Log message:
devel/ruby-regexp_parser: update to 2.10.0

2.10.0 (2024-12-25)

Added

* #referenced_expressions

    - like #referenced_expression, but for multiplexing backrefs
    - returns the Group expressions that are being referenced

Fixed

* fixed #char & #codepoint errors for single-digit hex escapes

    - e.g. \xA

Log message:
devel/ruby-regexp_parser: update to 2.9.3

2.9.3 (2024-11-29)

Fixed

* fixed positive lookbehinds with character ">" being treated as named
  groups

    - e.g. (?<=foo>)
    - thanks to Daniel Vandersluis

Log message:
devel/ruby-regexp_parser: update to 2.9.2

2.9.2 (2024-05-15)

Fixed

* made the MFA requirement for changes to this gem visible on rubygems
  thanks to Geremia Taglialatela

Log message:
devel/ruby-regexp_parser: update to 2.9.1

2.9.1 (2024-05-11)

Fixed

* fixed unnecessary $LOAD_PATH searches at load time
  thanks to Koichi ITO

Log message:
devel/ruby-regexp_parser: update to 2.9.0

2.9.0 (2024-01-07)

Added

* all expressions now respond to #negative? / #negated?

    - previously only sets, props, and posix classes did

* implemented #negative? / #negated? for more applicable expressions

    - \B, \D, \H, \S, \W, (?!...), (?<!...)

Fixed

* fixed missing support for grapheme cluster break unicode properties

    - e.g. /\p{Grapheme_Cluster_Break=Extend}/

Log message:
devel/ruby-regexp_parser: update to 2.8.3

2.8.3 (2023-12-04)

Fixed

* fixed scanner errors for insignificant leading zeros in numerical group
  refs
  - e.g. (a)\k<01>, (a)\g<-01>, (a)?(?(01)b|c)
  - thanks to Markus Schirp for the report

Log message:
devel/ruby-regexp_parser: update to 2.8.2

2.8.2 (2023-10-10)

Fixed

* handle a corner case where parsing redundant number escapes raised an error

   - e.g. parse(/\99/), which in Ruby is a valid Regexp that matches 99
   - thanks to Markus Schirp for the report

Log message:
devel/ruby-regexp_parser: update to 2.8.1

2.8.1] (2023-06-10)

Fixed

* support for extpict unicode property, added in Ruby 2.6
* support for 10 unicode script/block properties added in Ruby 3.2

Log message:
devel/ruby-regexp_parser: update to 2.8.0

2.8.0 (2023-04-17)

Added

o Regexp::Expression::Shared#ends_at
  * e.g. parse(/a +/x)[0].ends_at # => 3
  * e.g. parse(/a +/x)[0].ends_at(include_quantifier = false) # => 1
o Regexp::Expression::Shared#{capturing?,comment?}
  * previously only available on capturing and comment groups
o Regexp::Expression::Shared#{decorative?}
  * true for decorations: comment groups as well as comments and whitespace
    in x-mode
o Regexp::Expression::Shared#parent
o new format argument :original for Regexp::Expression::Base#to_s
  * includes decorative elements between node and its quantifier
  * e.g. parse(/a (?#comment) +/x)[0].to_s(:original) # => "a \ 
(?#comment) +"
  * using it is not needed when calling Root#to_s as Root can't be
    quantified

o support calling Subexpression#{each_expression,flat_map} with a
  one-argument block
  * in this case, only the expressions are passed to the block, no indices
o support calling test methods at Expression class level
  * capturing?, comment?, decorative?, referential?, terminal?
  * e.g. Regexp::Expression::CharacterSet.terminal? # => false

Fixed

o Regexp::Expression::Shared#full_length with whitespace before quantifier
  * e.g. parse(/a +/x)[0].full_length used to yield 2, now it yields 3
o Subexpression#to_s output with children with whitespace before their
  quantifier
  * e.g. parse(/a + /x).to_s used to yield "a+  ", now it yields \ 
"a + "
  * calling #to_s on sub-nodes still omits such decorative interludes by
    default
      - use new #to_s format :original to include it
      - e.g. parse(/a + /x)[0].to_s(:original) # => "a +"
o fixed Subexpression#te behaving differently from other expressions
  * only Subexpression#te used to include the quantifier
  * now #te is the end index without quantifier, as for other expressions
o fixed NoMethodError when calling #starts_at or #ts on empty sequences
  * e.g. Regexp::Parser.parse(/|/)[0].starts_at
  * e.g. Regexp::Parser.parse(/[&&]/)[0][0].starts_at
o fixed nested comment groups breaking local x-options
  * e.g. in /(?x:(?#hello)) /, the x-option wrongly applied to the
    whitespace
o fixed nested comment groups breaking conditionals
  * e.g. in /(a)(?(1)b|c(?#hello)d)e/, the 2nd conditional branch included
    "e"
o fixed quantifiers after comment groups being mis-assigned to that group
  * e.g. in /a(?#foo){3}/ (matches 'aaa')
o fixed Scanner accepting two cases of invalid Regexp syntax
  * unmatched closing parentheses ()) and k-backrefs with number 0 (\k<0>)
  * these are a SyntaxError in Ruby, so could only be passed as a String
  * they now raise a Regexp::Scanner::ScannerError
o fixed some scanner errors not inheriting from
  Regexp::Scanner::ScannerError
o reduced verbosity of inspect / pretty print output

Log message:
devel/ruby-regexp_parser: update to 2.7.0

2.7.0 (2023-02-08)

Added

* Regexp::Lexer.lex now streams tokens when called with a block
  - it can now take arbitrarily large input, just like Regexp::Scanner
  - this also slightly improves Regexp::Parser.parse performance
  - note: Regexp::Parser.parse still does not and will not support streaming
* improved performance of Subexpression#each_expression
* minor improvements to Regexp::Scanner performance
* overall improvement of parse performance: about 10% for large Regexps

Fixed

* parsing of octal escape sequences in sets, e.g. [\141]
  - thanks to Randy Stauner for the report