements Your comments may be brief or lengthy, but basically the Review Manager needs your evaluation of the library. If you identify problems along the way, please note if they are minor, serious, or showstoppers. Here are some questions you might want to answer in your review: 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. For information about submitting a Formal Review, see http://www.boost.org/more/formal_review_process.htm The Review manager is: Tom Brinkman reportbase(a)yahoo.com

9 56

matching algorithm
by Gunnar W. Klau 12 Oct '04

12 Oct '04

Hello, does anybody know whether there's an implementation to compute (maximum weight) matchings in general graphs? Thanks for any help, --gunnar--

2 1

Re: [boost] Re: [review] Formal review of
by brangdon＠cix.compulink.co.uk 12 Oct '04

12 Oct '04

In-Reply-To: <loom.20041011T172801-139(a)post.gmane.org> nesotto(a)cs.auc.dk (Thorsten Ottosen) wrote (abridged): > If there is no null in the tree, then why test for it? The test is to prevent moving a null dog to the tree. > I suspect having null is reasonable, but I explicitly requested to see > the code. There is no real code; the example was invented to illustrate the importance of post-conditions. Which is another way of putting what it says in the .pdf file you referenced: the Null Object pattern is not universally applicable. The point is general and not about the specific code. Are you claiming that Null Object is, in fact, applicable in every situation? If not, just imagine I picked an example where you agree its not applicable. If, on the other hand, you believe no such example is possible, then you surprise me. Regardless I don't think further discussion is appropriate here, because it is not crucial to the Smart Container proposal. As I said in my original message: ... this is no more an argument against ptr_vector<Dog> than it is against std::vector<Dog>. In saying that I was supporting your proposed library. Given that ptr_vector<T> behaves more like vector<T> than vector<T*>, the restriction on NULLs is logical and reasonable. -- Dave Harris, Nottingham, UK

1 0

[lambda] cw83-release vs. cw83-debug failures...
by Rene Rivera 12 Oct '04

12 Oct '04

I became a bit curious as to why two Lambda tests fail in the only release configuration for CW8.3, but don't fail in the debug tests. As seen in: http://www.meta-comm.com/engineering/boost-regression/developer/lambda.html And the really strange part is the lack of information from Boost.Test about it. It only outputs: Run output [2004-10-11 02:46:48 UTC]: **** no errors detected EXIT STATUS: 1 Which seems like a rather contradictory statement from the test system I don't really know what is going on here. But I've narrowed it down to some strange interaction when the "--" and "++" pre-ops are used. If I make the following changes the tests pass... Index: bind_tests_advanced.cpp =================================================================== RCS file: /cvsroot/boost/boost/libs/lambda/test/bind_tests_advanced.cpp,v retrieving revision 1.8 diff -u -r1.8 bind_tests_advanced.cpp --- bind_tests_advanced.cpp 27 Jul 2004 03:43:33 -0000 1.8 +++ bind_tests_advanced.cpp 12 Oct 2004 03:02:38 -0000 @@ -110,7 +110,7 @@ int i = 1; BOOST_TEST(unlambda(_1 + _2)(i, i) == 2); - BOOST_TEST(unlambda(++var(i))() == 2); + { bool t = unlambda(++var(i))() == 2; BOOST_TEST(t); } BOOST_TEST(call_with_100(_1 + 1) == 101); Index: operator_tests_simple.cpp =================================================================== RCS file: /cvsroot/boost/boost/libs/lambda/test/operator_tests_simple.cpp,v retrieving revision 1.8 diff -u -r1.8 operator_tests_simple.cpp --- operator_tests_simple.cpp 27 Jul 2004 03:43:33 -0000 1.8 +++ operator_tests_simple.cpp 12 Oct 2004 03:02:38 -0000 @@ -175,11 +175,11 @@ BOOST_TEST(i == -1); i = 0; - BOOST_TEST((++_1)(i) == 1); + { bool t = (++_1)(i) == 1; BOOST_TEST(t); } BOOST_TEST(i == 1); i = 0; - BOOST_TEST((--_1)(i) == -1); + { bool t = (--_1)(i) == -1; BOOST_TEST(t); } BOOST_TEST(i == -1); i = 0; =================================================================== Any ideas, clues, etc. ?? -- -- Grafik - Don't Assume Anything -- Redshift Software, Inc. - http://redshift-software.com -- rrivera/acm.org - grafik/redshift-software.com - 102708583/icq

1 0

[ANN] Boost testing mailing list
by Misha Bergal 12 Oct '04

12 Oct '04

The boost-testing mailing list is now operational: The setup, procedures and tools necessary for running Boost regression tests are discussed on this list. The list main participants are regression runners - people who run Boost tests on a variety of compilers and platforms, and the maintainers of tools for collecting and processing test results. Important: questions relevant to a wider audience, including questions about Boost.Test framework or test results for a particular library, should be posted to main development list. You can subscribe to the list at http://lists.boost.org/mailman/listinfo.cgi/boost-testing. -- Misha Bergal MetaCommunications Engineering

3 4

filesystem::is_directory return true when exception is disabled and this directory does't exist
by Liao Weijie 12 Oct '04

12 Oct '04

// the snip of code in filesystem::is_directory begin--- DWORD attributes = ::GetFileAttributesA( ph.native_directory_string().c_str() ); if ( attributes == 0xFFFFFFFF ) boost::throw_exception( filesystem_error( "boost::filesystem::is_directory", ph, fs::detail::system_error_code() ) ); return (attributes & FILE_ATTRIBUTE_DIRECTORY) != 0; //-------------------------------------------------------end-- When GetFileAttributesA failed but the exception is disabled, the (attributes & FILE_ATTRIBUTE_DIRECTORY) will always be true. Is it a bug?

4 4

Suggestion for multi_array query information
by Anders Edin 11 Oct '04

11 Oct '04

Hello, I have a short suggestion about the interface for multi_array. It is possible to query a multi_array<T,N> for information about the storage: // queries ... const size_type* shape() const; const index* strides() const; const index* index_bases() const; ... It appears multi_array uses a boost::array internally to store this information, and uses the data() method to return a pointer in the above methods. Wouldn't it be better to return a reference to the array directly: // queries ... const array<size_type, N>& shape() const; const array<index, N>& strides() const; const array<index, N>& index_bases() const; ... With this interface the information about the number of elements of shapes and strides is not lost during the query. -- Anders Edin, Sidec Technologies AB Fogdevreten 2A, 171 77 Stockholm, Sweden Phone: +46850884525, Fax: +46850884522

2 1

Boost.String Algorithms patch
by Daniel Schlyder 11 Oct '04

11 Oct '04

The attached patch fixes invalid #include in boost/algorithm/string_regex.hpp and unused parameter warnings in boost/algorithm/string/formatter.hpp boost/algorithm/string/detail/find_format.hpp -- Daniel Schlyder http://bitblaze.com/

2 1

MPL - mpl::sequence to std::sequence
by Jaap Suter 11 Oct '04

11 Oct '04

Hi, what is the easiest way to convert an mpl::sequence to an std::sequence? I'm using such conversions a lot when unit-testing meta-code. I just wrote two versions of algorithms, a compile-time and the runtime one, and then compare their outputs. I recall hearing something about mpl::for_each having defined an operator () for runtime purposes, but the details don't seem to be documented anywhere. I came up with the attached code, of which the "to_compile_time" might actually be of general interest. It's a function that converts a runtime integer to a compile time integer. Code-generation-wise it's a horror and it could be more generic, but it gets the job done for me. I'd like to hear if there are more elegant solutions for this though, as it seems like too much code. Thanks, Jaap Suter

5 11

Preprocessor headers docs
by Aleksey Gurtovoy 11 Oct '04

11 Oct '04

Paul, Are these (from http://tinyurl.com/4hp7d) automatically generated: libs/preprocessor/doc/headers/arithmetic.hpp.html: filename contains more than one dot character ('.') libs/preprocessor/doc/headers/arithmetic/add.hpp.html: filename contains more than one dot character ('.') libs/preprocessor/doc/headers/arithmetic/dec.hpp.html: filename contains more than one dot character ('.') libs/preprocessor/doc/headers/arithmetic/div.hpp.html: filename contains more than one dot character ('.') libs/preprocessor/doc/headers/arithmetic/inc.hpp.html: filename contains more than one dot character ('.') libs/preprocessor/doc/headers/arithmetic/mod.hpp.html: filename contains more than one dot character ('.') libs/preprocessor/doc/headers/arithmetic/mul.hpp.html: filename contains more than one dot character ('.') libs/preprocessor/doc/headers/arithmetic/sub.hpp.html: filename contains more than one dot character ('.') ... ? If so, can we adopt a convention that doesn't involve multiple dots [1] and re-generate them? [1] E.g. "libs/preprocessor/doc/headers/arithmetic/add_hpp.html", or even, since these are in their own subdirectory anyway ("headers"), plain "libs/preprocessor/doc/headers/arithmetic/add.html". -- Aleksey Gurtovoy MetaCommunications Engineering

1 1