libs/spirit/doc/x3/changelog.qbk
[/==============================================================================
Copyright (C) 2018-2021 Nikita Kniazev
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)
===============================================================================/]
[section Changelog]
[/////////////////////////////////////////////////////////////////////////////]
[section:spirit_3_10 Spirit V3.10 (Boost V1.78.0)]
* Fortified against `min` and `max` macros. [gh_pr 697]
* Producing a variant attribute in an alternative parser branch was leading to a
compilation error when parsed into a superset variant attribute due to
swapped parameters to `is_substitute` trait. [gh_pr 702] [gh 701] [gh 679]
[endsect]
[/////////////////////////////////////////////////////////////////////////////]
[section:spirit_3_0_9 Spirit V3.0.9 (Boost V1.77.0)]
*Breaking change*: Optional parser no longer unflattens sequence, parsers like
`int_ >> -(int_ >> int_)` into `tuple<int, optional<int>, optional<int>>`
will produce a compilation error in sequence parser. This was a bugfeature
which might be reimplemented in the future.
* Fixed list parser with attribute-less subject in sequence parser. [gh_pr 663]
* Fixed optional parser with container attribute in sequence parser. [gh_pr 665]
* Fixed `error_handler` newline/whitespace handling issues. [gh_pr 669] [gh_pr 670]
* Fixed `seek[...]` directive going past the end of input introduced in [gh_pr 30],
triggered when pre-skip leaves no data. [gh_pr 658] [gh_pr 683]
* Removed rollback in rules to fix a regression introduced in [gh_pr 670]. [gh_pr 686]
[endsect]
[/////////////////////////////////////////////////////////////////////////////]
[section:spirit_3_0_8 Spirit V3.0.8 (Boost V1.76.0)]
* Fixed assertion in standard parser with an input containing characters from
extended charset (on values above 0x7F) [gh_pr 620]
* Enhanced debug output to print unicode code points as hex [gh_pr 634]
* Skip rule definition binder creation in `parse_rule` [gh_pr 646]
* Replaced Boost.Math with std alternatives [gh_pr 648]
* Fixed unqualified `get<typename>` call in `parse_rule`
* Removed unused header `no_case_string_parse.hpp`
* Make literal binary parser constructor non-templated
* Fixed unqualified `get<typename>` call in `annotate_on_success::on_success` [gh 526]
* Fixed multiple times missed `force_attribute` parameter. Adresses compilation
errors and slow compilation due to rule definition injection into context
when `force_attribute` is set to `true` [gh_pr 651]
* Skip rule definition injection into context when a rule placeholder constructed
and immediately consumed (like `auto const grammar = rule<tag>{} = parser`) [gh_pr 650]
[endsect]
[/////////////////////////////////////////////////////////////////////////////]
[section:spirit_3_0_7 Spirit V3.0.7 (Boost V1.74.0)]
* Ceased dependence on Boost.Locale (replaced `locale::conv::utf_to_utf`
with `x3::to_utf8`), and Boost.Filesystem [gh_pr 596]
* Fixed infinity recursion in `is_substitute` trait that was blocking recursive
sequences via containers [gh 461] [gh_pr 606]
* Fixed sequence parser in alternative into container [gh 434] [gh_pr 605]
* Fixed some corner cases in alternative parser with container attribute,
notably atribute-less parsers in alternative parser with container attribute
were appending an empty value [gh 394] [gh 603] [gh_pr 604]
* Attribute type of alternative/sequence parsers now calculated without MPL,
what removes MPL limits on alternative/sequence parser chain length [gh 515] [gh_pr 607]
* Reverted [gh_pr 534] that made possible parsing `optional<T>` into `T` [gh_pr 608]
* Removed unneeded `reserve` before `insert` [gh_pr 611]
* Removed `value_initialize` trait [gh_pr 613]
* Removed unwanted move due to wrong `pseudo_attribute` usage, that was also
causing bogus GCC `-Wununinitlized` warning [gh_pr 613]
* Removed unused `boost/spirit/home/x3/support/utility/unrefcv.hpp` header
[endsect]
[/////////////////////////////////////////////////////////////////////////////]
[section:spirit_3_0_6 Spirit V3.0.6 (Boost V1.73.0)]
* ADL used to find and match `unused_type` streaming operators for every type
in `boost::spirit::x3` namespace [gh_pr 553] [gh 552]
* Ceased dependence on Boost.TTI
* Rule placeholders now a non-trivial type because its copy constructor checks
for a static initialization fiasco [gh_pr 571]
* Most parsers are now constexpr-constructible [gh_pr 570]
[endsect]
[/////////////////////////////////////////////////////////////////////////////]
[section:spirit_3_0_5 Spirit V3.0.5 (Boost V1.72.0)]
* `any_char` parser were accepting characters outside of encoding. [gh 520] [gh_pr 521] sehe
* Added workaround `move_to` overload for `boost::optional` type. [gh 533] [gh_pr 534] Duzy Chan
* Unicode out-of-bounds category lookup. [gh_pr 528] Christian Mazakas
* Mismatched pragma warning push/pop in `escape_char.ipp`. [gh 532] [gh_pr 535] stefan301
* Remove bool->char->bool conversion in `[no_]case_compare`. [gh_pr 536]
* `<cctype>` functions were called with negative values. [gh_pr 543]
* Fixed `parse_rule` definition check. [gh_pr 549] (regressed in [gh_pr 457])
[endsect]
[/////////////////////////////////////////////////////////////////////////////]
[section:spirit_3_0_4_boost_1_71 Spirit V3.0.4 (Boost V1.71.0)]
* Explicitly delete all non implemented assignment operators. [gh_pr 489] Romain Geissler
* Unpublish `x3/support/utility/testing.hpp`. [gh_pr 492]
* Removed unused `is_char` and `is_string` traits. [gh_pr 507]
* Simplified `string_traits`. [gh_pr 507]
* `raw` parser attribute materialization issue. [gh 509]
* Removed static assert guarding unwanted attribute materialization that
happens when rule parser producing an attribute called without an actual
attribute. See [gh 511] for the discussion.
[endsect]
[/////////////////////////////////////////////////////////////////////////////]
[section:spirit_3_0_4 Spirit V3.0.4 (Boost V1.70.0)]
* Integer parsers should now handle user defined types. [gh_pr 429]
*[/^] [*Note:] The `check_overflow` trait default implementation now relies on
`std::numeric_limits<T>::is_bounded` instead of `boost::integer_traits<T>::is_integral`.
* Removed use of deprecated `boost/detail/iterator.hpp` header. [gh_pr 432]
* Removed sequence into plain parsing. Now it triggers a compile time error
instead of silently parsing the the sequence and taking the last value as a result. [gh_pr 439]
* Fixed `parse_rule` instantiation with `BOOST_SPIRIT_INSTANTIATE` when:
* A rule has no attribute. [gh_pr 455]
* An actual attribute is not of type a rule was declared with. [gh_pr 456] [gh 457]
*[/^] A [*huge] thanks goes out to [@https://github.com/Xeverous Xeverous]
for reporting a regression in [gh 453].
* Fixed unneeded attribute synthesization and no-transformation attribute
reference pass-through in rules. [gh 444] [gh_pr 449] [gh_pr 452]
* Removed broken 1.0#INF parser. It was poorly documented and never worked. [gh 415] [gh_pr 458] [trac 8699]
* The undocumented `make_attribute` trait was removed due to bugs [gh_pr 449]
and to simplify attribute transformation. [gh_pr 460]
*[/^] If you were using it to workaround bugs in attribute transformation -
they should not be needed anymore as of this release.
* The integer value parser now respects `std::numeric_limits<T>::digits10` value. [gh_pr 469]
* Fixed underflow check for a `(Min % Base) == 0` corner case. [gh_pr 469]
* Fixed an output value on overflow of `IgnoreOverflowDigits=true` integer parser. [gh_pr 470]
* Container attribute elements were copyied, but not moved. [gh_pr 472]
* Special handling of references in `transform_attribute` was removed. [gh_pr 480]
[endsect]
[/////////////////////////////////////////////////////////////////////////////]
[section:spirit_3_0_3 Spirit V3.0.3 (Boost V1.69.0)]
* Drop own FP routines in favor of `boost::math`. [gh_pr 392] [trac 13531]
* Missing visibility mark on exception types. [gh_pr 409]
* to_utf8: Fixed `wchar_t` handling on Windows. [gh_pr 413] [gh 395]
[endsect]
[/////////////////////////////////////////////////////////////////////////////]
[section:spirit_3_0_2 Spirit V3.0.2 (Boost V1.68.0)]
* Small list parser optimization. [gh_pr 368]
* Pass container attributes through sequence parser or unary
ending down to a sequence parser. [gh_pr 370] [trac 12085]
* More fine grained sequence attribute check message. [gh_pr 371]
* Removed redundant check in `skip_over`. [gh_pr 373]
* Workaround constexpr in noexcept VS2015 bug in entire `x3::variant`. [gh_pr 379]
* Fixed `calc4b` example compilation. [gh_pr 384]
* Minor code improvements. [gh_pr 374]
[endsect]
[/////////////////////////////////////////////////////////////////////////////]
[section:spirit_3_0_1 Spirit V3.0.1 (Boost V1.67.0)]
There was no version bump and changelog for Boost V1.56.0-1.66.0 releases.
* Removed `with_context`. [gh_pr 239]
* Added noexcept to `x3::variant` and `forward_ast`. [gh_pr 241]
* CR+LF lines wrongly counted in `error_handler::position()`. [gh_pr 248]
* Fixed `unused_type` attribute case in `parse_into_container`. [gh_pr 266]
* Fixed parsing into associative containers. [gh_pr 289]
* Fixed overflow problem in `uint_parser<signed T>`. [gh_pr 297]
* Added VS2015 Update 3 support by using workarounds. [gh_pr 308]
* Fixed include guard names collision with Qi. [gh_pr 313]
* Added parsing into range. [gh_pr 316] [trac 12928]
* Changed iterator concept static assert from `ForwardIterator` to
`ReadableIteratorConcept && ForwardTraversalConcept`. [gh_pr 320]
* Reenabled `fusion::map` support. [gh_pr 330]
* Fixed `string("...")` and `attr("...")` to single item sequence. [gh_pr 337]
* Dereference a single item view instead of unwrapping sequence. [gh_pr 340]
* Prevent `parse_nan` from dereferencing out of range iterator. [gh_pr 351]
* Use traits to test if container is empty. [gh_pr 355]
[endsect]
[endsect]
