Changes

0.088		2023-02-28	T. R. Wyant
    Remove support for (**{ ... code ... }).  This was introduced in
    Perl 5.37.8 along with a single-splat version.  The double-splat
    version was removed without deprecation in Perl 5.37.9, so it is
    being removed without deprecation here as well, per my stated policy
    about development functionality.  The single-splat version still
    exists (and is documented) in Perl 5.37.9, and in this package.

0.087		2023-01-28	T. R. Wyant
    Add support for code in optimized regex, a.k.a. (*{...}).  This
    involved making the recognition of backtracking control more
    specific, since it also uses (?*...).

    If (*{...}) and (**{...}) are removed from Perl before Perl 5.38.0,
    support for them will be removed from this package.

0.086		2022-12-13	T. R. Wyant
    Add width(), which returns the number of characters matched. Note
    that an indefinite upper boumd is represented as IEEE 754 Inf if
    that appears to be supported; otherwise by a singleton object
    overloaded to allow stringification, numification, and numeric
    tests.

    Use width() to enhance the detection of variable-width look-behinds.

    Serious clean-up on accepts_perl() subsystem.

0.085		2022-04-16	T. R. Wyant
    Remove 'postderef' argument to PPIx::Regexp->new().  Postfix
    dereference is always recognized.

0.084		2022-04-02	T. R. Wyant
    Require PPI 1.238 for postfix deref support, and recode the
    postfix deref logic in terms of 1.238's functionality.

    Parse '@{[ ... ]}' as code, not interpolation.  This is more in line
    with what it actually represents, and allows correct versioning of
    postfix dereferences.  But is is an incompatible change.

0.083		2022-03-17	T. R. Wyant
    Correct and optimize the computation of logical column position (the
    one that takes account of tabs).

0.082		2021-11-29	T. R. Wyant
    Add --version to eg/predump, and document all options with double
    dashes.

    Silence 'uninitialized' warning generated by /(?<=.{35})/.
    Thanks to Brian Fraser for reporting this.

    Try to quell weird Win32 test failures which seem to occur only in
    tests where I am using 'use open' to put the standard handles into
    UTF-8 mode. The fix (I hope) is to do this to the Test::Harness
    handles at run time instead of to the standard handles at compile
    time.

    Add file CONTRIBUTING.

0.081		2021-10-22	T. R. Wyant
    Any use of the postderef argument is now fatal.

    Correct generation of 'provides' metadata.  Thanks to Favio Poletti
    for blogging
    https://github.polettix.it/ETOOBUSY/2021/06/15/the-real-pause-workaround/,
    and ultimately to Joel Berger for the pointer to
    https://metacpan.org/pod/CPAN::Meta::Spec#no_index

    Add YAPE::Regex to SEE ALSO

0.080		2021-04-16	T. R. Wyant
    All uses of the postderef argument to new() now warn.

0.079		2021-03-26	T. R. Wyant
    Get prerequisites up to snuff, and add xt/author/prereq.t to ensure
    they stay that way.

    Add rt.cpan.org back to bug reporting methods. Long live RT!

0.078		2021-01-28	T. R. Wyant
    Allow CPAN to index Script_Run, Atomic_Script_Run, since they made
    it into a production release.

    Allow {,3} and { 0 , 3 } as quantifiers, requiring at least Perl
    5.33.6. Previously these parsed as literals. This parse will be
    retracted if it does not make it into 5.34.0.

0.077		2021-01-14	T. R. Wyant
    Add Travis CI testing.

    Use GitHub as bug tracker. R.I.P. rt.cpan.org.

0.076		2020-11-28	T. R. Wyant
    Correct (I hope) detection of \K in nested assertions.

    Variable-length look-behind is version 5.029009.
    
    Look-behinds quantified longer than 255 characters are an error, and
    are made into unknown tokens or structures. I ended up refactoring
    the PPIx::Regexp::Token::GroupType class initialization for the
    latter two changes.

0.075		2020-10-08	T. R. Wyant
    Warn on first use of attribute 'postderef'.

0.074		2020-09-08	T. R. Wyant
    Remove PPIx::Regexp::StringTokenizer itself and all documentation
    referring to it or the 'parse' argument to PPIx::Regexp->new().

0.073		2020-07-28	T. R. Wyant
    Remove prototypes from testing subroutines defined in t/*.t.

0.072		2020-05-20	T. R. Wyant
    Drop dependency on List::MoreUtils.  Thanks to Graham Ollis
    (plicease) for bringing the issues with this module to my attention.

0.071		2020-03-28	T. R. Wyant
    Recognize wildcard Unicode names (Perl 5.31.10).

    Try to get correct line number in derived PPI. This is done by
    injecting "\n" as needed. The initial #line directive becomes "#line
    2", but is suppressed if I need to generate line 1.

    Improve normalization of content for ppi(). This involves the
    un-bracketing of things like ${foo}.

    Deprecate new() argument postderef. At this stage it is only
    documented as deprecated. In the first release after October 1 2020
    it will warn on the first use. Eventually it will be retracted, and
    postfix dereferences will always be recognized. This is the default
    behavior now.

    Add dump argument/option 'short' which, if true, causes leading
    'PPIx::Regexp::' to be removed from class names.

0.070		2020-02-27	T. R. Wyant
    Add index_locations option to PPIx::Regexp->new(). This defaults to
    true if the regexp is specified as a PPI::Element object. The
    locations are consistent with the containing PPI::Document.

    Add methods location(), column_number(), line_number(),
    logical_filename(), logical_line_number(), and
    visual_column_number() to PPIx::Regexp::Element. All return undef if
    the locations could not be determined.

    Add method statement() to PPIx::Regexp::Element. This returns the
    PPI statement containing the regexp element, or nothing if none.

    Add method is_matcher() to PPIx::Regexp::Element. This classifies
    objects as to whether they actually match something in the target
    string. Possible returns are true (they do), false but defined (they
    do not) and undef (no clue).

    Add methods first_token() and last_token() to PPIx::Regexp::Node.

    Add methods next_token() and previous_token() to
    PPIx::Regexp::Element.

0.069		2020-02-07	T. R. Wyant
    The PPIx::Regexp->new() 'parse' option is now fatal. This selected
    either string or regex parse. I consider the string parse a failed
    experiment. This is the latest step in removing it in favor of the
    PPIx::QuoteLike package.

0.068		2020-01-21	T. R. Wyant
    Expose PPIx::Regexp::Util::is_ppi_regexp_element()

    explain() on [[=x=]] now calls it a Character Equivalence.
    It's still a PPIx::Regexp::Token::CharClass::POSIX::Unknown (and
    therefore an error), though.

0.067		2019-08-30	T. R. Wyant
    \K was retracted in Perl 5.31.3, but only inside look-around
    assertions.

0.066		2019-08-16	T. R. Wyant
    Fix broken POD, and add tests to ensure it remains fixed.

0.065		2019-05-25	T. R. Wyant
    Quash undef error in __is_ppi_regexp_element() when passed a
    PPI::Token::Regexp::Transliterate

    Support proper version for qr'\N{name}'. Until 5.29.10 this
    construction failed to parse because it did not interpolate. But
    PPIx::Regexp blithely ignored this detail. As of 5.29.10, something
    like m'\N{LATIN CAPITAL LETTER L}' matches identically to m'L'. So I
    implemented introduction as of that version.

    Have explain() recognize Unicode property wildcards.

0.064		2019-04-01	T. R. Wyant
    Empty \p{} should be an error.
    
    \x{} and \x{ non-hex } should be errors under "use re 'strict'"

    \o{} should be an error

    \o{ non-octal } should be an error under "use re 'strict'"

    Support wildcard Unicode property values. These were added in
    5.29.9.

    Add eg/find-variable-length-lookarounds

    Add convenience method extract_regexps(). This is a static method
    on PPIx::Regexp that takes as its argument a PPI::Document and
    manufactures PPIx::Regexp objects out of anything that parses to a
    regexp of some sort.

    Don't run illegal character tests before Perl 5.18 unless we're
    author testing, because they are noisy. I think the issue is not the
    Perl version per se, but the version of Unicode; Perl5180delta says
    it shipped with Unicode 6.2.

0.063		2018-11-08	T. R. Wyant
    Silence weird-character parse tests and make them no longer
    author-only.

    Further deprecate 'parse' argument to new(). You now get a warning
    on each use.

0.062		2018-08-12	T. R. Wyant
    Remove tokenizer method prior(). This is the last step in its
    deprecation.

0.061		2018-07-09	T. R. Wyant
    Only standalone graphemes and non-characters allowed as delimiters
    starting with Perl 5.29.0.
    
    Non-ASCII delimiters started working in 5.8.3, so that is what
    perl_version_introduced() returns for them.

    Collateral with all this, accept word characters as delimiters, but
    only with at least one space between the operator and the expression
    -- that is, 'qr xyx' is OK, but 'qrxyx' is not.

0.060		2018-06-16	T. R. Wyant
    \N{} now parses as the unknown token, not NoOp, regardless of the
    setting of 'use re qw< strict >;'. \N{} became unconditionally fatal
    in 5.28.0 (5.27.1, actually). The policy when the parse changes is
    to use the most-modern parse. Hence this change.

    As a side effect of this, the unknown token's explain() method now
    returns something -- normally the associated error.

    Add method remove_insignificant(). If the invocant isa Node, this
    returns a clone of the invocant with non-significant elements
    removed. Otherwise it returns either the invocant or nothing.

0.059		2018-05-08	T. R. Wyant
    Install @CARP_NOT everywhere so that warnings and exceptions
    generated in the bowels of the system appear to come from the point
    where the system is entered.

    Further deprecate string (versus regexp) parsing. The first use of
    the 'parse' argument to new() will result in a warning.  If the
    value of the argument is 'guess' or 'string', the warning refers to
    PPIx::QuoteLike.

0.058		2018-04-26	T. R. Wyant
    Prefer /[0-9]/ over /\d/ for numeric checks. The latter can match
    non-ASCII digits.

    Explain the negated POSIX character classes. Also tweak some of the
    asserted explanations -- mostly for readability and parallel
    construction with the negated explanations, but it turns out
    [[:digit:]] is NOT equivalent to [0-9].

0.057		2018-04-17	T. R. Wyant
    Allow ->asserts( 'a*' ). This modification actually allows wild
    cards in asserts() on all match semantic modifiers, but it is
    probably only useful in the case of 'a*', because that is the only
    one that can be doubled.

    Explain grouping structure as 'Grouping', not 'Capture or grouping'.

    Caret modifier was not turning off /n. This was complicated by the
    fact that (?^) was introduced in 5.13.6, but (?n) was not introduced
    until 5.21.8. The solution was to include -n in the expansion of the
    caret if and only if /n had been seen in the scope of the caret.

    Recognize caret in /(?^)x/.

    Acknowledge Regexp::Parsertron in SEE ALSO

0.056		2018-03-07	T. R. Wyant
    Support removal of unescaped literal left curlys after left parens,
    which was deprecated in 5.27.8. No actual change in output yet,
    since deprecation is not tracked, but the perl_version_removed()
    logic is there.


    Add next_element() and kin. These are analogous to next_sibling()
    and kin, but will cross over from content proper into structure
    (beginning and end delimiters, etc) and vice versa.

    Correct requirements_for_perl() for impossible regular expression.
    It now returns '! $]' when the components of the regexp are valid, but
    none are valid under any specific version of Perl. It used to think all
    Perls were OK when this happened.

    Add the alpha_assertions introduced in 5.27.9.

    Handle 5.27.9's change from +script_run to *script_run, and support
    *sr as a synonym.

0.055		2018-02-08	T. R. Wyant
    Tokenizer method prior() is now fatal. This was documented as
    package-private, but as it WAS documented, I am putting it through a
    deprecation cycle anyway. Six months from now it will be removed.

    Add Script_Run classes as subclasses to their superclass docs. This
    was missed in the last update.

0.054		2018-01-29	T. R. Wyant
    Add support for (+script_run:...). This is an experimental feature
    added in Perl 5.27.8. It imposes on any matches it contains the
    additional restriction that everything matched has to belong to the
    same Unicode script. This support will be retracted if the
    functionality does not make it into Perl 5.28.

    Add method scontent().  This returns significant content only. That
    is, if called on the parse of '/ f u b a r /x', it returns
    '/fubar/x'.

0.053		2017-10-30	T. R. Wyant
    Recognize \px as Unicode char class. At least, when the x is C, L,
    M, N, P, S or Z.

    The 'parse' argument to new() is now deprecated.

0.052		2017-09-07	T. R. Wyant
    RT 122715: Clarify Node->find_parents() documentation. Thanks to
    Salvatore Bonaccorso for letting me know about this problem.

    Further deprecate tokenizer method prior() in favor of
    prior_significant_token().


    Add requirements_for_perl(). This is analogous to the
    CPAN::Meta::Requirements method requirements_for_module(), though
    the output is formatted differently. Also put in the actual
    requirements for an un-escaped literal left curly after a constant,
    which was removed in 5.25.1 and reinstated in 5.27.1.

    Add accepts_perl(). This is analogous to
    CPAN::Meta::Requirements->accepts_module(). I decided that
    CPAN::Meta;:Requirements was overkill, but this may turn out to be
    the wrong decision, so I will be careful what I expose.

    Document behavior of perl_version_introduced() and
    perl_version_removed() when a feature is re-introduced after
    removal, or re-removed after re-introduction.

    \N{} (empty curlys) removed in 5.27.1.

0.051		2017-01-29	T. R. Wyant
  Support whitespace inside [] if /xx in effect.
    Starting with Perl 5.25.9, a space or tab appearing inside a bracketed
    character class is not significant if /xx is asserted.

  Further deprecate tokenizer method prior()

  Add 'provides' data to ExtUtils::MakeMaker output

  SOME unescaped litaral '{' removed in 5.025001.
    After '.', Unicode classes, and bracketed classes (including extended)
    they are still legal.

  Make /\b{/ an error
    Perl fails to parse the above, because once it sees the '\b{' it wants
    to find one of the extended boundary assertions (like \b{wb}), and
    declares an error when it does not. So we check for this and rebless the
    curly into an unknown token, not a literal.

0.050		2016-05-06	T. R. Wyant
  Parse bracketed substitution with embedded comment. This is something
    like s{foo}\n#{bar}\n{baz} which is equivalent to s/foo/baz/. PPI
    gets this wrong, and we're not smart enough to fix up the PPI parse,
    but if given this as text, we now parse it correctly.

  We now recognize postfix dereferences by default, since Perl does
    beginning with 5.24. In other words, default new() argument
    'postderef' to true.

  Unterminated substitutions (i.e. 's//') should no longer cause an
    exception. Instead they parse as an unknown token.

0.049		2016-04-19	T. R. Wyant
  Robustify PPIx::Regexp->perl_version_removed()
    The problem here was that if the expression being parsed was
    sufficiently badly-formed, $self->delimiters() would be undef, throwing
    a warning.

  Correct dump of embedded modifiers (eg: (?i:...))

0.048		2016-02-29	T. R. Wyant
  Add option 'strict', like 'use re "strict"'
    In the presence of strict(), I opted to set perl_version_introduced
    to the version of Perl where the construct became an error.

  Parse '\N{}' as no-op.
    The previous parse was a character class ('\N') followed by two
    literals ('{' and '}'). But perl5238delta said that it had been
    ignored up to that time. Starting with 5.23.8 it is an error if 'use
    re strict' is in effect.

  Quash 'NOT a POSIX class ...' warning under 5.23.8

  Add Makefile targets authortest and testcover.

0.047		2016-01-29	T. R. Wyant
  Recognize \b{lb}, introduced in 5.23.7. If this is retracted before
    5.24, it will be removed outright.

0.046		2016-01-08	T. R. Wyant
  Add GitHub repository to mmetadata.

0.045		2015-12-31	T. R. Wyant
  Deprecate tokenizer method prior() in favor of
    prior_significant_token(). This is not part of the public interface,
    so I suppose I could have just slam-dunked it, but ...

  Add ability to parse strings as well as regexes
    The new functionality is controlled by the new new() argument
    'parse', whose permitted values are 'regex' (the default), 'string',
    or 'guess'.  String parsing, and the 'string' and 'guess' values of
    'parse', are experimental.

0.044		2015-12-08	T. R. Wyant
  Allow nesting of \Q with \U, \L, and \F
    The perlop docs say these nest with each other. Playing with Perl
    suggests that \U, \L and \F supersede each other, but thet they as a
    group nest with \Q in either order, so that if you specify \Q and
    one of the \U, \L, \F group you need two \Es to turn them all back
    off.

  Restrict recognition of back references in replacement strings to
    \number form, since Perl itself does not recognize \g{...} or
    \k{...} there.

  Recognize postfix dereference if desired. This is controlled by the
    Boolean argument 'postderef' passed to PPIx::Regexp->new(). The
    default is false, but will become true if postfix dereference
    becomes mainstream Perl 5.

  Add explain() and supporting methods main_structure() and
    in_regex_set(). The explain() method returns a brief explanation of
    what the element does.

0.043		2015-11-18	T. R. Wyant
  Do not end regex set prematurely on finding '])'
    The problem is that '])' can occur within an extended bracketed
    character class if it contains grouping parentheses and the last
    item in a group is a regular bracketed character class and there is
    no white space between the end of the character class and the end of
    the group.

  Record parse failure if switch condition is unknown
    The structure was being reblessed to
    PPIx::Regexp::Structure::Unknown, but the number of parse failures
    was not being incremented.

  Parse \U and friends as meta-characters inside \Q...\E
    This turns out to be what Perl itself does, as shown by
    $ perl -E 'say qr{\Q\Ufoo}'

  Clear error when lexer identifies unknown token. Those who peruse the
    changes in this release will see that a bunch of refactoring was
    done as part of this.

  Parse white space inside bracketed character classes inside extended
    bracketed character classes (whew!) as literals, except for the
    space character itself and the horizontal tab. This tracks the
    corresponding change in Perl 5.23.4. This will be reverted if the
    corresponding Perl change does not make it into 5.24.0.

  Beginning with version 0.035, PPIx::Regexp was incorrectly reporting
    the sense of modifiers when the same token both asserted and negated
    modifiers (e.g. '(?x-i:...)'). This release should correct the
    problem.

  Document policy when Perl changes in such a way that the proper parse
    for a regular expression changes. In this case the more modern parse
    is preferred.

0.042		2015-10-09	T. R. Wyant
  Report error rather than failing when parsing a string consisting
    wholly of white space.

  Group types were not being recognized if they contained the delimiter
    character for the regexp (e.g. in qr<(?\<foo)> the look-behind
    assertion was not recognized as such).

  Correct mis-parse of ' s///'. Leading white space is supposed to be
    acceptable, but the leading whitespace token caused
    PPIx::Regexp::Lexer not to recognize the substitution as such.

  Tokenizer was failing when the string to be parsed was so bad it was
    trying to return the whole thing as a single
    PPIx::Regexp::Token::Unknown.

  PPIx::Regexp::Dumper now displays a message if a structure is missing
    its end delimiter.

  RT 107331 Produce parse error in the presence of trailing cruft.
    Thanks to Klaus Rindfrey for catching this.
    The tokenizer now does a preliminary scan for delimiting brackets
    and modifiers. Anything after the modifiers except for white space
    is now made into a PPIx::Regexp::Token::Unknown, resulting in a
    parse failure being reported. The previous implementation simply
    assumed a valid expression, and in the case of the expression in the
    ticket blithely mismatched the delimiters and returned a parse
    without failures, but which was manifestly bogus.

  Tweak documentation in PPIx::Regexp.

0.041		2015-07-02	T. R. Wyant
  Report \C (match octet) as removed in 5.23.0.

  Accept non-ASCII whitespace under /x. The Whitespace object can be
    multiple characters; the perl_version_introduced() becomes
    '5.021001' if any of them is a code point above 127.

  The perl_version_removed() method now returns '5.021001' when called
    on a PPIx::Regexp object produced by parsing '?foo?' (match once
    without explicit 'm'). The object produced by parsing 'm?foo?' still
    returns the minimum Perl version.

0.040		2015-05-31	T. R. Wyant
  Do not parse unadorned parentheses as capture groups when /n is in
    effect. Instead, they are parsed as PPIx::Regexp::Structure. Named
    captures appear to be unaffected by /n.

  Made a verbose dump a little more so. Specifically, dump
    max_capture_group where relevant, and display dumped values a bit
    more informatively.

  Report /n (no captures) as having been added in 5.21.8.

0.039		2015-04-02	T. R. Wyant
  Recognize nested subscripts in interpolation.
    Thanks to Andy Lester for finding this, which actually manifested in
    Perl-Critic-Policy-Variables-ProhibitUnusedVarsStricter. The problem is
    that the actual heuristics for finding the end of an interpolation are
    undocumented, and I missed this rather-obvious case.

  Add \b{g} (= \b{gcb})

0.038		2015-03-09	T. R. Wyant
  Make \b{foo} into an unknown token (and therefore an error. This
    applies to \b{anything}, where 'anything' is anything bur 'gcb',
    'wb', or 'sb'.

  Handle the boundary assertions introduced in Perl 5.21.9: '\b{gcb}'
    (grapheme cluster boundary), '\b{wb}' (word boundary), '\b{sb}'
    (sentence boundary), and the corresponding '\B{...}' constructions.
    Similar-looking things like '\b{foo}' are not recognized as
    assertions, and end up being literals.  This is less general than I
    usually make things, but was done against the possibility that
    (e.g.) '\b{foo}' might be introduced later, requiring
    perl_version_released() to return a different number. Any of these
    retracted prior to Perl 5.22.0 will simply be removed from
    PPIx::Regexp.

0.037		2014-11-12	T. R. Wyant
  Have PPIx::Regexp::Structure::RegexSet POD recognize that the Perl
    docs (specifically perlrecharclass) now call this construction
    Extended Bracketed Character Classes, not sets.

  Correctly mark the replacement portion of s///ee as code. Prior to
    this release it was parsed as though no /e were present.

  Make available the number of times a given modifier is asserted
    (except for the match semantics modifiers which get handled
    differently). See PPIx::Regexp::Token::Modifier->asserted() and
    PPIx::Regexp::Tokenizer->modifier() for details.

0.036		2014-01-04	T. R. Wyant
  Retract the "Allow non-ASCII white space under /x" change introduced
    in version 0.033. I misread perl5170delta, and implemented early.

  Change to explicit character class to recognize white space under /x.
    I was previously using \s, which matched too much. Thanks to Nobuo
    Kumagai for finding and reporting this.

0.035		2013-11-15	T. R. Wyant
  Properly handle multi-character modifiers like /ee. We now handle /eie
    as being the same as /eei. Thanks to Anonymous Monk for finding
    this.

  Properly handle \g and \k back references that do not correspond to an
    actual capture group. They are now reblessed into the unknown token,
    and counted as errors. Thanks to Anonymous Monk for finding this.

  Add method error() to PPIx::Regexp::Element. This should return an
    error message when the element is in error -- normally when it has
    been blessed into the unknown token or structure.

  Add method modifier_asserted() to PPIx::Regexp::Element. This walks
    the parse tree backward to determine if the given modifier is in
    effect for the element.

0.034		2013-05-11	T. R. Wyant
  Correct spelling and grammar errors in POD and comments. RT #85050.
    Thanks David Steinbrunner for catching these.

  Allow interpolation in regex sets. It implies Perl 5.17.9 or higher.

  Allow non-ASCII white space under /x. It implies Perl 5.17.9 or
    higher.

  Fix problems with Regex Set functionality under Perl 5.6.2. CPAN
    testers RULE!

0.031		2013-01-31	T. R. Wyant
  Have PPIx::Regexp::Token::Code (and offspring) become
    PPIx::Regexp::Token::Unknown inside a regex set.

0.030		2013-01-22	T. R. Wyant
  Add Regex Sets, which were added to Perl as an experimental feature in
    5.17.8. This is experimental in Perl, therefore the parse may
    change.

  Ditch PPIx::Regexp::Token::GroupType method __expect_after_match() in
    favor of the more general __match_setup(). This is done without
    deprecation because __expect_after_match() was documeted as
    package-private, but noted in the change log because it _was_
    documented.

0.029		2013-01-14	T. R. Wyant
  Add method unescaped_content() to PPIx::Regexp::Element().

  Rewrite the tokenizing code in PPIx::Regexp::Token::GroupType and
    offspring to use regular expressions specific to the regexp
    delimiter, and escaping only that delimiter. Thanks again to
    Alexandr Ciornii for finding more of these.

  Fix mis-parse of /(\?|I)/ as a branch reset (it's really an
    alternation). There may be more of these lurking. Thanks to Alexandr
    Ciornii for finding this one.

  Add options -files and -objectify to eg/predump.

0.028		2012-06-06	T. R. Wyant
  Replace all uses of YAML::Any with YAML, since they come in the same
    distro, and YAML does not suffer from deprecation warnings.

0.027		2012-05-28	T. R. Wyant
  Eliminate unescaped literal "{" characters in regexps in
    PPIx::Regexp::Token::Backreference and
    PPIx::Regexp::Token::CharClass::Simple. These are deprecated in 5.17.0.

0.026		2012-02-24	T. R. Wyant
  Add support for \F (fold case), added in 5.15.8.

0.025		2012-01-04	T. R. Wyant
  Tolerate leading and trailing white space around the regular
    expression.  These are still round-trip safe, since the white space
    is tokenized.

  Make Changes file conform to CPAN::Changes, and add
    xt/author/changes.t to ensure continued compliance.

0.024		2011-12-17	T. R. Wyant
  Reinstate author test xt/author/manifest.t, which was clobbered
    shortly before the release of 0.021_10.

0.023		2011-12-08	T. R. Wyant
  Correct address of FSF in the version of the GPL distributed in
    LICENSES/Copying. Thanks to Petr Pisar for picking this up.

0.022		2011-11-24	T. R. Wyant
  Correct various documentation errors.

  Don't initialize effective modifiers with '^', since that wrongly
    asserts that /d has been seen somewhere along the line.

  Implement negation of match-semantic modifiers (e.g. 'no re /u;') by
    setting the relevant datum to undef.

  Support for default modifiers. This includes:

    * default_modifiers argument to new() in PPIx::Regexp,
      PPIx::Regexp::Tokenizer, and PPIx::Regexp::Dumper

    * Public method modifier_asserted() on PPIx::Regexp, to return
      whether a given modifier is actually in effect. The results of the
      modifier() method are unchanged.

  Require Test::More 0.88 for installation. Eliminate all the 'eval
    { require ... }' logic in favor of 'use Test::More 0.88'.

  Have Makefile.PL make use of {BUILD_REQUIRES} if it is available.

  Fix PPIx::Regexp::Token::Whitespace->can_be_quantified() to return
    false.

0.021		2011-07-22	T. R. Wyant
  Modified tokenizer to correctly handle a back slash used as a
    delimiter. I believe.

  PPIx::Regexp::Dumper now dumps the results of ppi() if that method is
    present and -verbose is asserted.

0.020		2011-04-02	T. R. Wyant
  Corrected perl_version_introduced():

  * \R is now 5.009005 (was 5.000).

0.019		2011-03-01	T. R. Wyant
  Various corrections to perl_version_introduced():

  * \X is now 5.006 (was 5.000);

  * \N{name} is now 5.006001 (was 5.006);

  * \N{U+xxxx} is now 5.008 (was 5.006).

  The \C is now parsed as a PPIx::Regexp::Token::CharClass::Simple. It
    was previously considered a PPIx::Regexp::Token::Literal.

  Ensure that \N{$foo} parses as a Unicode literal, not a quantified \N.
    The ordinal() method returns undef for this.

  Understand the /aa modifier, introduced with 5.13.10.

  Report perl_version_introduced() of 5.013010 for the new semantic
    modifiers when modifying the entire expression.

  Correct handling of interpolations like ${^foo} and $#{foo}.

0.018		2011-02-16	T. R. Wyant
  Override ppi() in PPIx::Regexp::Token::Interpolation to provide the
    proper PPI when variable names are bracketed.

  Properly parse bracketed variable names (I hope!), which may not be
    subscripted.

  Take account of possible '$' or '@' casts before a symbol in an
    interpolation (e.g. $$foo{bar}, which is equivalent to $foo->{bar}).

  Add the /a modifier to PPI::Regexp::Token::Modifiers, legal only in
    the (?:...) construction. This was introduced in Perl 5.13.9.

  When parsing an interpolation from a replacement string (rather than a
    regular expression), take subscripts at face value rather than
    trying to disambiguate them from quantifiers and character classes,
    which they can't be in this context.

0.016		2011-01-05	T. R. Wyant
  The PPIx::Regexp::Token::Code perl_version_introduced() method now
    returns the minimum Perl version (currently set to 5.000) if it is
    used to represent the substitution portion of s///e.

0.015		2010-10-25	T. R. Wyant
  Documented intent to revoke support for features introduced in a
    development Perl which do not make it to a production release. This
    is necessary because in this case the syntax could be reused with
    different semantics.

  Added support for Perl 5.13.6 (?^...) construction.

  Added support for Perl 5.13.6 d, l, and u modifiers.

  Fixed inconsistency in perl_version_introduced() results between
    PPIx::Regexp::Token::Modifier and
    PPIx::Regexp::Token::GroupType::Modifier.

  Corrected PPIx::Regexp::Constant RE_CAPTURE_NAME docs, somehow missed
    back at 0.010_01.

0.014		2010-10-14	T. R. Wyant
  Recognize \o{...} as a PPIx::Regexp::Token::Literal, with
    perl_version_introduced() of 5.0013003.

  Terminate \0.. through \7.. after three characters, as Perl does.
    These two were brought to my attention by Brian D. Foy's "The
    Effective Perler" for October 11 2010,
    http://www.effectiveperlprogramming.com/blog/697

  Correct the PPIx::Regexp::Token::Literal ordinal() method for '\b'. As
    a literal, this is a back space.

0.013		2010-10-10	T. R. Wyant
  Declare a parse failure if characters are found between the '}' and
    the ')' of (?{...}) and (??{...}), and rebless the tokens to
    ::Unknown. Perl does not accept anything here, so I think I should
    not either.

  Whitespace tweak in the PPIx::Regexp::Dumper test output for the
    failures test.

  Replace the PPI logic in PPIx::Regexp::Token::Code with a call to
    $tokenizer->find_matching_delimiter(). This is actually the way Perl
    works, as a look at toke.c and regcomp.c makes clear.

  Push the perl_version_introduced() back to 5.0 at the request of
    Alexandr Ciornii, for the potential benefit of Perl::MinimumVersion.
    This was done mostly by reading the various perlre, perldelta, and
    perlop documents, so these should be taken with a HUGE grain of
    salt.

0.012		2010-09-26	T. R. Wyant
  Track all the features reported as introduced (or removed) in Perl
    5.010 back to Perl 5.009005, and report them as such.

  Report modifier /r as having been introduced in Perl 5.013002, rather
    than the default of 5.006.

0.011		2010-09-16	T. R. Wyant
  Remove dependencies on Params::Util and Readonly. The latter was
    requested by ADAMK for the benefit of Padre. It involved changing
    the symbols exported from PPIx::Regexp::Constant, but these were
    documented as private, so ...

  Parse POSIX character classes [=a=] and [.a.] as
    PPIx::Regexp::Token::CharClass::POSIX::Unknown, which counts as a
    parse failure since these are not supported by Perl.

  Make the PPI::Document created by PPIx::Regexp::Token::Code->ppi() be
    read only. This means we need PPI 1.116. Cache the document, and
    ensure the cached result is returned on subsequent calls.

0.010		2010-08-06	T. R. Wyant
  Fix fatal error in PPIx::Regexp::Token::Code->ppi().

  Move author tests from xt/ to xt/author/.

0.009		2010-08-03	T. R. Wyant
  Recognize s/.../.../ee as being different from s/.../.../e. In
    particular, the replacement portion of the former is _not_ a Perl
    expression: it's an interpolatble string, which later gets
    eval{}'ed.

0.008		2010-07-01	T. R. Wyant
  Promote methods can_be_quantified() and is_quantifier() from
    PPIx::Regexp::Token to PPIx::Regexp::Element, so all classes inherit
    them. They still return true and false respectively.

  Override can_be_quantified() to return false on PPIx::Regexp,
    PPIx::Regexp::Structure::Quantifier,
    PPIx::Regexp::Structure::Regexp, and
    PPIx::Regexp::Structure::Replacement.

  Override is_quantifier() to return true on
    PPIx::Regexp::Structure::Quantifier.

  Modify PPIx::Regexp::Dumper to be able to display can_be_quantified
    and is_quantifier for PPIx::Regexp::Node objects when dumping
    verbosely.

  Convert internal data to Readonly in PPIx::Regexp::Lexer,
    PPIx::Regexp::Token::CharClass::Simple,
    PPIx::Regexp::Token::Structure, and PPIx::Regexp::Tokenizer.

  Remove leftover boilerplate in PPIx::Regexp::Token::CharClass::Simple.

  Explicitly require a minimum Perl of 5.006.

  Centralized dependencies in inc/PPIx/Regexp/Meta.pm.

  Removed claim that PPIx::Regexp is alpha code. Docs still say that the
    interface can be changed, but now it will go through a deprecation
    cycle.

0.007		2010-04-28	T. R. Wyant
  PPIx::Regexp::Lexer no longer fails when encountering expressions like
    m{)}. Instead, it marks the right parenthesis as an unmatched
    delimiter.

  Fixed RT 56864 - PPIx::Regexp::Lexer fails in Perl::Critic under Perl
    5.13.0. This was due to the value of a returned $+[0] getting
    transmogrified before the caller saw it. I never did isolate what
    triggered the bug.

  You can now get a tokenizer trace by setting environment value
    PPIX_REGEXP_TOKENIZER_TRACE to a non-zero numeric value. This is
    unsupported, though.

0.006		2010-02-26	T. R. Wyant
  Parse \N{...} in accordance with perl5115delta. The curlys must
    contain an alpha followed by alphanumerics, spaces, parens, colons,
    or dashes. \N{ without a matching } is a character class (if legal)
    followed by a literal '{'.

  Parse \N inside a character class as PPI::Regexp::Token::Unknown,
    since Perl 5.11.5 considers this a compile error. A \N{...} inside a
    character class is still OK.

  Add method match() to PPIx::Regexp::Tokenizer. This is analogous to
    capture(), but returns the entire matched string.

0.005		2009-12-26	T. R. Wyant
  Recognize \N (without curlys), back-ported from Perl 6 into 5.11.

  Recognize unicode characters as \N{[[:alpha:]] ... rather than
    \N{[\w\s:] ... This is per the 5.11 documentation, but I think Perl
    always worked this way.

  Recognize loose matching of Unicode character classes, and allow '='
    in lieu of a single ':' in a Unicode character class (this from Perl
    5.11.3).

  PPIx::Regexp::Dumper now produces the proper output when called with
    perl_version => 1, test => 1.

  Describe the typical content of the object in the documentation for
    PPIx::Regexp::Structure::NamedCapture and
    PPIx::Regexp::Token::GroupType::NamedCapture.

0.004		2009-11-09	T. R. Wyant
  Have PPIx::Regexp::Token::Literal correctly recognize when
    charnames::vianame() is unavailable, and decouple this from the
    handling of \N{U+hhhh}.

  Add dependency on Task::Weaken, since depending on Scalar::Util
    appears not to cut it.

  Correct the assignment of the license type in Makefile.PL.

0.003		2009-11-05	T. R. Wyant
  Have PPIx::Regexp::Token::Literal recognize \N{U+hhhh} (where hhhh
    represents hex digits), and provide its ordinal (hhhh). Remove
    recognition of \N. (. = any character), which Perl does not do.

  Fix $re->flush_cache() so that it actually removes $re and only
    $re from the cache.

  Add delimiters() method to PPIx::Regexp::Main and PPIx::Regexp.
    Support this in eg/prenav.

  Increase test coverage and remove dead code.

  Count tests in t/parse.t and t/unit.t

0.002		2009-10-28	T. R. Wyant
  In verbose mode, have PPIx::Regexp::Dumper dump the absolute capture
    number referred to by a numbered reference.

  Have eg/preslurp pass its -verbose option to PPIx::Regexp::Dumper

  Don't use Test::More::isa_ok for the t/basic.t class heritage tests,
    since some versions of Test::More require a reference for the first
    argument of isa_ok().

0.001		2009-10-21	T. R. Wyant
  Initial release.

# ex: set textwidth=72 autoindent :