July 2025 - ruby-core - ml.ruby-lang.org

[ruby-core:122796] [Ruby Feature#17473] Make Pathname to embedded class of Ruby
by Eregon (Benoit Daloze) 16 Jul '25

16 Jul '25

Issue #17473 has been updated by Eregon (Benoit Daloze). hsbt (Hiroshi SHIBATA) wrote in #note-24: > >I have merged https://github.com/ruby/pathname/pull/53, it is clearly better for the maintenance of the gem at least. > > I reverted that. Please don't merge without consensus. At least, no one strongly against migrate C to Ruby in Dev Meetings. > > The transition from C to Ruby should be done gradually, in reviewable chunks. That PR is actually easy to review as explained in its description, it restores pathname.rb as it used to be and includes all changes to pathname.rb since then, see https://github.com/ruby/pathname/pull/53/files/3736eab91f783cb087f873b99c25…. IOW, it only adds code in pathname.rb and removes code in pathname.c, and then has separate commits for all other changes. Making multiple smaller PRs does not make sense, especially since the Ruby code is just the one from before the migration to C. I remade the PR as https://github.com/ruby/pathname/pull/57, please review it and share your concerns there. What is the plan for the code in the gem and the code in core? Will they be sync'd, how? From what I have seen changes in core have been done seemingly without caring about the gem, so it seemed the intention was to not sync but effectively fork? We also cannot make the gem empty in general as the gem needs to keep working on Ruby 3.4 and older at least. ---------------------------------------- Feature #17473: Make Pathname to embedded class of Ruby https://bugs.ruby-lang.org/issues/17473#change-114077 * Author: hsbt (Hiroshi SHIBATA) * Status: Closed * Assignee: akr (Akira Tanaka) ---------------------------------------- pathname is one of most useful utility class of Ruby. I'm happy to use Pathname without require it. Any thought? -- https://bugs.ruby-lang.org/

1 0

[ruby-core:122795] [Ruby Feature#17473] Make Pathname to embedded class of Ruby
by Eregon (Benoit Daloze) 16 Jul '25

16 Jul '25

Issue #17473 has been updated by Eregon (Benoit Daloze). Dan0042 (Daniel DeLorme) wrote in #note-23: > I don't understand this at all. What's the problem with having `require "fileutils"` inside #mkpath ? Those dependencies are part of the stdlib anyway, so it's not like they could be missing. Is it for some reason forbidden to call `require` in a method of a core class? I had the exact same thought, this seems the obvious solution. ---------------------------------------- Feature #17473: Make Pathname to embedded class of Ruby https://bugs.ruby-lang.org/issues/17473#change-114076 * Author: hsbt (Hiroshi SHIBATA) * Status: Closed * Assignee: akr (Akira Tanaka) ---------------------------------------- pathname is one of most useful utility class of Ruby. I'm happy to use Pathname without require it. Any thought? -- https://bugs.ruby-lang.org/

1 0

[ruby-core:122345] [Ruby Feature#21386] Introduce `Enumerable#join_map`
by matheusrich (Matheus Richard) 16 Jul '25

16 Jul '25

Issue #21386 has been reported by matheusrich (Matheus Richard). ---------------------------------------- Feature #21386: Introduce `Enumerable#join_map` https://bugs.ruby-lang.org/issues/21386 * Author: matheusrich (Matheus Richard) * Status: Open ---------------------------------------- ### Problem The pattern `.map { ... }.join(sep)` is extremely common in Ruby codebases: ```ruby users.map(&:name).join(", ") ``` It’s expressive but repetitive (both logically and computationally). This pattern allocates an intermediate array and does two passes over the collection. Real-world usage is widespread: - [Open source Ruby projects using this pattern](https://github.com/search?q=lang%3Aruby+%2F%5C.map%5Cs*%5C%7B%5B%5… - [Within Rails](https://github.com/search?q=lang%3Aruby+%2F%5C.map%5Cs*%5C%7B%5B%5E%… - [Within Ruby itself](https://github.com/search?q=lang%3Aruby+%2F%5C.map%5Cs*%5C%7B%5B%5E… ### Proposal Just like `filter_map` exists to collapse a common `map + compact`, this proposal introduces `Enumerable#join_map`, which maps and joins in a single pass. ```ruby users.join_map(", ", &:name) ``` A Ruby implementation could look like this: ```ruby module Enumerable def join_map(sep = "") return "" unless block_given? str = +"" first = true each do |item| str << sep unless first str << yield(item).to_s first = false end str end end ``` The name `join_map` follows the precedent of `filter_map`, emphasizing the final operation (`join`) over the intermediate (`map`). ### Prior Art Some other languages have similar functionality, but with different names or implementations: #### Elixir Elixir has this via [the `Enum.map_join/3` function](https://hexdocs.pm/elixir/1.12/Enum.html#map_join/3): ```elixir Enum.map_join([1, 2, 3], &(&1 * 2)) "246" Enum.map_join([1, 2, 3], " = ", &(&1 * 2)) "2 = 4 = 6" ``` #### Crystal Crystal, on the other hand, [uses `Enumerable#join` with a block](https://crystal-lang.org/api/1.16.3/Enumerable.html#join%28separator…: ```crystal [1, 2, 3].join(", ") { |i| -i } # => "-1, -2, -3" ``` #### Kotlin Kotlin has a similar [function called `joinToString`](https://kotlinlang.org/api/core/kotlin-stdlib/kotlin.collections/join-to-string.html) that can take a transformation function: ```kotlin val chars = charArrayOf('a', 'b', 'c') println(chars.joinToString() { it.uppercaseChar().toString() }) // A, B, C ``` -- https://bugs.ruby-lang.org/

9 11

[ruby-core:122793] [Ruby Feature#17473] Make Pathname to embedded class of Ruby
by hsbt (Hiroshi SHIBATA) 16 Jul '25

16 Jul '25

Issue #17473 has been updated by hsbt (Hiroshi SHIBATA). >I have merged https://github.com/ruby/pathname/pull/53, it is clearly better for the maintenance of the gem at least. I reverted that. Please don't merge without consensus. At least, no one strongly against migrate C to Ruby in Dev Meetings. The transition from C to Ruby should be done gradually, in reviewable chunks. ---------------------------------------- Feature #17473: Make Pathname to embedded class of Ruby https://bugs.ruby-lang.org/issues/17473#change-114074 * Author: hsbt (Hiroshi SHIBATA) * Status: Closed * Assignee: akr (Akira Tanaka) ---------------------------------------- pathname is one of most useful utility class of Ruby. I'm happy to use Pathname without require it. Any thought? -- https://bugs.ruby-lang.org/

1 0

[ruby-core:122791] [Ruby Feature#17473] Make Pathname to embedded class of Ruby
by Dan0042 (Daniel DeLorme) 15 Jul '25

15 Jul '25

Issue #17473 has been updated by Dan0042 (Daniel DeLorme). > * akr: To use the following four methods, require "pathname" is needed for a while > * find, mkpath, rmtree, mktmpdir > * rationale: they require other libraries (find gem, fileutils, tmpdir) > * in future, they (except mktmpdir) should be rewritten in pure C (or rbinc) without dependency to find and fileutils I don't understand this at all. What's the problem with having `require "fileutils"` inside #mkpath ? Those dependencies are part of the stdlib anyway, so it's not like they could be missing. Is it for some reason forbidden to call `require` in a method of a core class? ---------------------------------------- Feature #17473: Make Pathname to embedded class of Ruby https://bugs.ruby-lang.org/issues/17473#change-114072 * Author: hsbt (Hiroshi SHIBATA) * Status: Closed * Assignee: akr (Akira Tanaka) ---------------------------------------- pathname is one of most useful utility class of Ruby. I'm happy to use Pathname without require it. Any thought? -- https://bugs.ruby-lang.org/

1 0

[ruby-core:122773] [Ruby Feature#21513] Raise on converting endless range to set
by viralpraxis (Iaroslav Kurbatov) 15 Jul '25

15 Jul '25

Issue #21513 has been reported by viralpraxis (Iaroslav Kurbatov). ---------------------------------------- Feature #21513: Raise on converting endless range to set https://bugs.ruby-lang.org/issues/21513 * Author: viralpraxis (Iaroslav Kurbatov) * Status: Open ---------------------------------------- Converting endless range to array raises: ``` shell ruby -e '(1..).to_a' -e:1:in 'Range#to_a': cannot convert endless range to an array (RangeError) from -e:1:in '<main>' ``` but converting to set does not: ``` shell ruby -e '(1..).to_set # hangs' ``` I think it makes to raise in both cases for consistency. Something like this should do the trick: ``` diff diff --git i/prelude.rb w/prelude.rb index f49cada637..11bfa3fc95 100644 --- i/prelude.rb +++ w/prelude.rb @@ -30,6 +30,10 @@ module Enumerable # Makes a set from the enumerable object with given arguments. # Passing arguments to this method is deprecated. def to_set(*args, &block) + if self.class == Range && self.end.nil? + raise RangeError, "cannot convert endless range to a set" + end + klass = if args.empty? Set else diff --git i/test/ruby/test_range.rb w/test/ruby/test_range.rb index f875c0ab40..27b968641b 100644 --- i/test/ruby/test_range.rb +++ w/test/ruby/test_range.rb @@ -1541,4 +1541,11 @@ def test_overlap? assert_not_operator((1...3), :overlap?, (3..4)) assert_not_operator((...3), :overlap?, (3..)) end + + def test_to_set + assert_equal(Set[], (1..-1).to_set) + assert_equal(Set[1, 2, 3], (1..3).to_set) + + assert_raise(RangeError) { (239..).to_set } + end end ``` If this patch is accepted, I'll open a PR. ``` shell ./ruby -v ruby 3.5.0dev (2025-07-14T20:34:32Z master a6d483971a) +PRISM [x86_64-linux] ``` -- https://bugs.ruby-lang.org/

3 5

[ruby-core:112664] [Ruby master Bug#19473] can't be called from trap context (ThreadError) is too limiting
by Eregon (Benoit Daloze) 15 Jul '25

15 Jul '25

Issue #19473 has been reported by Eregon (Benoit Daloze). ---------------------------------------- Bug #19473: can't be called from trap context (ThreadError) is too limiting https://bugs.ruby-lang.org/issues/19473 * Author: Eregon (Benoit Daloze) * Status: Open * Priority: Normal * Backport: 2.7: UNKNOWN, 3.0: UNKNOWN, 3.1: UNKNOWN, 3.2: UNKNOWN ---------------------------------------- Simple reproducer: ``` $ ruby -ve 'm=Mutex.new; trap(:HUP) { m.synchronize { p :OK } }; Process.kill :HUP, Process.pid; sleep 0.1' ruby 3.2.1 (2023-02-08 revision 31819e82c8) [x86_64-linux] -e:1:in `synchronize': can't be called from trap context (ThreadError) from -e:1:in `block in <main>' from -e:1:in `kill' from -e:1:in `<main>' ``` Expected behavior: ``` $ ruby -ve 'm=Mutex.new; trap(:HUP) { m.synchronize { p :OK } }; Process.kill :HUP, Process.pid; sleep 0.1' truffleruby 22.3.1, like ruby 3.0.3, GraalVM CE Native [x86_64-linux] :OK $ ruby -ve 'm=Mutex.new; trap(:HUP) { m.synchronize { p :OK } }; Process.kill :HUP, Process.pid; sleep 0.1' jruby 9.4.0.0 (3.1.0) 2022-11-23 95c0ec159f OpenJDK 64-Bit Server VM 17.0.6+10 on 17.0.6+10 +jit [x86_64-linux] :OK ``` This exception is highly problematic, for instance it breaks `Timeout.timeout` in `trap`: https://github.com/ruby/timeout/issues/17#issuecomment-1142035939 I suppose this behavior is because *sometimes* it's problematic to lock a Mutex in trap, e.g., if it's already locked by the main thread/fiber. However, that's just one case, not all, so we should not raise an exception early like that. There seems to be no valid reason to prevent *all* `Mutex#synchronize` in `trap`. After all, if the Mutex for instance is only used in `trap`, it's well-defined AFAIK. For instance a given trap handler does not seem executed concurrently: ``` $ ruby -ve 'trap(:HUP) { puts "in trap\n"+caller.join("\n")+"\n\n"; sleep 0.1 }; pid = Process.pid; Process.wait fork { 20.times { Process.kill :HUP, pid } }; sleep 1' ruby 3.2.1 (2023-02-08 revision 31819e82c8) [x86_64-linux] in trap -e:1:in `wait' -e:1:in `<main>' in trap -e:1:in `wait' -e:1:in `<main>' in trap -e:1:in `wait' -e:1:in `<main>' in trap -e:1:in `wait' -e:1:in `<main>' in trap -e:1:in `wait' -e:1:in `<main>' in trap -e:1:in `wait' -e:1:in `<main>' ``` And if the trap handler using the Mutex is never called while the Mutex is held by the main thread/fiber, there is also no problem. -- https://bugs.ruby-lang.org/

8 27

[ruby-core:122782] [Ruby Feature#17473] Make Pathname to embedded class of Ruby
by Eregon (Benoit Daloze) 15 Jul '25

15 Jul '25

Issue #17473 has been updated by Eregon (Benoit Daloze). I have merged https://github.com/ruby/pathname/pull/53, it is clearly better for the maintenance of the gem at least. Now we need to figure out if we want to keep the copy in core in sync with the gem or not (if not it's going to be a "fun" gotcha that the pathname gem is faster than core Pathname). ---------------------------------------- Feature #17473: Make Pathname to embedded class of Ruby https://bugs.ruby-lang.org/issues/17473#change-114061 * Author: hsbt (Hiroshi SHIBATA) * Status: Closed * Assignee: akr (Akira Tanaka) ---------------------------------------- pathname is one of most useful utility class of Ruby. I'm happy to use Pathname without require it. Any thought? -- https://bugs.ruby-lang.org/

1 0

[ruby-core:122781] [Ruby Feature#17473] Make Pathname to embedded class of Ruby
by Eregon (Benoit Daloze) 15 Jul '25

15 Jul '25

Issue #17473 has been updated by Eregon (Benoit Daloze). Discussion at https://github.com/ruby/dev-meeting-log/blob/master/2025/DevMeeting-2025-07… I am disappointed, https://github.com/ruby/pathname/pull/53 makes the code clearly more maintainable and a lot easier to read, there is no value to have it in C. I don't see much value to have Pathname as a core class, `require "pathname"` is no big deal and since it depends on other gems & stdlib it's awkward. Having a partially-defined Pathname in core (missing some methods) seems the worst outcome. As pathname is already a gem, it also means we'll have two unsynchronized copies of the code, and with enough time maybe subtle behavior differences, this seems something nobody wants. @matz this decision is adding a lot of work to CRuby, and other Ruby implementations. Pathname or most of it should be defined in Ruby, having multiple copies of it in C, Java, Ruby, etc means useless duplication and things will get out of sync needlessly. > @nobu: if Pathname is embeded to the core, I prefer the C implementation to Ruby impl. Why? I see only disadvantages to that. > @ioquatix for the purpose of JIT isn't it better to write more Ruby in Ruby? > @shyouhei nobody wants to JIT a path manipulation. https://github.com/ruby/pathname/pull/53#issuecomment-2989134661 shows the Ruby code is faster than the C extension for Pathname. Even from a performance point of view it's bad to implement Pathname methods in C (for most methods). ---------------------------------------- Feature #17473: Make Pathname to embedded class of Ruby https://bugs.ruby-lang.org/issues/17473#change-114060 * Author: hsbt (Hiroshi SHIBATA) * Status: Closed * Assignee: akr (Akira Tanaka) ---------------------------------------- pathname is one of most useful utility class of Ruby. I'm happy to use Pathname without require it. Any thought? -- https://bugs.ruby-lang.org/

1 0

[ruby-core:122742] [Ruby Bug#21512] Socket.tcp_with_fast_fallback('example.com', 80, '127.0.0.1') fails with unknown socket domain: ipv4 (SocketError)
by osyoyu (Daisuke Aritomo) 15 Jul '25

15 Jul '25

Issue #21512 has been reported by osyoyu (Daisuke Aritomo). ---------------------------------------- Bug #21512: Socket.tcp_with_fast_fallback('example.com', 80, '127.0.0.1') fails with unknown socket domain: ipv4 (SocketError) https://bugs.ruby-lang.org/issues/21512 * Author: osyoyu (Daisuke Aritomo) * Status: Open * ruby -v: ruby 3.5.0dev (2025-07-13T04:37:15Z master f03cc0b514) +PRISM [x86_64-linux] * Backport: 3.2: UNKNOWN, 3.3: UNKNOWN, 3.4: UNKNOWN ---------------------------------------- The following code is expected to open a socket bound to the local IPv4 interface, but raises an SocketError. ```ruby require 'socket' Socket.tcp_with_fast_fallback('localhost', 8080, '127.0.0.1') ``` Expected behavior: Returns a `Socket` object Actuai behavior: Causes a SocketError (unknown socket domain: ipv4) I believe this is because Socket.tcp_with_fast_fallback is calling Addrinfo.getaddrinfo with `:ipv4` / `:ipv6` as its third argument, instead of `Socket::AF_INET` / `Socket::AF_INET6`. https://github.com/ruby/ruby/blob/b2a7b7699261d2a4ef8a9d5d38d3fb9dc99c8253/… -- https://bugs.ruby-lang.org/

3 3