Spirit Change Log

1.8.7

• Integrated the Spirit V1.8.x code base with Spirit V2. Spirit V1.8.x is now called Spirit Classic. Even if the directory structure has changed (the Spirit Classic headers are now moved to the '$BOOST_ROOT/boost/spirit/home/classic' directory), we created forwarding headers allowing to compile existing applications without any change. These forwarding headers are deprecated, though, which will result in corresponding warnings generated for each of the headers. The forwarding headers are expected to be removed in the future.
  The recommended way of using Spirit Classic is now to include header files from the directory '$BOOST_ROOT/boost/spirit/include'. All files of Spirit Classic have now a 'classic_' prefixed to their name. For example the include
  
    #include <boost/spirit/core/core.hpp>
  
  now should be written as:
  
    #include <boost/spirit/include/classic_core.hpp>
  
  To avoid namespace conflicts with the new Spirit V2 library we moved Spirit Classic into the namespace boost::spirit::classic. This change will be automatically deactivated whenever the deprecated include files are being used. This ensures full backwards compatibility for existing applications.
  For more details about this change please consult the documentation.

1.8.6

• Fixed a integer overflow bug preventing to fail parsing on certain large integers. This bug was reported and fixed by Michael Andersen Nexø

1.8.5

• For performance reasons, leaf_node_d/token_node_d have been changed to implicit lexems that create leaf nodes in one shot. The old token_node_d is still available and called reduced_node_d, now.
• It's now possible to phoenix::bind (version 1) symbols::add.

1.8.4

• Fixed no_actions bug where no_action is applied recursively.
• Fixed the regex_p parser for Boost >= V1.33.0
• Implemented a workaround for namespace issues VC++ has with Spirit's file_iterators
• Fixed bug in tree match policies that prevented using gen_pt/ast_node_d, reported by Jascha Wetzel.
• Made position_iterator usable with wchar_t based strings.

1.8.3

• Config correction for Sun C++ by Steve Clamage (see this link).
• Fixed multi_pass_iterator for 64 platforms, where sizeof(int) != sizeof(ptr_type).Fixed bug that prevents the use of closures with grammars with multiple entry points, reported by David Pierre
• Fixed bug that prevented embedding of grammars with multiple entry points, reported by David Pierre
• Added '\0' to the set of valid escaped characters for escape_ch_p.
• Fixed a switch_p bug when used with a phoenix::actor as the conditional expression.
• __LINE__ macro now gets expanded in BOOST_SPIRIT_ASSERT_EXCEPTION
• Fixed a bug in the intersection parser reported by Yusaku Sugai
• The symbol parser uses the null character internally. Checks were added so that:
  • tst.add asserts if string contains the null character
  • tst.find doesn't match null characters in the input
• Fixed match_attr_traits.ipp to allow non-POD to pass through. The previous version taking in the ellipsis "..." does not allow PODs to pass through.
• Allow evaluation to int as condition to if_p parser.
• Applied performance improvement changes to the PT/AST code as suggested by Stefan Slapeta.
• Fixed several problems with AST tree node directives (inner_node_d[], discard_first_node[], discard_last_node[] and infix_node_d[]).

1.8.2

Maintenance release (almost the same as 1.8.1 plus a few fixes here and there)

• Added specializations to str_p and ch_p to allow str_p('c') and ch_p("c") thus fixing some non-bugs
• Fixed bug where a match<T> is a variant.
• added Jamfile/Jamrules from CVS to spirit-1.8.1/
• added boost-build.jam from boost to spirit-1.8.1/
• disabled template multi-threading in libs/spirit/test/Jamfile
• added a boost-header-include rule (from spirit-header-include) pointing to miniboost in libs/spirit/test/Jamfile
• Fixed if_p inconsistency

1.6.2

The Spirit 1.6.2 release is a bug-fix release only, no new features were introduced.

• wchar_t friendly implementation of graph_p
• Modified escape_char_parser::parse() to use a static parser instead of a rule. This will make it more friendly to use in trees. It should also be a little more efficient.
• Moved to Boost Software license 1.0.
• workaround for Error 322 name lookup in base class specialization finds type
• fixed limit_d bug
• [numerics] Workaround for aC++
• Fixed a bug in the switch_p parser.
• Fixed a EOI problem in multi_pass
• added Jamfile/Jamrules from CVS to spirit-1.6.1/
• added boost-build.jam from boost to spirit-1.6.1/
• disabled template multi-threading in libs/spirit/test/Jamfile
• added a boost-header-include rule (from spirit-header-include) pointing to miniboost in libs/spirit/test/Jamfile

1.8.1 (Released with Boost 1.32.0)

The Spirit 1.8.1 release is a bug-fix release only, no new features were introduced.

• Spirit now requires at least Boost 1.32.0 to compile correctly
• Removed the support for the older iterator adaptor library and
• Moved to use the new MPL library
• Spirit was moved to use the Boost Software License 1.0.
• Fixed several parsers to support post-skips more correctly.
• Fixed a no_node_d[] bug.
• Fixed a bug in shortest_d[].
• Fixed a bug in limit_d[].
• Fixed parser traversal meta code.
• Fixed several bugs in switch_p.
• Fixed AST generating problems, in particular with the loops related parsers.
• Fixed several bugs in the multi_pass iterator.
  • Fixed a problem, when the used base iterator returned a value_type and not a reference from its dereferencing operator.
  • Fixed iterator_traits problem
  • Fixed an EOI problem
  • Fixed a bug, when used with std::cin
• Found a bug in grammar.ipp when BOOST_SPIRIT_SINGLE_GRAMMAR_INSTANCE is defined
• Rewritten safe_bool to use CRTP - now works also on MWCW, fixed several bugs with the implementation.
• Fixed and extended the debug diagnostics printed by the parse tree code.

1.8.0 (Released with Boost 1.31.0; Includes unreleased 1.7.1)

• Fixed a wchar_t problem in the regex_p parser.
• removed code and workarounds for old compilers (VC6/7 and Borland)
• Changed license to the new boost license.
• Modified escape_char_parser::parse() to use a static parser instead of a rule. This will make it more friendly to use in trees. It should also be a little more efficient.

1.7.1 (Unreleased; becomes 1.8.0)

• Added a full suite of predefined actors.
• Moved rule_alias and stored_rule from core/non-terminal to dynamic.
  Made as_parser a public API in meta/as_parser.hpp
• Separated Core.Meta into its own module
• Refactored Utility module
  Moved some files into Utility.Parsers

• utilities
  • parsers
    • chset, regex, escape_char
      confix, list, distinct
      functor_parser
• support
  • scoped_lock
    flush_multi_pass
    grammar_def
• actors
  • assign

• Stored rules
• Added the switch_p and select_p dynamic parsers.
• Multiple scanner support for rules.
• The Rule's Scanner, Context and Tag template parameters can be specified in any order now. If a template parameter is missing, it will assume the defaults. See test/rule_tests.cpp.
• Introduced the possibility to specify more than one start rule from a grammar.
• Added an implementation of the file_iterator iterator based on the new Boost iterator_adaptors (submitted originally by Thomas Witt).

[The transition to the new iterator_adaptors should be complete now.]

• Added an implementation of the fixed_size_queue iterator based on the new Boost iterator_adaptors.
• wchar_t friendly implementation of graph_p
• made the copy-constructor and assignment-operator of parser_error_base public to clear VC7.1 C4673 warning. Added copy-constructor and assignment operator to parser_error for clarity of intent.

1.7.0

• assign(string) semantic action now works in VC6
• parsers need not be default constructible
• simplified aggregation of binary and unary parsers (more compiler friendly)
• epsilon workarounds for VC++
• match's attribute now uses boost.optional
• subrules can now have closures
• project wide 64 bit compatibility
• dynamic_parser_tag, reissue of rule.set_id(id);
• numerous primitives improvements and workarounds for problematic compilers
• proper complement (~) of single char parser primitives and chsets
• intuitive handling of lexeme_d
• wide_phrase_scanner_t typedef
• dynamic parser improvements (better support for more compilers)
• complete rewrite of the file_iterator (using boost.iterator_adapters). Supports memory maps wherever available
• multi_pass updates (compatibility with more compilers (e.g VC7) and more)
• position_iterator improvements
• better phoenix support for more compilers
• phoenix new_(...) construct
• new lazy_p parser
• utility.distinct parser (undocumented)
• chset operators improvements
• confix_p streamlining and improvements
• numerous Boost integration improvements

Bug fixes (1.7.0 and 1.6.0)

• Fixed. Using MSVC++6 (SP5), calling the assign action with a string value on parsers using the file_iterator will not work.
• Fixed: using assign semantic action in a grammar with a multi_pass iterator adaptor applied to an std::istream_iterator resulted in a failure to compile under msvc 7.0.
• Fixed: There is a bug in the "range_run<CharT>::set (range<CharT> const& r)" function in the "boost\spirit\utility\impl\chset\range_run.ipp".
• Fixed: handling of trailing whitespace bug (ast_parse/pt_parse related)
• Fixed: comment_p and end of data bug
• Fixed: Most trailing space bug:
• Fixed:
  chset<>::operator~(range<>) bug
  operator&(chset<>, range<>) bug
  operator&(range<>, chset<>) bug
• Fixed: impl::detach_clear bug
• Fixed: mismatch closure return type bug
• Fixed: access_node_d[] and access_match_d[] iterator bugs
• Fixed a bug regarding threadsafety of Phoenix/Spirit closures.
• Added missing include files to miniboost

Copyright © 1998-2005 Joel de Guzman, Hartmut Kaiser
Use, modification and distribution is subject to 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)