- Boost - lists.boost.org

Boost inspection notification (2007-03-18/RC_1_34_0) *X*
by Rene Rivera 18 Mar '07

18 Mar '07

Boost Inspection Report Run Date: 16:04:16 UTC, Sunday 18 March 2007 An inspection program <http://www.boost.org/tools/inspect/index.html> checks each file in the current Boost CVS for various problems, generating this as output. Problems detected include tabs in files, missing copyrights, broken URL's, and similar misdemeanors. Totals: 11447 files scanned 908 directories scanned (including root) 196 problems reported Problem counts: 1 files with invalid line endings 0 bookmarks with invalid characters 3 invalid urls 104 broken links 21 unlinked files 22 file/directory names issues 1 files with tabs 9 violations of the Boost min/max guidelines 35 usages of unnamed namespaces in headers (including .ipp files) Summary: archive (3) bind (1) boost-root (3) build (2) date_time (1) doc (2) filesystem (37) graph (2) iostreams (3) lambda (3) math (8) more (1) mpl (1) multi_array (2) ptr_container (1) python (10) regex (1) regression (32) serialization (1) test (81) type_traits (1) Details: *R* invalid (cr only) line-ending *A* invalid bookmarks, invalid urls, broken links, unlinked files *N* file/directory names issues *T* tabs in file *M* uses of min or max that have not been protected from the min/max macros, or unallowed #undef-s *U* unnamed namespace in header |archive| boost/archive/basic_streambuf_locale_saver.hpp: *N* name exceeds 31 characters boost/archive/impl/xml_wiarchive_impl.ipp: *U* unnamed namespace at line 53 boost/archive/iterators/remove_whitespace.hpp: *U* unnamed namespace at line 57 |bind| boost/bind/placeholders.hpp: *U* unnamed namespace at line 25 |boost-root| index.htm: *A* broken link: doc/hmtl/date_time.html *A* broken link: doc/hmtl/date_time/details.html#changes *A* broken link: more/BoostCon07.html |build| tools/build/v2/test/test_system.html: *A* unlinked file tools/build/v2/tools/python.jam: *R* |date_time| libs/date_time/xmldoc/date_time_docs_howto.html: *A* unlinked file |doc| doc/html/boost_math/inverse_complex.html: *A* unlinked file doc/html/jam.html: *A* unlinked file |filesystem| libs/filesystem/doc/do-list.htm: *A* invalid URL (hardwired file): file://?/ *A* invalid URL (hardwired file): file://?/UNC/ libs/filesystem/doc/i18n.html: *A* broken link: convenience.htm#basic_recursive_directory_iterator *A* broken link: exception.htm *A* broken link: operations.htm *A* broken link: operations.htm#Do-the-right-thing *A* broken link: operations.htm#is_directory *A* broken link: operations.htm#is_file *A* broken link: operations.htm#status libs/filesystem/doc/index.htm: *A* broken link: ../build/Jamfile *A* broken link: convenience.htm *A* broken link: fstream.htm *A* broken link: operations.htm#create_directory *A* broken link: operations.htm#create_hard_link *A* broken link: operations.htm#current_path *A* broken link: operations.htm#directory_iterator *A* broken link: operations.htm#equivalent *A* broken link: operations.htm#file_size *A* broken link: operations.htm#initial_path *A* broken link: operations.htm#is_file *A* broken link: operations.htm#is_symlink *A* broken link: operations.htm#status *A* broken link: operations.htm#symlink_status *A* broken link: path.htm#Canonical *A* broken link: path.htm#Grammar *A* broken link: path.htm#Normalized *A* broken link: path.htm#default_name_check *A* broken link: path.htm#name_check_mechanism *A* broken link: path.htm#normalize *A* broken link: path.htm#operator_eq *A* broken link: path.htm#synopsis libs/filesystem/doc/portability_guide.htm: *A* broken link: path.htm#name_check_typedef libs/filesystem/doc/tr2_proposal.html: *A* invalid URL (hardwired file): file:///C|/boost/site/libs/filesystem/doc/operations.htm#complete_note |graph| libs/graph/example/file_dependencies.cpp: *M* violation of Boost min/max guidelines on line 139 libs/graph/test/serialize.cpp: *T* |iostreams| libs/iostreams/doc/acknowledgments.html: *A* unlinked file libs/iostreams/doc/concepts/multi-character.html: *A* unlinked file libs/iostreams/doc/installation.html: *A* broken link: ../../../tools/build/v1/build_system.htm |lambda| boost/lambda/core.hpp: *U* unnamed namespace at line 62 boost/lambda/detail/lambda_functors.hpp: *U* unnamed namespace at line 25 boost/lambda/exceptions.hpp: *U* unnamed namespace at line 24 |math| libs/math/test/common_factor_test.cpp: *M* violation of Boost min/max guidelines on line 146 *M* violation of Boost min/max guidelines on line 147 *M* violation of Boost min/max guidelines on line 193 *M* violation of Boost min/max guidelines on line 194 |more| more/separate_compilation.html: *A* broken link: ../libs/config/test/link/test/Jamfile.v2 |mpl| boost/mpl/alias.hpp: *U* unnamed namespace at line 17 |multi_array| boost/multi_array/base.hpp: *U* unnamed namespace at line 69 libs/multi_array/test/generative_tests.hpp: *U* unnamed namespace at line 57 |ptr_container| libs/ptr_container/doc/tutorial_example.html: *A* unlinked file |python| libs/python/doc/building.html: *A* broken link: ../../../tools/build/v1/build_system.htm *A* broken link: ../../../tools/build/v1/build_system.htm#user_globals *A* broken link: ../../../tools/build/v1/build_system.htm#variants libs/python/doc/tutorial/doc/html/python/hello.html: *A* broken link: ../../../../../example/tutorial/Jamrules libs/python/doc/tutorial/index.html: *A* broken link: ../../../LICENSE_1_0.txt libs/python/doc/v2/May2002.html: *A* broken link: ../../../../tools/build/v1/build_system.htm *A* broken link: ../../../../tools/build/v1/gen_aix_import_file.py libs/python/doc/v2/faq.html: *A* broken link: ../../../../tools/build/v1/build_system.htm |regex| libs/regex/performance/input.html: *A* unlinked file |regression| regression/.htaccess: *N* leading character of ".htaccess" is not alphabetic tools/regression/test/test-cases/incremental/bjam.log.1: *N* name contains more than one dot character ('.') tools/regression/xsl_reports/test/test-components/test_is_test_log_complete/compile-fail.xml: *N* file path will exceed 100 characters in a directory tree with a root of the form boost_X_XX_X/ tools/regression/xsl_reports/test/test-components/test_is_test_log_complete/completed.xml: *N* file path will exceed 100 characters in a directory tree with a root of the form boost_X_XX_X/ tools/regression/xsl_reports/test/test-components/test_is_test_log_complete/expected/compile-fail.xml: *N* file path will exceed 100 characters in a directory tree with a root of the form boost_X_XX_X/ tools/regression/xsl_reports/test/test-components/test_is_test_log_complete/expected/completed.xml: *N* file path will exceed 100 characters in a directory tree with a root of the form boost_X_XX_X/ tools/regression/xsl_reports/test/test-components/test_is_test_log_complete/expected/lib.xml: *N* file path will exceed 100 characters in a directory tree with a root of the form boost_X_XX_X/ tools/regression/xsl_reports/test/test-components/test_is_test_log_complete/expected/misfire.xml: *N* file path will exceed 100 characters in a directory tree with a root of the form boost_X_XX_X/ tools/regression/xsl_reports/test/test-components/test_is_test_log_complete/expected/no-run.xml: *N* file path will exceed 100 characters in a directory tree with a root of the form boost_X_XX_X/ tools/regression/xsl_reports/test/test-components/test_is_test_log_complete/expected/run_pyd.xml: *N* file path will exceed 100 characters in a directory tree with a root of the form boost_X_XX_X/ tools/regression/xsl_reports/test/test-components/test_is_test_log_complete/expected/test-case.xml: *N* file path will exceed 100 characters in a directory tree with a root of the form boost_X_XX_X/ tools/regression/xsl_reports/test/test-components/test_is_test_log_complete/test-case.xml: *N* file path will exceed 100 characters in a directory tree with a root of the form boost_X_XX_X/ tools/regression/xsl_reports/test/test-components/test_is_test_log_complete/test-driver.xsl: *N* file path will exceed 100 characters in a directory tree with a root of the form boost_X_XX_X/ tools/regression/xsl_reports/test/test-components/test_re_match/expected/test_re_match.xml: *N* file path will exceed 100 characters in a directory tree with a root of the form boost_X_XX_X/ tools/regression/xsl_reports/test/test-components/test_test_case_status/explicit-failures-markup.xml.test: *N* file path will exceed 100 characters in a directory tree with a root of the form boost_X_XX_X/ *N* name contains more than one dot character ('.') *N* name exceeds 31 characters tools/regression/xsl_reports/test/test-components/test_test_structure/explicit-failures-markup.xml.test: *N* file path will exceed 100 characters in a directory tree with a root of the form boost_X_XX_X/ *N* name contains more than one dot character ('.') *N* name exceeds 31 characters tools/regression/xsl_reports/xsl/html/issues_legend.html: *A* unlinked file tools/regression/xsl_reports/xsl/html/library_developer_legend.html: *A* unlinked file tools/regression/xsl_reports/xsl/html/library_user_legend.html: *A* unlinked file tools/regression/xsl_reports/xsl/html/make_tinyurl.html: *A* unlinked file tools/regression/xsl_reports/xsl/html/summary_developer_legend.html: *A* unlinked file tools/regression/xsl_reports/xsl/html/summary_user_legend.html: *A* unlinked file tools/regression/xsl_reports/xsl/v2/html/issues_legend.html: *A* unlinked file tools/regression/xsl_reports/xsl/v2/html/library_developer_legend.html: *A* unlinked file tools/regression/xsl_reports/xsl/v2/html/library_user_legend.html: *A* unlinked file tools/regression/xsl_reports/xsl/v2/html/make_tinyurl.html: *A* unlinked file tools/regression/xsl_reports/xsl/v2/html/summary_developer_legend.html: *A* unlinked file tools/regression/xsl_reports/xsl/v2/html/summary_user_legend.html: *A* unlinked file |serialization| libs/serialization/src/basic_xml_grammar.ipp: *U* unnamed namespace at line 43 |test| boost/test/floating_point_comparison.hpp: *U* unnamed namespace at line 206 *U* unnamed namespace at line 228 boost/test/impl/cpp_main.ipp: *U* unnamed namespace at line 42 boost/test/impl/exception_safety.ipp: *U* unnamed namespace at line 400 boost/test/impl/framework.ipp: *U* unnamed namespace at line 199 boost/test/impl/plain_report_formatter.ipp: *U* unnamed namespace at line 45 boost/test/impl/progress_monitor.ipp: *U* unnamed namespace at line 38 boost/test/impl/results_collector.ipp: *U* unnamed namespace at line 106 boost/test/impl/results_reporter.ipp: *U* unnamed namespace at line 48 boost/test/impl/unit_test_log.ipp: *U* unnamed namespace at line 79 boost/test/impl/unit_test_monitor.ipp: *U* unnamed namespace at line 35 boost/test/impl/unit_test_parameters.ipp: *U* unnamed namespace at line 50 boost/test/results_collector.hpp: *U* unnamed namespace at line 40 boost/test/test_tools.hpp: *U* unnamed namespace at line 262 boost/test/utils/iterator/token_iterator.hpp: *U* unnamed namespace at line 166 boost/test/utils/named_params.hpp: *U* unnamed namespace at line 216 boost/test/utils/runtime/cla/dual_name_parameter.ipp: *U* unnamed namespace at line 43 boost/test/utils/runtime/cla/modifier.hpp: *U* unnamed namespace at line 34 boost/test/utils/runtime/env/modifier.hpp: *U* unnamed namespace at line 34 boost/test/utils/runtime/file/config_file.hpp: *U* unnamed namespace at line 169 *U* unnamed namespace at line 64 *U* unnamed namespace at line 74 boost/test/utils/runtime/file/config_file_iterator.hpp: *U* unnamed namespace at line 68 boost/test/utils/trivial_singleton.hpp: *U* unnamed namespace at line 52 *U* unnamed namespace at line 61 libs/test/build/msvc71_proj/config_file_iterator_test.vcproj: *N* name exceeds 31 characters libs/test/doc/components/prg_exec_monitor/compilation.html: *A* broken link: ../../../build/Jamfile libs/test/doc/components/prg_exec_monitor/index.html: *A* broken link: ../../../../../boost/test/cpp_main.hpp libs/test/doc/components/test_tools/index.html: *A* broken link: ../../tests/boost_check_equal_str.html libs/test/doc/components/test_tools/reference/BOOST_CHECK_CLOSE.html: *A* broken link: BOOST_CHECK_CLOSE_FRACTION.html libs/test/doc/components/test_tools/reference/BOOST_CHECK_MESSAGE.html: *A* broken link: BOOST_MESSAGE.html libs/test/doc/components/test_tools/reference/BOOST_CHECK_SMALL.html: *A* broken link: BOOST_CHECK_CLOSE_FRACTION.html libs/test/doc/components/test_tools/reference/tools_list.html: *A* broken link: ../../btl1.gif *A* broken link: BOOST_CHECK_CLOSE_FRACTION.html libs/test/doc/components/utf/compilation.html: *A* broken link: ../../../build/Jamfile libs/test/doc/components/utf/components/index.html: *A* broken link: ../../btl1.gif libs/test/doc/components/utf/components/test_case/abstract_interface.html: *A* broken link: ../../../btl1.gif libs/test/doc/components/utf/components/test_case/auto_register_facility.html: *A* broken link: ../../../btl1.gif libs/test/doc/components/utf/components/test_case/boost_function_tc.html: *A* broken link: ../../../../../../../boost/test/unit_test_suite_ex.hpp *A* broken link: ../../../btl1.gif libs/test/doc/components/utf/components/test_case/class_tc.html: *A* broken link: ../../../btl1.gif libs/test/doc/components/utf/components/test_case/function_tc.html: *A* broken link: ../../../btl1.gif libs/test/doc/components/utf/components/test_case/index.html: *A* broken link: ../../../btl1.gif libs/test/doc/components/utf/components/test_case/param_boost_function_tc.html: *A* broken link: ../../../../../../../boost/test/unit_test_suite_ex.hpp *A* broken link: ../../../btl1.gif libs/test/doc/components/utf/components/test_case/param_class_tc.html: *A* broken link: ../../../btl1.gif libs/test/doc/components/utf/components/test_case/param_function_tc.html: *A* broken link: ../../../btl1.gif libs/test/doc/components/utf/components/test_case/tc_template.html: *A* broken link: ../../../btl1.gif libs/test/doc/components/utf/components/test_log/custom_log_formatter.html: *A* broken link: ../../../btl1.gif libs/test/doc/components/utf/components/test_log/index.html: *A* broken link: ../../../btl1.gif libs/test/doc/components/utf/components/test_result/index.html: *A* broken link: ../../../btl1.gif libs/test/doc/components/utf/components/test_suite/index.html: *A* broken link: ../../../btl1.gif libs/test/doc/components/utf/index.html: *A* broken link: getting_started/index.html libs/test/doc/components/utf/parameters/build_info.html: *A* broken link: ../../btl1.gif libs/test/doc/components/utf/parameters/catch_system_errors.html: *A* broken link: ../../btl1.gif libs/test/doc/components/utf/parameters/detect_memory_leaks.html: *A* broken link: ../../btl1.gif libs/test/doc/components/utf/parameters/index.html: *A* broken link: ../../btl1.gif libs/test/doc/components/utf/parameters/log_format.html: *A* broken link: ../../btl1.gif libs/test/doc/components/utf/parameters/log_level.html: *A* broken link: ../../btl1.gif libs/test/doc/components/utf/parameters/no_result_code.html: *A* broken link: ../../btl1.gif libs/test/doc/components/utf/parameters/output_format.html: *A* broken link: ../../btl1.gif libs/test/doc/components/utf/parameters/random.html: *A* broken link: ../../btl1.gif libs/test/doc/components/utf/parameters/report_format.html: *A* broken link: ../../../../../LICENSE_1_0.txt *A* broken link: ../../btl1.gif libs/test/doc/components/utf/parameters/report_level.html: *A* broken link: ../../btl1.gif libs/test/doc/components/utf/parameters/show_progress.html: *A* broken link: ../../btl1.gif libs/test/doc/examples/unit_test_example1.html: *A* broken link: ../../example/unit_test_example1.cpp libs/test/doc/examples/unit_test_example2.html: *A* broken link: ../../example/unit_test_example2.cpp libs/test/doc/examples/unit_test_example3.html: *A* broken link: ../../example/unit_test_example3.cpp libs/test/doc/examples/unit_test_example4.html: *A* broken link: ../../example/unit_test_example4.cpp libs/test/doc/examples/unit_test_example5.html: *A* broken link: ../../example/unit_test_example5.cpp *A* broken link: ../../example/unit_test_example5.input libs/test/doc/tests/auto_unit_test_test.html: *A* broken link: ../../test/auto_unit_test_test.cpp libs/test/doc/tests/auto_unit_test_test_mult.html: *A* broken link: ../../test/auto_unit_test_test_mult1.cpp *A* broken link: ../../test/auto_unit_test_test_mult2.cpp libs/test/doc/tests/unit_test_suite_ex_test.html: *A* broken link: ../../test/unit_test_suite_ex_test.cpp libs/test/doc/tutorials/hello_the_testing_world.html: *A* broken link: ../../../../../LICENSE_1_0.txt *A* broken link: ../execution_monitor/index.html libs/test/doc/tutorials/new_year_resolution.html: *A* broken link: ../../../../../../LICENSE_1_0.txt |type_traits| libs/type_traits/cxx_type_traits.htm: *A* unlinked file

1 0

[Report] 36 regressions on RC_1_34_0 (2007-03-18)
by Douglas Gregor 18 Mar '07

18 Mar '07

Boost Regression test failures Report time: 2007-03-17T04:53:32Z This report lists all regression test failures on release platforms. Detailed report: http://engineering.meta-comm.com/boost-regression/CVS-RC_1_34_0/developer/i… 36 failures in 4 libraries iostreams (6) optional (6) parameter (1) python (23) |iostreams| bzip2_test: msvc-7.1 msvc-8.0 gzip_test: msvc-7.1 msvc-8.0 zlib_test: msvc-7.1 msvc-8.0 |optional| optional_test: msvc-6.5 msvc-6.5 msvc-6.5_stlport4 msvc-7.0 optional_test_ref_fail2: msvc-7.1 msvc-8.0 |parameter| python_test: gcc-cygwin-3.4.4 |python| builtin_converters: gcc-3.4.5_linux gcc-4.0.3_linux gcc-4.1.0_linux gcc-4.1.0_linux_x86_64 import_: cw-9.4 gcc-mingw-3.4.2 gcc-mingw-3.4.5 intel-vc71-win-9.1 msvc-6.5 msvc-6.5 msvc-6.5_stlport4 msvc-7.0 msvc-7.1 msvc-7.1 msvc-7.1 msvc-7.1_stlport4 msvc-8.0 msvc-8.0 msvc-8.0 map_indexing_suite: gcc-3.4.5_linux gcc-4.0.3_linux gcc-4.1.0_linux gcc-4.1.0_linux_x86_64

1 0

SOC Boost.Process (whats happening)
by Shams 18 Mar '07

18 Mar '07

Hi, Does the mentor/developer know whats the status of this project? The last edit in trac shows it was about 4 months ago. https://www.boost-consulting.com:8443/trac/soc/browser/boost/soc/2006?order… Any progress so far? Thanks Shams --

3 2

[sorting] Radix quicksort examples - sort within sort
by Sam Schetterer 18 Mar '07

18 Mar '07

Earlier, I loaded an example of how normal radix sort is useful. However, radixsort is almost impossible to use on data with varying lengths, making it unsiuted for strings. Luckily, the same "sort within sort" effect can still be achieved through radix quicksort. Although radix_quicksort is a string sort, it can be used for data of any type by using the create_id_string function, which is like sprintf, but instead of printing literal values for numbers (33 becomes "33"), the bytes are printed onto the string(33 becomes '!' (33 is the ascii value for '!')). Therefore, you get a fast sort that can accomplish what is called "sort within sort" k times faster than normal quicksort, where k is the number of bytes in the string. (normal quicksort is k(2 N log N), while radix quicksort is k + 2 N log N). Unfortunantly, it is not obvious how to implement radix quicksort, so I loaded an example of how to create an id string and use radix quicksort.

2 1

[sorting] Documentation is outdated
by Sam Schetterer 18 Mar '07

18 Mar '07

For all of you who have downloaded the sorting library from vault, just a notice, but the documentation is pretty old and has not been updated. Please do not judge the content of the library by the documentation.

1 0

[sorting] Mistake in code
by Sam Schetterer 18 Mar '07

18 Mar '07

For those of you who downloaded the recent radix sort, where I ask why msvc gives 31 errors on radixsort in one spot, never mind that. I made a dumb mistake and forgot to make the variable temp an iterator instead of a raw pointer. Sorry.

1 0

Runtime Dynamic Dispatch Library (0.1.11)
by Dean Michael Berris 18 Mar '07

18 Mar '07

Hi Everyone, I've been working on this library for the past few months, and I've uploaded the zip archive to the vault ( http://tinyurl.com/2bu427 ). First, let me describe what the library is for, what problem it's supposed to solve, and how it's different from existing implementations. The dispatcher library aims to provide a homogeneous function registry associating keys to callbacks. The type of the key and the function signature are provided as template parameters. The function registry (or dispatcher) can be given an index validation and routing strategy to validate indexes during runtime. The dispatcher is not a container, but it does contain an implementation of a strategized_map which encapsulates an STL map and allows the user to define an index validation and routing strategy on top of the existing STL map implementation. The intended use of the library is given as an inlined example snippet (also discussed in the documentation also in the zip archive): using namespace boost::dispatch ; dispatcher<void(std::string)> d; // register some functions to d here... d[0] = a_function; d[1] = another_function; d[3] = some_other_function; // note index '2' is not defined invoke_(d)("Hello, World!") << 0 << 1 << alt_(2, 3); The intention is to be able to delay implementations to loaded dll's or external modules, while writing statically type safe functions like below: template <class DispatcherType> void some_function(DispatcherType & d) { using namespace boost::dispatch; try { invoke_(d)("a", "b", "c") // bind the three arguments to successive invocations << alt_("handler1", "handler2") << "cleanup"; } catch (std::exception & e) { std::cerr << e.what() << '\n'; }; }; // client code... { using namespace boost::dispatch; dispatch<void(std::string, std::string, std::string), std::string> d; // load the implementations using dlopen, or something like that d["handler2"] = loaded_function_2; d["cleanup"] = cleanup_function; some_function(d); } More details available in the contained documentation. Comments and suggestions (and some help with Boost.PP) will be very much appreciated. Have a great day everyone, and I certainly hope this library would be useful to others who need this functionality. -- Dean Michael C. Berris http://cplusplus-soup.blogspot.com/ mikhailberis AT gmail DOT com +63 928 7291459

1 0

Re: [boost] SQL library with full embedded SQL-syntax
by Phil Endecott 18 Mar '07

18 Mar '07

Lyfar Dmitriy wrote: > Phil Endecott wrote: >> Thinking just about SQL backends, the basic question is whether we want to >> >> (a) hide the SQL, making the database accesses look more like C++ >> function calls and data structures, or >> >> (b) make the SQL explicit, making the C++ code look more like SQL. >> >> I definitely vote for (a), and the possibility of moving an application >> to something like stxxl is one of my motivations for that. > > If you interested in such libraries -- take a look at database template > library(dtl) also. http://dtemplatelib.sourceforge.net/ Thanks, I hadn't seen that. > What do you think about containers overhead? In my experience, which is based on PostgreSQL, you can implement STL-compatible containers for results with essentially no overhead. The most ugly part, as I have mentioned before, is mapping from a row to a tuple or struct. As far as I can see, in dtl you can use a struct (good!) but you have to define how each field should be filled from the row, and they have some magic with operator== to do so. Interesting, but it looks a bit kludgy. > Does the (a) variant means that the parsing of the result of query is > something like parsing an Variant type? Is it always opportune? With my PostgreSQL interface, I ask for the backend to pass the results in binary format, and it tells me the type for each column. I then do any necessary conversion (it may need its endianness changing; date/time formats need some effort; strings are trivial) and check that the type is what the C++ code said it expected. This is low-overhead and safe. > The people who wants to have type-safety of queries and platform > independence will select the (b) variant, I hope :) I don't understand why you say that. I want (and have) type safety. I would have portability if it were possible, but it isn't. Regards, Phil.

1 0

[intrusive] Review
by Steven Watanabe 17 Mar '07

17 Mar '07

AMDG I think that Intrusive should be accepted into Boost. The documentation is good. I had no problems compiling with msvc 8.0 The design and implementation are pretty straightforward. I have needed embedded linked lists at least once in the last year. So, I definitely have a use for this library. intrusive.qbk line 136 [*Boost.Intrusive]container missing space intrusive.qbk lines 169-170 because the object can be destroyed before erasing it from the container. Dangling participle intrusive.qbk line 172 However swapping can be used to implement move-capabilities missing comma after however. intrusive.qbk line 263-264 A class that the user must add as a base class or as a member to its own class its should be his. A user is a person not a thing. intrusive.qbk line 306-307 it can store the number of stored elements to achieve constant-time size information, it can offer debugging facilities... Using store twice sounds a little odd. These are both complete sentences and should therefore be separated by a semicolon instead of a comma. islist is a terrible name. I automatically read this as a predicate (is X a list) the second template parameter of ilist_base_hook &c. could be an mpl::bool_ or better it could be unique type that expresses the meaning better. struct safe_mode {}; struct fast_mode {}; intrusive.qbk line 531 typedef boost::intrusive::ilist_base_hook<Foo>::value_traits<Foo> FooValueTraits; should be typedef boost::intrusive::ilist_base_hook<Tag>::value_traits<Foo> FooValueTraits; intrusive.qbk line 564 Sometimes a 'is-a' relationship a should be an intrusive.qbk line 689-690 That's why you better avoid them in public interfaces of libraries. "That's why you should avoid them in public interfaces of libraries."? ilist_hook.hpp line 121 ilist_base_hook& ilist_base_hook::operator=(const ilist_base_hook& ); I'm not sure that forbidding assignment of nodes that are linked is a good idea. ilist_hook.hpp line 163 bool ilist_base_hook::linked() const; Since this is only allowed in safe mode it should fail to compile instead of asserting at runtime. On further reflection the safe mode and non-safe mode variants are sufficiently different to warrant two specializations. config_begin line 12, config_end.hpp line 15 #ifdef _MSC_VER should be #ifdef BOOST_MSVC I suggest this implementation for the default size_type parameter. As it stands, ADL pulls in namespace std for the containers. (ADL pulls in namespace detail, too but there are no fully generic functions in namespace boost::intrusive::detail) struct default_size_type {}; template<class T> struct get_size_type { typedef T type }; template<> struct get_size_type<default_size_type> { typedef std::size_t type }; template< class ValueTraits , bool ConstantTimeSize = true , class SizeTypeParam = default_size_type> class ilist : private detail::size_holder<ConstantTimeSize, typename get_size_type<SizeTypeParam>::type> { typedef typename get_size_type<SizeTypeParam>::type SizeType; //... }; swap_nodes doesn't handle adjacent nodes correctly. This isn't a problem now but you might make a note of it just in case it does become a problem. For slist you might consider adding a pointer to the last element so that swap can execute in constant time. clone_from can use insert_after to be optimally efficient. I'm unclear about the usefulness of clone_from. First of all it doesn't look more efficient than what I would write by hand. Second, it only provides the basic guarantee although the strong guarantee can be achieved as follows for ilist (assuming that destroyer completely reverses the effects of cloner). template <class Cloner, class Destroyer> void clone_from(const ilist &src, Cloner cloner, Destroyer destroyer) { class cleanup { public: cleanup(Destroyer& d) : destroyer(d) {} ~cleanup() { value.clear_and_destroy(destroyer); } ilist value; Destroyer& destroyer; } temp(destroyer); for(const_iterator b(src.begin()), e(src.end()) ; b != e; ++b){ temp.value.push_back(*cloner(*b)); } this->swap(temp.value); } The principal use I can think of for clone_from is a container that can hold objects of different types. You might change doc_clone_from.cpp to use a polymorphic clone function instead of new so that it doesn't seem quite so contrived. member_hooks do not work with virtual inheritance. This blows up. #include <boost/intrusive/ilist.hpp> struct Y {}; struct X : virtual public Y { boost::intrusive::ilist_member_hook<X> hook; int i; }; typedef boost::intrusive::ilist_member_hook<X>::value_traits<&X::hook> value_traits; int main() { X x; boost::intrusive::ilist<value_traits> list; list.push_back(x); list.front().i = 1; } I don't think that current is a good name for the function that gets an iterator from the value_type. How about something a little more descriptive like get_iterator, iterator_from, or to_iterator? I would prefer to use CRTP for hooks template<class Derived, class Tag, bool SafeMode, class Pointer> struct ilist_base_hook; Or at least use a metafunction instead of a member template for the value_traits. template<class Hook, class T> struct base_hook_value_traits; The base class hooks should be in a nested namespace and be brought into namespace intrusive with using to prevent unwanted ADL. The BOOST_STATIC_ASSERT in ilist needs to be INTERNAL ONLYed. I compared the performance of sort for ilist on a std::random_shuffled set of one million ints with msvc 8.0. (code is attached) INTRUSIVE .51 sec LIST .89 sec LIST POINTER 1.14 sec VECTOR .20 sec VECTOR STABLE .23 sec VECTOR POINTER .46 sec VECTOR POINTER STABLE .42 sec In Christ, Steven Watanabe // sort_benchmark.cpp // // Copyright (c) 2007 // Steven Watanabe // // Distributed under the Boost Software License, Version 1.0. (See // accompanying file LICENSE_1_0.txt or copy at // http://www.boost.org/LICENSE_1_0.txt) #include <ctime> #include <vector> #include <list> #include <algorithm> #include <iostream> #include <boost/lambda/lambda.hpp> #include <boost/lambda/bind.hpp> #include <boost/intrusive/ilist.hpp> //#define INTRUSIVE //#define LIST #define VECTOR //#define POINTER #define STABLE struct Y {}; struct X { #ifdef INTRUSIVE boost::intrusive::ilist_member_hook<X> hook; #endif int i; bool operator<(const X& other) const { return(i < other.i); } X(int val) : i(val) {} }; #ifdef POINTER typedef X* x_type; #define PREDICATE *boost::lambda::_1 < *boost::lambda::_2 #define TRANSFORM &boost::lambda::_1 #else typedef X x_type; #define PREDICATE boost::lambda::_1 < boost::lambda::_2 #define TRANSFORM boost::lambda::_1 #endif #ifdef STABLE #define SORT std::stable_sort #else #define SORT std::sort #endif int main() { #ifdef LIST typedef std::list<x_type> list_t; #elif defined(VECTOR) typedef std::vector<x_type> list_t; #elif defined(INTRUSIVE) typedef boost::intrusive::ilist<boost::intrusive::ilist_member_hook<X>::value_traits<&X::hook> > list_t; #endif list_t list; std::vector<X> x; for(int i = 0; i < 1000000; ++i) { x.push_back(X(i)); } std::random_shuffle(x.begin(), x.end()); std::for_each(x.begin(), x.end(), boost::lambda::bind(&list_t::push_back, boost::lambda::var(list), TRANSFORM)); { std::clock_t start_time = std::clock(); #if defined(LIST) || defined(INTRUSIVE) list.sort(PREDICATE); #else SORT(list.begin(), list.end(), PREDICATE); #endif std::cout << "time for sort: " << static_cast<double>(std::clock() - start_time) / CLOCKS_PER_SEC << std::endl; } list_t::const_iterator iter(list.begin()); list_t::const_iterator end(list.end()); list.erase(list.begin(), list.end()); }

4 5

[sorting] How radixsort is useful
by Sam Schetterer 17 Mar '07

17 Mar '07

Many people are asking me how radixsort is useful. Well, I have just posted a program that does what is called a "sort within sort". It sorts a class containing an int array. The classes are first sorted by the first int in the array, and then the second. They are then outputed to the standard output in proper order. I also wrote an operator< to show how much slower that function is than operator[] for a "sort within sort". Also, if you have other ideas or questions about radix sort and radix quicksort, post them and I will write a sample program showing how radixsort and radix quicksort are useful for that purpose.

1 0