Dec 2025 Meeting

calendar_today December 11, 2025 confirmation_number 12 tickets discussed Source English

DevMeeting-2025-12-11

https://bugs.ruby-lang.org/issues/21689

DateTime and location

  • 2025/12/11 (Thu) 13:00-17:00 JST @ Online

Next Date

  • 2026/01/14 (Wed) 13:00-17:00 JST @ Online

Announce

About release timeframe

Ordinary tickets

[Feature #21619] logger context API (chucke)

  • Blog post detailing motivation here
    • logger doesn’t have a context API to enrich payload for structured logging
    • other languages have some of the kind (the post details examples from java, go, python)
  • PR with proposal here
    • some public API feedback to address.
    • sonots (gem maintainer) didn’t reply since I opened the proposal.

Preliminary discussion:

logger = Logger.new(STDOUT)
logger.with_context("user_id" => 1) do
  logger.info("foo")
    #=> I, [2025-11-28T17:57:50.947152 #3705893] [user_id=1] INFO -- : foo
    # if using a hypothetical json formatter
    #=> { "message" => "foo", "user_id" => 1, ....
  end
end
logger.info("foo", context: { "user_id" => 1 }) #=> same thing, but single call only

Discussion:

  • tagomoris: should I take over logger gem?
  • mame: I believe no one will be against it

Conclusion:

  • tagomoris: I will contact on sonots (maybe in person).

[Feature #21704] Expose rb_process_status_new to C extensions. (ioquatix)

  • Required for improved io_uring process_wait hook (it returns full status).
  • https://man7.org/linux/man-pages/man3/io_uring_prep_waitid.3.html
  • Is it acceptable? https://github.com/ruby/ruby/pull/15213

Discussion:

  • mame: was ko1 against?
  • ko1: Not patricular against to this one but I guess it can lead to other requests? That if any can be handled as separate tickets so I’m okay.
  • n0kada: I don’t see practial info that people want to add here though.
  • akr: wait, 
      VALUE rb_process_status_new(rb_pid_t pid, int status, int error);

    Why do we need error here? Process::Status doesn’t have that bit.

  • mame: nobu made the API.
  • n0kada: me remembers nothing.
  • mame: It seems the function was a refactoring effort of old rb_last_status_set()
  • akr: I guess rb_last_status_set() doesn’t fit to Sam’s needs because the function is thread-local.

Conclusion:

  • mame: Rather than exposing this existing function, which isn’t well thought-out, it would be better to have a more fluent API.
  • akr: I will reply

[Feature #21701] Enumerator.produce accepts an optional size keyword argument (knu)

  • Enumerator#to_set now rejects infinite sequences, and the current situation of Enumerator::Producer that its size is fixed to infinity creates an inconsistency that prevents finite produce enumerators from being converted to sets.

Preliminary discussion:

# Infinite enumerator
enum = Enumerator.produce(1, size: Float::INFINITY, &:succ)
enum.size  # => Float::INFINITY

# Finite enumerator with known/computable size
abs_dir = File.expand_path("./baz") # => "/foo/bar/baz"
traverser = Enumerator.produce(abs_dir, size: -> { abs_dir.count("/") + 1 }) {
  raise StopIteration if it == "/"
  File.dirname(it)
}
traverser.size  # => 4

abs_dir = "/"
traverser = Enumerator.produce(abs_dir, size: -> { abs_dir.count("/") + 1 }) {
  raise StopIteration if it == "/"
  File.dirname(it)
}
p traverser.size #=> 2
p traverser.to_a #=> ["/"]

Discussion:

  • knu: produce can in fact be used to create an infinite enumerator, but after using it for a while I found that it is not always the case.
  • mame: I don’t like it when the provided size and actual number of iteration differs.
  • akr: produce’s size being infinity now seems like a wrong idea to me

(everybody nods)

  • akr: does “size being infinite” mean the enumerator does not stop, or can stop?
  • knu: size can be different from actual number of iteration. For instance, SELECT COUNT(*) returns a fixed number, then later SELECT * can accumulate different number of rows because of INSERT / DELETE in between the twp operations.
  • akr: Then I guess it was not a wise idea for #to_set to check that info…

Conclusion:

  • matz: Go ahead

[Feature #21552] allow String.strip and similar to take a parameter similar to String.delete (shugo)

  • pull request: https://github.com/ruby/ruby/pull/15400
  • Is it acceptable to take character selectors like String#delete?

Discussion:

  • shugo: I find it handy.
  • shugo: Also python has this feature.
  • matz: I’m not super against it. This breaks nothing.
  • mame: Not a big fan though.
  • ko1: It can be misused with String#delete_suffix and so on.
  • mame: "aabbccddee".count("a-c") #=> 6

Conclusion:

  • matz: Go ahead.
  • naruse: OK to ship it w/4.0.0.

[Bug #21049] Reconsider handling of the numbered parameters and “it” parameter in Binding#local_variables (mame)

  • An API like Binding#numbered_parameter_get was actually requested. I’d like to discuss how to proceed.

Discussion:

  • mame: local_variable_get already does not return it.
  • mame: a way to get it has been postponed, untl requested from someone.
  • mame: now RubyMine’s authors want to get that bit from their debugger.
  • matz: #it_get sounds esoteric to me…
Binding#local_variables #=> [:a, :b, :c]

Binding#implicit_parameters #=> [:_1, :_2, :_3, ...]
Binding#implicit_parameters #=> [:it]
Binding#implicit_parameter_get(:_1) #=> obj
Binding#implicit_parameter_get(:_9) #=> implicit parameter '_1' is not defined for #<Binding:...> (NameError)
Binding#implicit_parameter_get(:it) #=> implicit parameter 'it' is not defined for #<Binding:...> (NameError)
Binding#implicit_parameter_defined?(:_1) #=> true or false
Binding#implicit_parameter_defined?(:it) #=> true or false

Binding#implicit_parameter_set: do not introduce this

foo do
  _2
  binding.implicit_parameters #=> [:_1, :_2]
  binding.eval("_2") #=> undefined local variable or method '_2'
end

Conclusion:

  • matz: Feature accepted, naming need brush-ups.

[Feature #19979] Reconsider adding &nil to method declarations, to signal it won’t accept a block (pabloh)

  • Adding this feature should be possible now that the syntax moratorium is over

Discussion:

  • matz: Super honestly speaking it was a design mistake, that a method takes a block implicitly.
  • akr: what abot allowing implicit blocks only when yield is called?
  • matz: that “only when” is not obvious, for instance consider def foo(...) = super(...)
  • ko1: that’s warned though.
  • matz: given that def foo(**nil) = /.../ is already allowed, def foo(&nil) = /.../ can be a compromise to me.
  • knu: It would be meddlesome when rubocop starts warning every possible method definitions I have already wrote.

Conclusion:

  • matz: This would be a 4.1 killer feature. Accepted.

[Feature #8948] Frozen regex (eregon)

  • This is to make all Regexps frozen and not just literals.
  • I made a PR, it passes CI: https://github.com/ruby/ruby/pull/14547
  • Is it OK to merge it for Ruby 4.0?

Discussion:

  • shyouhei: I remember it was not merged before because of test breakage.
  • ko1: eregon also fixed those tests.

  • mame: Looking at the test changes, it can be problematic when previously Marshal.dump-ed (using 3.x) regular expression cannot be Marshal.load-ed in 4.x.

  • akr: Perhaps it is the first time we try freezing objects with “actual” identities, that is something different from Integer etc?
2008-04-12 /srv/gems/Kanocc-0.1.0/lib/kanocc/token.rb:  class Token < Regexp
2009-05-16 /srv/gems/animehunter-mongo-0.9/lib/mongo/types/regexp_of_holding.rb:      class RegexpOfHolding < Regexp
2016-07-21 /srv/gems/apipie-params-0.0.5/lib/apipie/params/descriptor.rb:      class Number < Regexp
2008-12-22 /srv/gems/bahuvrihi-tap-0.12.0/lib/tap/test/regexp_escape.rb:    class RegexpEscape < Regexp
2025-02-04 /srv/gems/blouson-4.0.0/lib/blouson/tolerant_regexp.rb:  class TolerantRegexp < Regexp
2013-05-04 /srv/gems/chomsky-0.1.1/lib/chomsky/parsers/blank.rb:    class Blank < Regexp
2018-12-07 /srv/gems/eleetscript-0.0.30/lib/lang/runtime/base_classes.rb:  class ESRegex < Regexp
2012-04-01 /srv/gems/hiwai-0.0.1/lib/hiwai/masked.rb:  class MaskedRegexp < Regexp
2022-04-27 /srv/gems/jekyll-bible_references-0.1.0/lib/jekyll/bible_references/scripture_pattern.rb:    class ScripturePattern < Regexp
2010-03-01 /srv/gems/kbaum-mongo-0.19/lib/mongo/types/regexp_of_holding.rb:  class RegexpOfHolding < Regexp
2010-08-31 /srv/gems/lgierth-rack-mount-0.6.13/lib/rack/mount/regexp_with_named_groups.rb:    class RegexpWithNamedGroups < Regexp
2010-01-28 /srv/gems/mongo-find_replace-0.18.3/lib/mongo/types/regexp_of_holding.rb:  class RegexpOfHolding < Regexp
2009-05-16 /srv/gems/mongodb-mongo-0.14.1/lib/mongo/types/regexp_of_holding.rb:  class RegexpOfHolding < Regexp
2024-10-12 /srv/gems/oktest-1.5.0/lib/oktest.rb:    class PartialRegexp < Regexp
2023-11-23 /srv/gems/opal-1.8.2/spec/opal/core/marshal/dump_spec.rb:class MarshalUserRegexp < Regexp
2009-05-16 /srv/gems/pahagon-mongo-abd-0.14.1/lib/mongo/types/regexp_of_holding.rb:  class RegexpOfHolding < Regexp
2012-07-14 /srv/gems/persistence-0.0.4/spec/persistence/object/flat/regexp_spec.rb:    class ::Persistence::Object::Flat::RegexpMock < Regexp
2015-03-05 /srv/gems/quandl_client-2.13.0/lib/quandl/pattern.rb:  class Pattern < Regexp
2011-08-30 /srv/gems/rack-mount-0.8.3/lib/rack/mount/regexp_with_named_groups.rb:    class RegexpWithNamedGroups < Regexp
2023-03-06 /srv/gems/resyma-0.1.2/lib/resyma/core/automaton/regexp.rb:    class RegexpConcat < Regexp
2023-03-06 /srv/gems/resyma-0.1.2/lib/resyma/core/automaton/regexp.rb:    class RegexpSelect < Regexp
2023-03-06 /srv/gems/resyma-0.1.2/lib/resyma/core/automaton/regexp.rb:    class RegexpRepeat < Regexp
2023-03-06 /srv/gems/resyma-0.1.2/lib/resyma/core/automaton/regexp.rb:    class RegexpSomething < Regexp
2023-03-06 /srv/gems/resyma-0.1.2/lib/resyma/core/automaton/regexp.rb:    class RegexpNothing < Regexp
2024-02-15 /srv/gems/rhodes-7.6.0/spec/framework_spec/app/spec/core/marshal/fixtures/marshal_data.rb:class UserRegexp < Regexp
2024-02-15 /srv/gems/rhodes-7.6.0/spec/framework_spec/app/spec/core/regexp/shared/new.rb:    class RegexpSpecsSubclass < Regexp
2017-09-09 /srv/gems/riel-1.2.1/lib/riel/regexp.rb:class NegatedRegexp < Regexp
2011-11-01 /srv/gems/shell_test-0.5.0/lib/shell_test/regexp_escape.rb:  class RegexpEscape < Regexp
2024-08-22 /srv/gems/sitediff-1.2.11/lib/sitediff/sanitize/regexp.rb:      class WithSelector < Regexp
2025-06-02 /srv/gems/strong_routes-2.0.1/lib/strong_routes/matcher.rb:  class Matcher < Regexp
2010-06-07 /srv/gems/tack-0.0.1/lib/tack/test_pattern.rb:  class TestPattern < Regexp
2011-01-09 /srv/gems/tap-test-0.7.0/lib/tap/test/shell_test/regexp_escape.rb:      class RegexpEscape < Regexp
2016-05-23 /srv/gems/tauplatform-1.0.3/spec/framework_spec/app/spec/core/marshal/fixtures/marshal_data.rb:class UserRegexp < Regexp
2016-05-23 /srv/gems/tauplatform-1.0.3/spec/framework_spec/app/spec/core/regexp/shared/new.rb:    class RegexpSpecsSubclass < Regexp
2010-05-21 /srv/gems/usher-0.8.3/lib/usher/route/static.rb:      class Greedy < Regexp
2013-07-26 /srv/gems/verbal-0.1.2/lib/verbal.rb:class Verbal < Regexp
2014-06-04 /srv/gems/verbal_expressions-0.1.5/lib/verbal_expressions.rb:class VerEx < Regexp
2024-11-11 /srv/gems/wallaby-core-0.3.2/lib/routes/wallaby/lazy_regexp.rb:  class LazyRegexp < Regexp
2018-06-17 /srv/gems/web-page-parser-1.2.2/lib/web-page-parser.rb:    class ORegexp < Regexp

module Mongo
  ...
  class RegexpOfHolding < Regexp

    attr_accessor :extra_options_str

    # +str+ and +options+ are the same as Regexp. +extra_options_str+
    # contains all the other flags that were in Mongo but we do not use or
    # understand.
    def initialize(str, options, extra_options_str)
      super(str, options)
      @extra_options_str = extra_options_str
    end
  end
end

# frozen_string_literal: true

module Wallaby
  # This is designed to delegate all the methods to {#lazy_regexp}
  # So that it doesn't need to load all the models for {Map.mode_map}
  # when the engine is mounted in `config/routes.rb`
  #
  # @see Map.resources_regexp
  # @see Map.id_regexp
  class LazyRegexp < Regexp
    delegate(*Regexp.instance_methods(false), to: :lazy_regexp)

    def initialize(source, **options)
      @lazy_source = source
      super(source.to_s, **options)
    end

    protected

    def lazy_regexp
      Map.try(@lazy_source)
    end
  end
end

Conclusion:

  • yui-knk: the Marshal glitch is a show-stopper to me.
  • akr: when a user class inherits Regexp, freezing it can be NG.

[Feature #21766] Pathname + FileUtils making sweet music together (eregon)

  • These methods are useful and in fact are expected by some to already exist on Pathname.
  • How should we add them given they depend on FileUtils?

Discussion:

  • akr: Feature wise pathname already have renaming etc. What is wanted here is just naming? That doesn’t amuse me.
  • hsbt: WDYT about require "fileutils" adds methods to Pathname?
  • akr: not in favour.

Conclusion:

  • akr: I don’t share your love for FileUtils.

[Misc #21769] Use “vX.Y.Z” instead of “vX_Y_Z” as tag names on ruby.git (k0kubun)

  • I want Ruby 4.0.0 to use v4.0.0 as its git tag name instead of v4_0_0. If we release rc1, can we call it v4.0.0-rc1 too?
  • Do we also want to drop v and make it 4.0.0 like @jeremyevans0 proposed?

Preliminary discussion:

  • v4_0_0_preview2
  • v4.1.1
  • v4.1.0
  • v4.0.2
  • v4.0.1
  • v4.0.0 <- selected
  • v4.0.0-preview3

or

  • v4_0_0_preview2
  • v3_5_10
  • 4.1.1
  • 4.1.0
  • 4.0.2
  • 4.0.1
  • 4.0.0

Discussion:

  • shyouhei: This is a CVS tradition.
  • n0kada: We just have had no needs to change previous tag scheme until now.
  • ko1: the motivation of transition?
  • k0kubun: I want to align version number and tag name.
  • n0kada: then v4.0.0 can be something you don’t want.
  • mame: tag 4.0.0 would appear at the very end of the list of tags, esp. in github web UI. This is NG to me.

Conclusion:

  • matz: v4.0.0 accepted.
  • hsbt: previous v4_0_0_preview2 should be renamed to: v4.0.0-preview2.

[Misc #21770] Stop bumping RUBY_PATCHLEVEL in release versions (k0kubun)

  • I want to update tool/merger.rb to skip updating RUBY_PATCHLEVEL for Ruby 4.0.0+. Is that okay?

Discussion:

  • shyouhei: (initial author of tool/merger.rb) I’m fine with that. Others?
  • hsbt: do we want to eliminate -1?
  • k0kubun: No, as @eregon says someone wants to detect if a version is under development or not.

Conclusion:

  • shyouhei: OK, let’s step forward.

[Misc #21768] Remove deprecated functions

  • Functions to be removed:
    • rb_clear_constant_cache deprecated for 3 years
    • postponed job APIs deprecated for 2 years
    • old APIs to allocate a data object deprecated for 5 years
    • rb_complex_polar deprecated for 7 years
    • rb_clone_setup and rb_dup_setup deprecated for 4 years
    • rb_gc_force_recycle deprecated as “removed soon”
    • taintedness/trustedness enums/macros deprecated for 4 years
    • RUBY_FL_DUPPED deprecated for 4 years
    • rb_iterate deprecated since 1.9
    • struct RData deprecated by struct RTypedData

Discussion:

  • shyouhei: I once tried killing rb_iterate in 3.0 era, but it turned out that there still were actual usage of the function.

Conclusion:

  • mame: Let’s try for 4.1

reminder

  • https://bugs.ruby-lang.org/issues/20409 Missing reporting some invalid breaks
  • https://github.com/ruby/ruby/pull/15189 Array#rfind, ok to merge? > matz
  • https://bugs.ruby-lang.org/issues/21520 Feature Proposal: Enumerator::Lazy#tee
module Enumerable
  def tee(...)
    each(...)
    self
  end
end

(1..).lazy.map { |x| puts x; x }.select { it.even? }.first(3)
(1..).lazy.select { p it; it.even? }.first(3)

[1, 2, 3].lazy.tee { p it }.map { it.to_s }
[1, 2, 3]     .tee { p it }.map { it.to_s }
[1, 2, 3]                  .map { p it; it.to_s }

[1, 2, 3].tee { p it }.map { p [it] } #=> ["1", "2", "3"]
# 1
# [1]
# 2
# [2]
# 3
# [3]

[1, 2, 3].map { it.to_s }.each { p [it] }
[1, 2, 3].lazy.map { it.to_s }.each { p [it] }
  • matz: let me consider more

misc.

  • https://github.com/ruby/ruby/pull/15424
  • -d
  • $DEBUG usage: https://gist.github.com/ko1/a15e9b1ec84e0bbe36503c0112662134

parser fixes

Looks like the fixes are ready for both parse.y and prism. When will we merge them? After 4.0.0 release?

  • https://bugs.ruby-lang.org/issues/18878 parse.y: Foo::Bar {} is inconsistently rejected
  • https://bugs.ruby-lang.org/issues/21168 Prism doesn’t require argument parentheses (in some cases) when a block is present but parse.y does
  • https://bugs.ruby-lang.org/issues/21669 Thoroughly implement void value expression check
    • matz: Let’s give it a try for Ruby 4.1

Which behavior is preferable?

  • https://bugs.ruby-lang.org/issues/21712 Prism and parse.y inconsistency in command call with block and .()
    • a b do end.() prism: OK, parse.y: SyntaxError->OK
    • a b do end.call() prism: OK, parse.y: OK
    • matz: It should parse

misc.

  • https://bugs.ruby-lang.org/issues/20959 Add a way to get codepage of console. Encoding.find("console")
  • https://bugs.ruby-lang.org/issues/21684 Does IO#pos clear the buffer?
  • https://bugs.ruby-lang.org/issues/21686 In combination with IO#ungetbyte, the write position may become unpredictable.
  • https://bugs.ruby-lang.org/issues/21690 Inconsistent rb_popcount64() definition
  • https://bugs.ruby-lang.org/issues/21693 Allow calling any callable object as a method
  • https://bugs.ruby-lang.org/issues/21374 FrozenError message is inconsistent when a singleton method is defined on a frozen object
  • https://bugs.ruby-lang.org/issues/21706 Add SIMD optimizations for string comparison operations
  • https://bugs.ruby-lang.org/issues/21715 Miscompilation on x86-64-v2 due to undefined behavior in search_nonascii in string.c
  • https://bugs.ruby-lang.org/issues/21709 Regexp interpolation is inconsistent with String interpolation
    • /#{ '\p{Hiragana}'.encode("US-ASCII") }\u1234/ #=> encoding mismatch in dynamic regexp : US-ASCII and UTF-8
    • /#{ 'p{Hiragana}'.encode("US-ASCII") }\u1234/ #=> no error
  • https://bugs.ruby-lang.org/issues/21723 binding.irb raises a LoadError under bundle exec
  • https://bugs.ruby-lang.org/issues/21721 Allow Queue and SizedQueue to be used as LIFO queues
  • https://bugs.ruby-lang.org/issues/21005 Update the source location method to include line start/stop and column start/stop details
irb(main):001> def foo = nil; method(:foo).source_location
=> ["(irb)", 1, 0, 1, 13]

irb(main):002> proc {}.source_location
=> ["(irb)", 2, 5, 2, 7]

irb(main):003> binding.source_location
=> ["(irb)", 3]

irb(main):004> b = binding
=> #<Binding:0x000073d738ed70a0>

irb(main):005> b.source_location
=> ["(irb)", 4]
  • https://bugs.ruby-lang.org/issues/21767 Consider procs which self is Ractor-shareable as Ractor shareable
  • https://bugs.ruby-lang.org/issues/21720 Add a native Binary Heap / Priority Queue to Ruby’s Standard Library (heapify, heappush, heappop)

box

autoload

  • https://bugs.ruby-lang.org/issues/21719 Thread deadlock with explicit require of a base clase in Linux Ruby 3.4
# start.rb
#
autoload :Target, "./target"
Thread.new { Target }
require "./target"

# target.rb
#
class Target
end