July 2005 - Boost - lists.preview.boost.org

[serialization][date_time] unresolved symbol 'special_value_from_string'
by Jonathan Turkanis 18 Jul '05

18 Jul '05

Hi, I'm trying to deserialize objects from an xml archive containing some gregorian dates. Using VC7.1 I get the linker error: serialization.obj : error LNK2019: unresolved external symbol "enum boost::date_time::special_values __cdecl boost::gregorian::special_value_from_string(class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> > const &)" (?special_value_from_string@gregorian@boost@@YA?AW4special_values@date_time@2@AB V?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@@Z) referenced in function __catch$??$load@Vxml_wiarchive@archive@boost@@@serialization@boost@@YAXAAVxml_wi archive@archive@1@AAVdate@gregorian@1@I@Z$0 Any ideas what's causing this? Jonathan

1 0

[filesystem] operations failues on gcc-3_4_4-cygwin
by Beman Dawes 18 Jul '05

18 Jul '05

I inadvertently upgraded my cygwin installation this morning to what __VERSION__ reports as "3.4.4 (cygming special) (gdc 0.12, using dmd 0.125)". The good news is that two freshly added dll tests which had been failing under gcc/cygwin 3.3.3 pass under 3.4.4. The bad news is that even though operations_test_dll is passing, statically linked operations_test fails. The reason for the static link failures is that the Windows function ::GetLastError() is returning 0 after several critical functions return error indications. I'm guessing from the "cygming" name that mingw is being used, and is the cause of the problem. Why the same problem doesn't appear when dynamically linking is unclear to me. My initial reaction is just to note the problem in the docs and move on. It seems to close to release to try workarounds, which would have to be very carefully done. Any other ideas? --Beman

1 0

[Boost-bugs] [ boost-Bugs-1239906 ] no Boolean conversion for boost::rational
by SourceForge.net 17 Jul '05

17 Jul '05

Bugs item #1239906, was opened at 2005-07-17 16:15 Message generated for change (Tracker Item Submitted) made by Item Submitter You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=107586&aid=1239906&group_… Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: None Group: None Status: Open Resolution: None Priority: 5 Submitted By: Daryle L. Walker (dlwalker) Assigned to: Nobody/Anonymous (nobody) Summary: no Boolean conversion for boost::rational Initial Comment: The boost::rational class template, in header file <boost/ rational.hpp>, does not support conversion to a Boolean value. It does have an "operator !" defined, though. It is a surprise for a type to have one of these operations but not the other. Currently code like "if ( !r )" works but not "if ( r )", which should be expected. A solution is to add a conversion operator to a type that can be used in Boolean contexts, but as little else as possible. A suggestion is to use a pointer-to-member (to data or function) type that is unique to each version of boost::rational, where the exact values are private. ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=107586&aid=1239906&group_… ------------------------------------------------------- SF.Net email is sponsored by: Discover Easy Linux Migration Strategies from IBM. Find simple to follow Roadmaps, straightforward articles, informative Webcasts and more! Get everything you need to get up to speed, fast. http://ads.osdn.com/?ad_id=7477&alloc_id=16492&op=click _______________________________________________ Boost-bugs mailing list Boost-bugs(a)lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/boost-bugs

1 0

Re: [boost] VC 7.1/8.0 regression testing, call for volunteers
by David Abrahams 17 Jul '05

17 Jul '05

"Carl Daniel" <cpdaniel(a)mvps.org> writes: > David Abrahams wrote: >> "Carl Daniel" <cpdaniel(a)mvps.org> writes: >>> Does anyone have a set of instructions for building and running the >>> regression tests WITHOUT using Python? Right now, that's the >>> sticking point - Python's not installed, and I'm unable to install >>> it. If I could put together a ZIP file with all the command line >>> tools pre-built and a simple BATch file to run them, that'd be a >>> workable solution. >> >> Somehow if they're letting us upload, build, and run our code it seems >> like it should be possible to upload, build, and run Python, no? > > True that - I could upload the sources and just build it in place. It's trivial; I would just go ahead and do it. The build system is set up to work with a development installation of Python. > Hopefully someone on the VC++ team can just install ActivePython on > the machine for me. If I build Python locally, is there a > straightforward way to identify all the files I'd need to push to > the machine to have it work there? Not that I know of. Why not just do the build remotely? > I'm assuming there'd be more to it than just python.exe. Yeah, all of the standard library, for example. And python.dll. Not sure what else right now. -- Dave Abrahams Boost Consulting www.boost-consulting.com

1 0

Boost regression notification (2005-07-17)
by Douglas Gregor 17 Jul '05

17 Jul '05

Boost regression test failures ------------------------------ Report time: 2005-07-17T14:58:38Z This report lists all regression test failures on release platforms. WARNING: The following libraries have failing regression tests but do not have a maintainer on file. Once a maintainer is found, add an entry to libs/maintainers.txt to eliminate this message. io utility Detailed report: http://engineering.meta-comm.com/boost-regression/CVS-HEAD/developer/issues… 245 failures in 13 libraries: io (1) multi_index (1) parameter (23) program_options (5) python (53) random (2) serialization (140) spirit (1) test (6) thread (1) type_traits (1) utility (1) wave (10) |io| ios_state_unit_test: intel-win32-8_1 |multi_index| test_special_set_ops: vc-8_0 |parameter| basics: borland-5_6_4 cw-8_3 gcc-3_4_3-sunos vc-6_5 vc-6_5-stlport vc-7_0 vc-8_0 macros: borland-5_6_4 cw-8_3 gcc-3_4_3-sunos vc-6_5 vc-6_5-stlport vc-7_0 vc-8_0 sfinae: borland-5_6_4 cw-8_3 cw-9_4 gcc-3_4_3-sunos intel-win32-8_1 vc-6_5 vc-6_5-stlport vc-7_0 vc-8_0 |program_options| cmdline_test_dll: mingw-3_4_2 positional_options_test_dll: mingw-3_4_2 unicode_test_dll: cw-9_5-darwin gcc-3_3-darwin gcc-3_3-darwin |python| andreas_beyer: intel-win32-8_1 args: intel-win32-8_1 auto_ptr: intel-win32-8_1 back_reference: intel-win32-8_1 ben_scott1: intel-win32-8_1 bienstman1: intel-win32-8_1 bienstman2: intel-win32-8_1 bienstman3: intel-win32-8_1 builtin_converters: intel-win32-8_1 callbacks: intel-win32-8_1 const_argument: intel-win32-8_1 crossmod_exception: intel-win32-8_1 data_members: intel-win32-8_1 defaults: intel-win32-8_1 dict: intel-win32-8_1 docstring: intel-win32-8_1 enum: intel-win32-8_1 exception_translator: intel-win32-8_1 extract: intel-win32-8_1 implicit: intel-win32-8_1 injected: intel-win32-8_1 iterator: intel-win32-8_1 keywords: intel-win32-8_1 list: intel-win32-8_1 long: intel-win32-8_1 map_indexing_suite: intel-win32-8_1 minimal: intel-win32-8_1 multi_arg_constructor: intel-win32-8_1 nested: intel-win32-8_1 numpy: intel-win32-8_1 object: intel-win32-8_1 opaque: intel-win32-8_1 operators: intel-win32-8_1 pearu1: intel-win32-8_1 pickle1: intel-win32-8_1 pickle2: intel-win32-8_1 pickle3: intel-win32-8_1 pickle4: intel-win32-8_1 pointer_vector: intel-win32-8_1 polymorphism: intel-win32-8_1 polymorphism2: intel-win32-8_1 properties: intel-win32-8_1 raw_ctor: intel-win32-8_1 return_arg: intel-win32-8_1 shared_ptr: intel-win32-8_1 slice: intel-win32-8_1 staticmethod: intel-win32-8_1 str: intel-win32-8_1 test_pointer_adoption: intel-win32-8_1 try: intel-win32-8_1 tuple: intel-win32-8_1 vector_indexing_suite: intel-win32-8_1 virtual_functions: intel-win32-8_1 |random| random_test: intel-win32-8_1 intel-win32-8_1 |serialization| test_class_info_load_binary_archive: gcc-3.3.6-linux gcc-3.4.4-linux vc-7_1 vc-7_1 vc-8_0 test_class_info_load_binary_archive_dll: gcc-3.3.6-linux gcc-3.4.4-linux vc-7_1 vc-7_1 vc-8_0 test_class_info_load_text_archive: cw-9_5-darwin gcc-3.3.6-linux gcc-3.4.4-linux vc-7_1 vc-7_1 vc-8_0 test_class_info_load_text_archive_dll: cw-9_5-darwin gcc-3.3.6-linux gcc-3.4.4-linux vc-7_1 vc-7_1 vc-8_0 test_class_info_load_text_warchive: cw-9_5-darwin gcc-3.3.6-linux gcc-3.4.4-linux vc-7_1 vc-7_1 test_class_info_load_text_warchive_dll: cw-9_5-darwin gcc-3.3.6-linux gcc-3.4.4-linux vc-6_5 vc-7_1 vc-7_1 vc-8_0 test_class_info_load_xml_archive: gcc-3.3.6-linux gcc-3.4.4-linux vc-7_1 vc-7_1 vc-8_0 test_class_info_load_xml_archive_dll: gcc-3.3.6-linux gcc-3.4.4-linux vc-7_1 vc-7_1 vc-8_0 test_class_info_load_xml_warchive: cw-9_5-darwin gcc-3.3.6-linux gcc-3.4.4-linux vc-7_1 vc-7_1 test_class_info_load_xml_warchive_dll: cw-9_5-darwin gcc-3.3.6-linux gcc-3.4.4-linux vc-7_1 vc-7_1 vc-8_0 test_const_load_fail1: borland-5_6_4 test_const_load_fail1_nvp: borland-5_6_4 test_const_load_fail2: borland-5_6_4 test_const_load_fail2_nvp: borland-5_6_4 test_const_save_fail1: vc-6_5 vc-6_5-stlport vc-7_0 test_const_save_fail2: vc-6_5 vc-6_5-stlport vc-7_0 test_const_save_fail3: vc-6_5 vc-6_5-stlport vc-7_0 test_demo: intel-win32-8_1 test_demo_dll: cw-9_4 cw-9_4 test_demo_exception: intel-win32-8_1 test_demo_exception_dll: cw-9_4 cw-9_4 test_demo_fast_archive_dll: vc-6_5 test_demo_pimpl_dll: vc-6_5 vc-7_0 test_demo_portable_archive: intel-win32-8_1 intel-win32-8_1 vc-7_0 vc-7_1 vc-7_1 vc-7_1 vc-7_1 vc-8_0 vc-8_0 vc-8_0 test_demo_portable_archive_dll: borland-5_6_4 intel-win32-8_1 intel-win32-8_1 test_demo_shared_ptr: cw-9_4 cw-9_4 cw-9_5-darwin intel-win32-8_1 test_demo_shared_ptr_dll: cw-9_4 cw-9_4 cw-9_5-darwin vc-6_5 vc-7_0 test_demo_xml_load: intel-win32-8_1 test_demo_xml_save: intel-win32-8_1 test_diamond_binary_archive: cw-8_3 cw-9_5-darwin test_diamond_binary_archive_dll: cw-9_5-darwin test_diamond_text_archive: cw-8_3 cw-9_5-darwin test_diamond_text_archive_dll: cw-9_5-darwin test_diamond_text_warchive: cw-8_3 cw-9_5-darwin test_diamond_text_warchive_dll: cw-9_5-darwin test_diamond_xml_archive: cw-8_3 cw-9_5-darwin test_diamond_xml_archive_dll: cw-9_5-darwin test_diamond_xml_warchive: cw-8_3 cw-9_5-darwin test_diamond_xml_warchive_dll: cw-9_5-darwin test_list_xml_archive: vc-8_0 test_map_binary_archive_dll: vc-6_5 vc-7_0 test_map_text_archive_dll: vc-6_5 vc-7_0 test_map_text_warchive_dll: vc-6_5 vc-7_0 test_mi_xml_archive_dll: vc-8_0 test_mult_archive_types: cw-9_4 cw-9_4 cw-9_5-darwin test_mult_archive_types_dll: cw-9_4 cw-9_4 cw-9_5-darwin vc-6_5 test_non_default_ctor_text_archive_dll: cw-9_5-darwin test_non_default_ctor_text_warchive_dll: cw-9_5-darwin test_non_default_ctor_xml_archive_dll: cw-9_5-darwin test_non_default_ctor_xml_warchive_dll: cw-9_5-darwin test_non_intrusive_text_archive_dll: cw-9_5-darwin test_non_intrusive_text_warchive_dll: cw-9_5-darwin test_non_intrusive_xml_archive_dll: cw-9_5-darwin test_non_intrusive_xml_warchive_dll: cw-9_5-darwin test_registered_xml_archive: vc-8_0 |spirit| owi_mt_tests: gcc-3_3-darwin |test| errors_handling_test: cw-9_5-darwin vc-8_0 test_fp_comparisons: borland-5_6_4 test_tools_test: cw-9_4 cw-9_4 cw-9_5-darwin |thread| test_mutex: vc-7_1 |type_traits| add_const_test: vc-8_0 |utility| compressed_pair_test: intel-win32-8_1 |wave| test_re2c_lexer: vc-8_0 vc-8_0 test_slex_lexer: vc-8_0 vc-8_0 testwave: cw-8_3 cw-9_5-darwin gcc-3.2.3-linux gcc-3.3.6-linux vc-8_0 vc-8_0

1 0

[date_time] day_of_week function
by Friedrich Wilckens 17 Jul '05

17 Jul '05

date_time::date::day_of_week() calls two functions from the calendar template parameter: the first to convert the internal integer representation into a ymd_type, and the second to calculate the day_of_week for this ymd_type. Both do some non-trivial arithmetic. I wonder why the following simpler and more efficient approach is not used. In date_time::date: day_of_week_type day_of_week() const { return calendar::day_of_week(days_); } In date_time::gregorian_calendar_base add a method: static unsigned short day_of_week(const date_int_type& d) { return (d + 1) % 7; } As far as I know, the sequence of weeks in the gregorian calendar is completely regular and this simple calculation is correct. This approach would also not limit flexibility of the date template. If for another calendar there is some irregularity, we could for this calendar implement static unsigned short day_of_week(const date_int_type& d) { return day_of_week(from_day_number(d)); } Is there any problem in this reasoning? Sincerely, Friedrich

2 1

The OSL4 sunpro-5_8u1-sunos smart_ptr failures
by Peter Dimov 17 Jul '05

17 Jul '05

... are caused by a missing -features=tmplife option. Without it, temporaries are destroyed at the end of the block. -- Peter Dimov http://www.pdimov.com

2 1

static templates in BOOST
by Steve Clamage 17 Jul '05

17 Jul '05

We recently corrected the way our compiler (Sun C++) handles static templates, and ran into a problem with some BOOST code. More than 100 tests in the suite depend on code like this: template <typename T> struct S { void foo(int); }; template<class T , class U> static // static template ... void smart_cast_reference(U& u) {} template <typename T> void S<T>::foo(int c) { smart_cast_reference< T& >(c); // used in global function } int main() { S<int> s; s.foo(1); } You can't use a static template in a global template function because it violates the One-Definition Rule, among possibly other rules. I have found that other compilers accept this code, but reject it when you put them in standard-conformning mode. (To see the rejection with some compilers, you need to build the executable.) How critical to BOOST is this style of coding? Could the static templates be made global? --- Steve Clamage, stephen.clamage(a)sun.com

9 24

weak_ptr triple states
by Jaap Suter 17 Jul '05

17 Jul '05

Hi, a weak_ptr can be said to have three states. 1. Never been assigned to (default constructed) 2. Assigned to, lock will return a shared_ptr to a valid object. 3. Assigned to, but the object pointed to is dead and lock will return null. As far as I know it is currently impossible to distinguish between case 1 and 3. Am I overlooking something? If not, can somebody explain the reason why it is not possible to distinguish the two cases? Admittedly, I cannot come up with any elegant design in which such a distinction is useful, so the question is merely theoretical. Thanks, Jaap Suter

4 5

Re: [boost] VC 7.1/8.0 regression testing, call for volunteers
by David Abrahams 16 Jul '05

16 Jul '05

"Carl Daniel" <cpdaniel(a)mvps.org> writes: > Aleksey Gurtovoy wrote: >> Carl Daniel writes: >>> I now have access to weekly builds of VC8 (currently testing the >>> 7/11/2005 build!). At the moment, it's not feasible to do boost >>> regression testing on those builds (it's a machine at MSFT), but I >>> could look into getting it set up. >> >> I think you are in a unique position w.r.t. weekly VC 8.0 builds, and >> IMO it would be a great service to both us and the VC8 team to test >> against those in addition to CTP/Beta 2. > > It looks like this might be possible - stay tuned. It'll be mid next week > at best before I can do the first run (assuming things get set up). I won't > be able to do them very often - perhaps only once/week because it won't be > possible to fully automate the process (the machine cannot make outgoing > connections to the internet, so I'll have to push tarballs to it, run the > tests, and then pull the results back. > > Does anyone have a set of instructions for building and running the > regression tests WITHOUT using Python? Right now, that's the sticking > point - Python's not installed, and I'm unable to install it. If I could > put together a ZIP file with all the command line tools pre-built and a > simple BATch file to run them, that'd be a workable solution. Somehow if they're letting us upload, build, and run our code it seems like it should be possible to upload, build, and run Python, no? -- Dave Abrahams Boost Consulting www.boost-consulting.com

1 0