October 2023 - ruby-core - ml.ruby-lang.org

[ruby-core:112211] [Ruby master Bug#19411] GC issue with moved objects

by luke-gru (Luke Gruber)

Issue #19411 has been reported by luke-gru (Luke Gruber). ---------------------------------------- Bug #19411: GC issue with moved objects https://bugs.ruby-lang.org/issues/19411 * Author: luke-gru (Luke Gruber) * Status: Open * Priority: Normal * ruby -v: 3.2.0 * Backport: 2.7: UNKNOWN, 3.0: UNKNOWN, 3.1: UNKNOWN, 3.2: UNKNOWN ---------------------------------------- This crashes: ```ruby class Obj def initialize @obj = 3 end end GC.stress = true r = Ractor.new do obj = receive p obj end obj = Obj.new r.send(obj, move: true) r.take ``` It only crashes with nested objects, if you remove the ivar set in `initialize` it works fine. Maybe missing `RB_GC_GUARD`? -- https://bugs.ruby-lang.org/

4 months

1
2
0 0

[ruby-core:112010] [Ruby master Bug#19375] File objects are currently shareable

by luke-gru (Luke Gruber)

Issue #19375 has been reported by luke-gru (Luke Gruber). ---------------------------------------- Bug #19375: File objects are currently shareable https://bugs.ruby-lang.org/issues/19375 * Author: luke-gru (Luke Gruber) * Status: Open * Priority: Normal * Backport: 2.7: UNKNOWN, 3.0: UNKNOWN, 3.1: UNKNOWN, 3.2: UNKNOWN ---------------------------------------- I don't know the internals of file.c but I don't think files are thread-safe. -- https://bugs.ruby-lang.org/

4 months, 1 week

1
2
0 0

[ruby-core:111789] [Ruby master Feature#19333] Setting (Fiber Local|Thread Local|Fiber Storage) to nil should delete value in order to avoid memory leaks.

by ioquatix (Samuel Williams)

Issue #19333 has been reported by ioquatix (Samuel Williams). ---------------------------------------- Feature #19333: Setting (Fiber Local|Thread Local|Fiber Storage) to nil should delete value in order to avoid memory leaks. https://bugs.ruby-lang.org/issues/19333 * Author: ioquatix (Samuel Williams) * Status: Open * Priority: Normal ---------------------------------------- As it stands, Fiber Locals, Thread Locals and Fiber Storage have no way of deleting key-value associations. ```ruby 100.times do |i| name = :"variable-#{i}" Thread.current[name] = 10 end ``` Because of this, dynamically generated associations can leak over time. This is worse for things like Threads that might be pooled (or maybe an argument against user-space pooling). In any case, having a way to delete those associations would allow application code to at least delete the associations when they no longer make sense. I propose that assigning `nil` to "locals" or "storage" should effectively delete them. e.g. ```ruby 100.times do |i| name = :"variable-#{i}" Thread.current[name] = 10 Thread.current[name] = nil # delete association end ``` A more invasive alternative would be to define new interfaces like `Thread::Local`, `Fiber::Local` and `Fiber::Storage::Local` (or something) which correctly clean up on GC. -- https://bugs.ruby-lang.org/

4 months, 1 week

3
8
0 0

[ruby-core:112446] [Ruby master Bug#19442] Remove USE_RINCGC flag

by eightbitraptor (Matthew Valentine-House)

Issue #19442 has been reported by eightbitraptor (Matthew Valentine-House). ---------------------------------------- Bug #19442: Remove USE_RINCGC flag https://bugs.ruby-lang.org/issues/19442 * Author: eightbitraptor (Matthew Valentine-House) * Status: Open * Priority: Normal * Backport: 2.7: UNKNOWN, 3.0: UNKNOWN, 3.1: UNKNOWN, 3.2: UNKNOWN ---------------------------------------- [GitHub PR #7317](https://github.com/ruby/ruby/pull/7313) Ruby doesn't compile when USE_RINCGC is disabled. It's also not being tested in CI. As @nobu has pointed out in comments on the PR, fixing it is simple. I think there are 2 approaches we could take: 1. Remove `USE_RINCGC` entirely and always run with incremental GC enabled. We have a precedent for this: `USE_RGENGC=0` was removed [in this commit](https://github.com/ruby/ruby/commit/62c2b8c74e47652fc5bbaf6150f4acd… almost 3 years ago. `USE_RINCGC=0` is in a similar state. It has been broken for almost a year (since [this commit](https://github.com/ruby/ruby/commit/dde164e968e382d50b07ad455946888…), and [disabled in CI for more than 2 years](https://github.com/ruby/ruby/commit/46d3ea2c2569e2e5a9ee3e7e206f07f0…. 2. We could fix `USE_RINCGC`. If we do this we should re-enable the CI task so that we continue to keep it working in the future. I am very much in favour of option 1, because I don't think this flag is being actively used, and removing it will simplify the code and reduce the cognitive overhead of working with the GC. However I am happy to defer to the team here. -- https://bugs.ruby-lang.org/

4 months, 1 week

6
9
0 0

[ruby-core:115208] [Ruby master Feature#19982] Bump required Visual Studio version to 2015 after 3.3

by kddnewton (Kevin Newton)

Issue #19982 has been reported by kddnewton (Kevin Newton). ---------------------------------------- Feature #19982: Bump required Visual Studio version to 2015 after 3.3 https://bugs.ruby-lang.org/issues/19982 * Author: kddnewton (Kevin Newton) * Status: Open * Priority: Normal ---------------------------------------- Visual Studio 2013 is finally being EOL'd next April. (https://learn.microsoft.com/en-us/visualstudio/productinfo/vs-servicing#old…) Let's drop support for it after Ruby 3.3 is released. If we bump the required version to Visual Studio 2015, we'll get better C99 support and be able to rely on some more standards. Personally I'm excited about `snprintf` being actually implemented. (https://learn.microsoft.com/en-us/cpp/overview/visual-cpp-language-conforma…) As support for the argument, CPython requires VS2017 (https://github.com/python/cpython/blob/main/PCbuild/readme.txt) and gnulib requires VS2015(https://www.gnu.org/software/gnulib/manual/html_node/Supported-Platf…. -- https://bugs.ruby-lang.org/

4 months, 2 weeks

5
5
0 0

[ruby-core:113951] [Ruby master Bug#19738] `ObjectSpace.each_object.to_a` crashes in `make runirb`.

by ioquatix (Samuel Williams)

Issue #19738 has been reported by ioquatix (Samuel Williams). ---------------------------------------- Bug #19738: `ObjectSpace.each_object.to_a` crashes in `make runirb`. https://bugs.ruby-lang.org/issues/19738 * Author: ioquatix (Samuel Williams) * Status: Open * Priority: Normal * Backport: 3.0: UNKNOWN, 3.1: UNKNOWN, 3.2: UNKNOWN ---------------------------------------- ``` > make runirb RUBY_ON_BUG='gdb -x ./.gdbinit -p' ./miniruby -I./lib -I. -I.ext/common ./tool/runruby.rb --extout=.ext -- --disable-gems -r irb -e 'IRB.start("make runirb")' irb(main):001:0> ObjectSpace.each_object.to_a Assertion Failed: ./vm_method.c:1366:callable_method_entry_or_negative:RB_TYPE_P(klass, T_CLASS) || RB_TYPE_P(klass, T_ICLASS) ruby 3.3.0dev (2023-06-19T19:19:45Z master 9ff4399dec) [arm64-darwin22] ``` I have no idea why. ---Files-------------------------------- error.txt (12.9 KB) -- https://bugs.ruby-lang.org/

4 months, 2 weeks

2
3
0 0

[ruby-core:114345] [Ruby master Bug#19831] warning message of linker with macOS Sonoma beta

by hsbt (Hiroshi SHIBATA)

Issue #19831 has been reported by hsbt (Hiroshi SHIBATA). ---------------------------------------- Bug #19831: warning message of linker with macOS Sonoma beta https://bugs.ruby-lang.org/issues/19831 * Author: hsbt (Hiroshi SHIBATA) * Status: Open * Priority: Normal * Backport: 3.0: UNKNOWN, 3.1: UNKNOWN, 3.2: UNKNOWN ---------------------------------------- Xcode 15 beta and macOS Sonoma beta show the following warnings with `make` ``` (snip) linking miniruby ld: warning: ignoring duplicate library '-lpthread' miniruby: replacing existing signature exe/ruby: replacing existing signature builtin_binary.inc updated compiling builtin.c linking static-library libruby.3.3-static.a linking ruby ld: warning: ignoring duplicate library '-lgmp' ld: warning: ignoring duplicate library '-ldl' ld: warning: ignoring duplicate library '-lobjc' ld: warning: ignoring duplicate library '-lpthread' ld: warning: ignoring duplicate library '-lpthread' (snip) ``` I'm not investigate that yet. It may be caused by `ld-prime` that is new linker of macOS platform. ``` $ pkgutil --pkg-info=com.apple.pkg.CLTools_Executables package-id: com.apple.pkg.CLTools_Executables version: 15.0.0.0.1.1690355577 volume: / location: / install-time: 1690967267 ``` -- https://bugs.ruby-lang.org/

4 months, 2 weeks

1
3
0 0

[ruby-core:111739] [Ruby master Feature#19324] Enumerator.product => Enumerable#product

by zverok (Victor Shepelev)

Issue #19324 has been reported by zverok (Victor Shepelev). ---------------------------------------- Feature #19324: Enumerator.product => Enumerable#product https://bugs.ruby-lang.org/issues/19324 * Author: zverok (Victor Shepelev) * Status: Open * Priority: Normal ---------------------------------------- I know it might be too late after introducing a feature and releasing a version, but I find `Enumerator.product` quite confusing, and can't find any justification in #18685. **Problem 1: It is `Array#product` but `Enumerator.product`** ```ruby [1, 2].product([4, 5]) # => [[1, 4], [1, 5], [2, 4], [2, 5]] # Usually, when we add methods to Enumerable/Enumerator which # already array had before, it is symmetric, say... [1, nil, 2, 3].compact #=> [1, 2, 3] [1, nil, 2, 3].lazy.compact.first(2) #=> [1, 2] # But not in this case: [1, 2].lazy.product([4, 5]).first(2) # undefined method `product' for #<Enumerator::Lazy: [1, 2]> (NoMethodError) # Because you "just" need to change it to: Enumerator.product([1, 2].lazy, [4, 5]).first(2) # => [[1, 4], [1, 5]] ``` No other method was "promoted" from Array this way And in general, I believe core methods tend to belong to the first object in the expression and not be free module methods, Elixir style. **Problem 2: It is one letter different from `Enumerator.produce`** I understand I might be biased here (as a person who proposed `produce`), and that method is not as popular (yet?) as I hoped, but still, two methods that do completely different things and differ by one letter, both being somewhat vague verbs (so it is easy to confuse them unless you did a lot of math and "product" is firmly set for set product in your head). I believe that EITHER of two problems would be concerning enough, but the combination of them seems to be a strong enough argument to make the change?.. (Maybe with graceful deprecation of module method in one next version, but, considering the Ruby 3.2 is just released, maybe vice versa, fix the problem in the next minor release?..) -- https://bugs.ruby-lang.org/

4 months, 2 weeks

6
17
0 0

[ruby-core:111988] [Ruby master Feature#19370] Anonymous parameters for blocks?

by zverok (Victor Shepelev)

Issue #19370 has been reported by zverok (Victor Shepelev). ---------------------------------------- Feature #19370: Anonymous parameters for blocks? https://bugs.ruby-lang.org/issues/19370 * Author: zverok (Victor Shepelev) * Status: Open * Priority: Normal ---------------------------------------- Just to clarify: are anonymous parameters delegation is planned to support in blocks? It would be a nice addition, if it is possible to implement: ```ruby # data in form [request method, URL, params]: [ [:get, 'https://google.com', {q: 'Ruby'}, {'User-Argent': 'Google-Chrome'}], [:post, 'https://gist.github.com', 'body'], # ... ].each { |method, *| request(method.to_s.upcase, *) } ``` ...and at the very least, consistent with what the method definition can have. If they are NOT planned to be implemented, I believe that at least error messages should be made much clearer, because currently, this would happen while running the code above: > no anonymous rest parameter (SyntaxError) I understand the reason (the `request` clause doesn't "see" anonymous parameter of the **block**, and claims that current **method** doesn't have them), but it looks honestly confusing and inconsistent. -- https://bugs.ruby-lang.org/

4 months, 2 weeks

5
9
0 0

[ruby-core:115188] [Ruby master Bug#19977] (nil..nil) === x can raise an exception, differing from Range#cover?

by kyanagi (Kouhei Yanagita)

Issue #19977 has been reported by kyanagi (Kouhei Yanagita). ---------------------------------------- Bug #19977: (nil..nil) === x can raise an exception, differing from Range#cover? https://bugs.ruby-lang.org/issues/19977 * Author: kyanagi (Kouhei Yanagita) * Status: Open * Priority: Normal * ruby -v: ruby 3.3.0dev (2023-10-27T03:13:17Z master f9f0cfe785) [arm64-darwin22] * Backport: 3.0: UNKNOWN, 3.1: UNKNOWN, 3.2: UNKNOWN ---------------------------------------- I investigated Range#=== and Range#cover?, and found that the only difference in behavior between them would be that `(nil..nil) === x` could throw an exception. ``` % ~/tmp/ruby-master/bin/ruby -v ruby 3.3.0dev (2023-10-27T03:13:17Z master f9f0cfe785) [arm64-darwin22] ``` ``` % ~/tmp/ruby-master/bin/ruby -e 'p (nil..nil) === "a"' -e:1:in `===': cannot determine inclusion in beginless/endless ranges (TypeError) p (nil..nil) === "a" ^^^ from -e:1:in `<main>' ``` ``` % ~/tmp/ruby-master/bin/ruby -e 'p (nil..nil).cover?("a")' true ``` Is this difference intended? According to NEWS, `Range#===` uses `cover?` since Ruby 2.6 (For `String`, since Ruby 2.7). Following this, `(nil..nil) === x` should not throw an exception in the same way as `(nil..nil).cover?(x)`. history: `(nil..nil) === "a"` throws an exception since https://github.com/ruby/ruby/commit/04a92a6. For "linear objects" (`Integer`, `Float`, `Numeric`, `Time`), it has beed fixed not to throw an exception on https://github.com/ruby/ruby/commit/fb17c83. related issues: * [Bug #15449] Range#=== is not using cover in Ruby 2.6 * [Bug #18580] Range#include? inconsistency for beginless String ranges * [Bug #19533] Behavior of ===/include? on a beginless/endless range (nil..nil) changed in ruby 3.2 -- https://bugs.ruby-lang.org/

4 months, 2 weeks

2
5
0 0

2024

2023

2022

ruby-core October 2023