- Boost - lists.boost.org

Metacomm STLport regressions generally failing
by Michael Stevens 18 Aug '04

18 Aug '04

Hi Aleksey, It seems that the STLport tests 'msvc- stlport' and 'intel- win32- 7.1- vc6- stlport- 4.5.3' for many libraries are failing with the same error. They report the library 'stlport_vc6_stldebug.lib' as missing. No idea why this should be! Thanks for the great work and very useful test runs. I'm grinding uBLAS back to VC6 compaitbility! Michael

1 0

Metacomm
by Michael Stevens 18 Aug '04

18 Aug '04

intel- win32- 7.1- vc6- stlport- 4.5.3 msvc- stlport -- ___________________________________ Michael Stevens Systems Engineering 34128 Kassel, Germany Phone/Fax: +49 561 5218038 Navigation Systems, Estimation and Bayesian Filtering http://bayesclasses.sf.net ___________________________________

1 0

RE: [boost] Re: ETI bug and iterators
by Yitzhak Sapir 18 Aug '04

18 Aug '04

> -----Original Message----- > From: Yitzhak Sapir > I can't send my own code, but the following code causes the > previously described error with an unpatched > minimum_category.hpp and an ICE when compiled with patched > and unpatched minimum_category.hpp: I meant an ICE with patched only. Sorry about that.

1 0

RE: [boost] Re: [boost.test] Patch for true64cxx65 support.
by Rozental, Gennadiy 18 Aug '04

18 Aug '04

> Because I get duplicated symbols when linking otherwise. > Therefore you > need to use either > > c_literal_string LOG_LEVEL = "BOOST_TEST_LOG_LEVEL"; > > or > > literal_string const LOG_LEVEL = "BOOST_TEST_LOG_LEVEL"; > > in the header file. > > Markus I would unconditionally add static to all parameter definitions instead. Gennadiy.

2 1

[inspect] Update on inspection tests...
by Rene Rivera 18 Aug '04

18 Aug '04

Here's a summary of the latest state of the Boost tree: 8101 files scanned 680 directories scanned 221 problems reported problem counts: 1 files with invalid line endings 0 bookmarks with invalid characters 1 invalid urls 99 broken links 46 unlinked files 60 file names too long 14 files with tabs As you can see some TABS have crept in. And a fair number of files break the 31 character limit. And for license and copyrights: 8101 files scanned 680 directories scanned 4657 problems reported problem counts: 3230 files missing Boost license info 1427 files missing copyrights Slowly but surely it's improving. More details at: http://www.boost.org/regression-logs/ -- -- Grafik - Don't Assume Anything -- Redshift Software, Inc. - http://redshift-software.com -- rrivera/acm.org - grafik/redshift-software.com - 102708583/icq

1 0

RE: [boost] [boost.test] Patch for true64cxx65 support.
by Rozental, Gennadiy 17 Aug '04

17 Aug '04

> Hello, > > attached is a patch which makes boost.test work for the > beforementioned > toolset. It contains a workaround for the internal compiler > error currently > preventing any tests from running. Could anyone please apply > it, if it's ok? > > Thanks, Markus Why do you need anything more then: +#if BOOST_WORKAROUND(__DECCXX_VER, BOOST_TESTED_AT(60590041)) +typedef const_string literal_string; +#else typedef const_string const literal_string; +#endif Gennadiy

2 1

Fw: Rational Discovery LLC Copyrights?
by John Maddock 17 Aug '04

17 Aug '04

Another permission for our records: ----- Original Message ----- From: "Greg Landrum" <Landrum(a)RationalDiscovery.com> To: "David Abrahams" <dave(a)boost-consulting.com>; "John Maddock" <john(a)johnmaddock.co.uk> Sent: Monday, August 16, 2004 3:37 PM Subject: Re: Rational Discovery LLC Copyrights? > At 07:14 AM 8/16/2004, David Abrahams wrote: > >"John Maddock" <john(a)johnmaddock.co.uk> writes: > > > > > Dave, > > > > > > I see that you committed two files copyright "Rational Discovery LLC" > > > (libs/python/test/crossmod_exception_a.cpp > > > libs/python/test/crossmod_exception_b.cpp). Do you know who we need to > > > contact to get permission to change these to the BSL? > > > >These are Greg Landrum's. I am not certain how to reach him, but I > >Cc'd all the likely emails I could find. > > Feel free to change the copyright/license statements as is necessary. > > Best Regards, > Greg Landrum > President > Rational Discovery LLC > > >

3 4

RE: [boost] Re: ETI bug and iterators
by Yitzhak Sapir 17 Aug '04

17 Aug '04

> -----Original Message----- > From: David Abrahams > > "Yitzhak Sapir" <yitzhak.sapir(a)actimize.com> writes: > > > Hello. > > > > I am in the process of upgrading code to use boost 1.31.0 > for MSVC6.5. > > During this process I happened upon minimum_category.hpp which > > contains "ETI workarounds". Now, it seems that the ETI workarounds > > cause the compilation to fail, with the error message: constant > > expression expected. Particularly, we are creating zip iterators to > > two STLPort vectors. In debug iterator mode, I get the following > > error. (minimum_category.hpp was modified a bit > > Post the error you get with an unmodified minimum_category.hpp, > please. It's hard to analyze when the line numbers are wrong. > ../../3rdParty\boost/iterator/detail/minimum_category.hpp(86) : error C2039: 'apply' : is not a member of '?$minimum_category_impl@$$00' ../../3rdParty\boost/iterator/detail/minimum_category.hpp(97) : see reference to class template instantiation 'boost::detail::minimum_category<struct boost::random_access_traversal_tag,struct boost::random_access_traversal_tag>' being compil ed ../../3rdParty\boost/mpl/aux_/preprocessed/msvc60/apply.hpp(88) : see reference to class template instantiation 'boost::detail::minimum_category<struct boost::mpl::arg<1>,struct boost::mpl::arg<2> >::apply<struct boost::random_access_travers al_tag,struct boost::random_access_traversal_tag>' being compiled ../../3rdParty\boost/mpl/aux_/preprocessed/msvc60/apply.hpp(101) : see reference to class template instantiation 'boost::mpl::aux::msvc_apply2<struct boost::detail::minimum_category<struct boost::mpl::arg<1>,struct boost::mpl::arg<2> > >::re sult_<struct boost::random_access_traversal_tag,struct boost::random_access_traversal_tag>' being compiled ../../3rdParty\boost/iterator/zip_iterator.hpp(173) : see reference to class template instantiation 'boost::mpl::apply2<struct boost::detail::minimum_category<struct boost::mpl::arg<1>,struct boost::mpl::arg<2> >,struct boost::random_access_ traversal_tag,struct boost::random_access_traversal_tag>' being compiled ../../3rdParty\boost/mpl/aux_/preprocessed/msvc60/apply.hpp(9) : see reference to class template instantiation 'boost::detail::tuple_impl_specific::tuple_meta_accumulate_impl<struct boost::tuples::cons<struct boost::random_access_traversal_t ag,struct boost::tuples::null_type>,struct boost::detail::minimum_category<struct boost::mpl::arg<1>,struct boost::mpl::arg<2> >,struct boost::random_access_traversal_tag>' being compiled ../../3rdParty\boost/mpl/apply_if.hpp(39) : see reference to class template instantiation 'boost::mpl::apply0<struct boost::detail::tuple_impl_specific::tuple_meta_accumulate_impl<struct boost::tuples::cons<struct boost::random_access_traver sal_tag,struct boost::tuples::null_type>,struct boost::detail::minimum_category<struct boost::mpl::arg<1>,struct boost::mpl::arg<2> >,struct boost::random_access_traversal_tag> >' being compiled ../../3rdParty\boost/iterator/zip_iterator.hpp(198) : see reference to class template instantiation 'boost::mpl::apply_if<struct boost::mpl::or_<struct boost::is_same<struct boost::tuples::cons<struct boost::random_access_traversal_tag,struc t boost::tuples::null_type>,struct boost::tuples::null_type>,struct boost::is_same<struct boost::tuples::cons<struct boost::random_access_traversal_tag,struct boost::tuples::null_type>,int>,struct boost::mpl::bool_<0>,struct boost::mpl::bool_<0>,str uct boost::mpl::bool_<0> >,struct boost::mpl::identity<struct boost::random_access_traversal_tag>,struct boost::detail::tuple_impl_specific::tuple_meta_accumulate_impl<struct boost::tuples::cons<struct boost::random_access_traversal_tag,struct boost ::tuples::null_type>,struct boost::detail::minimum_category<struct boost::mpl::arg<1>,struct boost::mpl::arg<2> >,struct boost::random_access_traversal_tag> >' being compiled > > but the problem is > > in the instantiation of boost::is_same<T1 or T2,int>). > ^^^^^^^^ > > Huh? That's not valid regardless of any ETI workarounds. > That was my attempt to shorthand "either boost::is_same<T1,int> or boost::is_same<T2,int>"

2 1

[boost.test] Patch for true64cxx65 support.
by Markus Schöpflin 17 Aug '04

17 Aug '04

Hello, attached is a patch which makes boost.test work for the beforementioned toolset. It contains a workaround for the internal compiler error currently preventing any tests from running. Could anyone please apply it, if it's ok? Thanks, Markus diff -ru boost-04-08-17-0400/boost/test/detail/basic_cstring/basic_cstring_fwd.hpp boost-cvs/boost/test/detail/basic_cstring/basic_cstring_fwd.hpp --- boost-04-08-17-0400/boost/test/detail/basic_cstring/basic_cstring_fwd.hpp 2004-07-19 14:28:17.000000000 +0200 +++ boost-cvs/boost/test/detail/basic_cstring/basic_cstring_fwd.hpp 2004-08-17 17:39:02.000000000 +0200 @@ -16,13 +16,21 @@ #ifndef BASIC_CSTRING_FWD_HPP_071894GER #define BASIC_CSTRING_FWD_HPP_071894GER +#include <boost/detail/workaround.hpp> + namespace boost { namespace unit_test { template<class CharT> class basic_cstring; typedef basic_cstring<char const> const_string; + +#if BOOST_WORKAROUND(__DECCXX_VER, BOOST_TESTED_AT(60590041)) +typedef const_string literal_string; +#else typedef const_string const literal_string; +#endif + typedef char const* const c_literal_string; } // namespace unit_test diff -ru boost-04-08-17-0400/boost/test/detail/unit_test_parameters.hpp boost-cvs/boost/test/detail/unit_test_parameters.hpp --- boost-04-08-17-0400/boost/test/detail/unit_test_parameters.hpp 2004-07-19 14:24:32.000000000 +0200 +++ boost-cvs/boost/test/detail/unit_test_parameters.hpp 2004-08-17 17:33:31.000000000 +0200 @@ -25,19 +25,41 @@ namespace unit_test { // framework parameters and there corresponding command-line arguments -literal_string LOG_LEVEL = "BOOST_TEST_LOG_LEVEL"; // --log_level -literal_string NO_RESULT_CODE = "BOOST_TEST_RESULT_CODE"; // --result_code -literal_string REPORT_LEVEL = "BOOST_TEST_REPORT_LEVEL"; // --report_level -literal_string TESTS_TO_RUN = "BOOST_TESTS_TO_RUN"; // --run_test -literal_string SAVE_TEST_PATTERN = "BOOST_TEST_SAVE_PATTERN"; // --save_pattern -literal_string BUILD_INFO = "BOOST_TEST_BUILD_INFO"; // --build_info -literal_string CATCH_SYS_ERRORS = "BOOST_TEST_CATCH_SYSTEM_ERRORS"; // --catch_system_errors -literal_string REPORT_FORMAT = "BOOST_TEST_REPORT_FORMAT"; // --report_format -literal_string LOG_FORMAT = "BOOST_TEST_LOG_FORMAT"; // --log_format -literal_string OUTPUT_FORMAT = "BOOST_TEST_OUTPUT_FORMAT"; // --output_format - -enum report_level { CONFIRMATION_REPORT, SHORT_REPORT, DETAILED_REPORT, NO_REPORT, UNDEF_REPORT }; -literal_string report_level_names[] = { "confirm" , "short" , "detailed" , "no" }; +#if BOOST_WORKAROUND(__DECCXX_VER, BOOST_TESTED_AT(60590041)) +// Needed because literal_string must not be const, you get multiple defined +// symbols otherwise. (see 'basic_cstring_fwd.hpp') As an alternative, +// 'c_literal_string' could be replaced by 'literal_string const' as well. +c_literal_string LOG_LEVEL = "BOOST_TEST_LOG_LEVEL"; // --log_level +c_literal_string NO_RESULT_CODE = "BOOST_TEST_RESULT_CODE"; // --result_code +c_literal_string REPORT_LEVEL = "BOOST_TEST_REPORT_LEVEL"; // --report_level +c_literal_string TESTS_TO_RUN = "BOOST_TESTS_TO_RUN"; // --run_test +c_literal_string SAVE_TEST_PATTERN = "BOOST_TEST_SAVE_PATTERN"; // --save_pattern +c_literal_string BUILD_INFO = "BOOST_TEST_BUILD_INFO"; // --build_info +c_literal_string CATCH_SYS_ERRORS = "BOOST_TEST_CATCH_SYSTEM_ERRORS"; // --catch_system_errors +c_literal_string REPORT_FORMAT = "BOOST_TEST_REPORT_FORMAT"; // --report_format +c_literal_string LOG_FORMAT = "BOOST_TEST_LOG_FORMAT"; // --log_format +c_literal_string OUTPUT_FORMAT = "BOOST_TEST_OUTPUT_FORMAT"; // --output_format +#else +literal_string LOG_LEVEL = "BOOST_TEST_LOG_LEVEL"; // --log_level +literal_string NO_RESULT_CODE = "BOOST_TEST_RESULT_CODE"; // --result_code +literal_string REPORT_LEVEL = "BOOST_TEST_REPORT_LEVEL"; // --report_level +literal_string TESTS_TO_RUN = "BOOST_TESTS_TO_RUN"; // --run_test +literal_string SAVE_TEST_PATTERN = "BOOST_TEST_SAVE_PATTERN"; // --save_pattern +literal_string BUILD_INFO = "BOOST_TEST_BUILD_INFO"; // --build_info +literal_string CATCH_SYS_ERRORS = "BOOST_TEST_CATCH_SYSTEM_ERRORS"; // --catch_system_errors +literal_string REPORT_FORMAT = "BOOST_TEST_REPORT_FORMAT"; // --report_format +literal_string LOG_FORMAT = "BOOST_TEST_LOG_FORMAT"; // --log_format +literal_string OUTPUT_FORMAT = "BOOST_TEST_OUTPUT_FORMAT"; // --output_format +#endif + +enum report_level { CONFIRMATION_REPORT, SHORT_REPORT, DETAILED_REPORT, NO_REPORT, UNDEF_REPORT }; + +#if BOOST_WORKAROUND(__DECCXX_VER, BOOST_TESTED_AT(60590041)) +// See comment on workaround above. +c_literal_string report_level_names[] = { "confirm" , "short" , "detailed" , "no" }; +#else +literal_string report_level_names[] = { "confirm" , "short" , "detailed" , "no" }; +#endif enum output_format { HRF /* human readable format */, XML /* XML */ };

1 0

Fw: Re your copyright in Boost C++ Library
by John Maddock 17 Aug '04

17 Aug '04

And yet another permission.... ----- Original Message ----- From: "Lars Gullik Bjønnes" <larsbj(a)lyx.org> To: "John Maddock" <john(a)johnmaddock.co.uk> Sent: Tuesday, August 17, 2004 1:57 PM Subject: Re: Re your copyright in Boost C++ Library > "John Maddock" <john(a)johnmaddock.co.uk> writes: > > >> | Hi, if you are receiving this message it is because there are one or > | more > >> | files in the Boost C++ libraries that contain your copyright > | declaration. > >> > >> I belive that my name (and copyright) should just be removed from > >> those files. > > > | Are you sure? > > If it is ok by you, yes. > (if not just add me to the blanket file. Or do that anyway...) > > > > | You can view the two files here: > > > | http://cvs.sourceforge.net/viewcvs.py/boost/boost/boost/detail/atomic_count… > | http://cvs.sourceforge.net/viewcvs.py/boost/boost/boost/detail/lwm_gcc.hpp?… > > > | cvs revision history suggests that the implementation originally came from > | you? > > It is really something of an super simple adaptation of code that was > already in boost. And I belive that this has be rewritten a bit since > even. (and it is even buggy... or so they say :-) ) > > -- > Lgb >

1 0