December 2024 - ruby-core - ml.ruby-lang.org

[ruby-core:111809] [Ruby master Bug#19338] Ruby hangs when defining new constant in ractor
by luke-gru (Luke Gruber) 13 May '25

13 May '25

Issue #19338 has been reported by luke-gru (Luke Gruber). ---------------------------------------- Bug #19338: Ruby hangs when defining new constant in ractor https://bugs.ruby-lang.org/issues/19338 * Author: luke-gru (Luke Gruber) * Status: Open * Priority: Normal * Backport: 2.7: UNKNOWN, 3.0: UNKNOWN, 3.1: UNKNOWN, 3.2: UNKNOWN ---------------------------------------- This code causes Ruby to hang: ``` ruby rs = [] 2.times do rs << Ractor.new do MYCONSTANT = 2 end end rs.each(&:take) ``` There is a problem when the warning is being outputted with multiple ractors. A thread is calling RB_VM_LOCK() while holding the VM lock in ractor.c (ractor_check_blocking()) If the code is changed to RB_VM_LOCK_ENTER() and RB_VM_LOCK_LEAVE() then it fixes it, but I don't know if there's a better way. Thanks! -- https://bugs.ruby-lang.org/

3 3

[ruby-core:118467] [Ruby master Feature#20610] Float::INFINITY as IO.select timeout argument
by akr (Akira Tanaka) 13 May '25

13 May '25

Issue #20610 has been reported by akr (Akira Tanaka). ---------------------------------------- Feature #20610: Float::INFINITY as IO.select timeout argument https://bugs.ruby-lang.org/issues/20610 * Author: akr (Akira Tanaka) * Status: Open ---------------------------------------- I propose IO.select accepts Float::INFINITY as a timeout argument. It behaves the same as nil which means IO.select will block indefinitely. Motivation: Currently, the Ruby convention to indicate no timeout is using nil. This practice often forces us to treat the nil case separately. Conceptually, no timeout can be thought of as an infinite timeout. So I propose to accept Float::INFINITY as a timeout. It makes us less conditionals when we need to calculate or compare timeouts. Assume `now` method as follows to read the following examples. ``` def now = Process.clock_gettime(Process::CLOCK_MONOTONIC) ``` Example 1: absolute timeout Sometimes we maintain timeout as an absolute clock. The following method takes a relative timeout as an argument. It invokes IO.select several times and raises if the timeout is reached. Assuming a user must specify a finite timeout, the following definition is possible. ``` # user_timeout is the required argument def method(..., user_timeout) abs_timeout = now + user_timeout loop { IO.select(rs, ws, es, (abs_timeout - now).clamp(0..)) raise "timeout" if abs_timeout < now ... } end ``` Consider we need to make user_timeout optional. If user_timeout is not given, no timeout occurs. The implementation is as follows. I think the following implementation is typical. It needs 3 more conditionals than the above. ``` # user_timeout is an optional argument. nil means no timeout. def method(..., user_timeout=nil) abs_timeout = user_timeout ? now + user_timeout : nil loop { IO.select(rs, ws, es, user_timeout ? (abs_timeout - now).clamp(0..) : nil) raise "timeout" if abs_timeout && (abs_timeout < now) ... } end ``` It is possible to reduce a conditional if we use Float::INFINITY. (`abs_timeout && (abs_timeout < now)` is changed to `abs_timeout < now`) ``` # user_timeout is an optional argument. nil means no timeout. def method(..., user_timeout=nil) abs_timeout = user_timeout ? now + user_timeout : Float::INFINITY loop { IO.select(rs, ws, es, abs_timeout != Float::INFINITY ? (abs_timeout - now).clamp(0..) : nil) raise "timeout" if abs_timeout < now ... } end ``` If IO.select accepts Float::INFINITY as a timeout argument (this proposal), we can reduce one more conditional as follows. ``` # user_timeout is an optional argument. nil means no timeout. def method(..., user_timeout=nil) abs_timeout = user_timeout ? now + user_timeout : Float::INFINITY loop { IO.select(rs, ws, es, (abs_timeout - now).clamp(0..)) raise "timeout" if abs_timeout < now ... } end ``` Example 2: minimum of timeouts Sometimes we need to choose the minimum of several timeouts. I think many event-driven programs use this strategy to determine the timeout for select function. If "no timeout" is represented as nil, `[t1, t2, t3, ...].compact.min` is the minimum. If "no timeout" is represented as Float::INFINITY, we can remove `compact`: `[t1, t2, t3, ...].min` However, Float::INFINITY must be converted to nil for IO.select. This proposal removes this conversion. Example 3: maximum of timeouts Sometimes we need to choose the maximum of several timeouts. We encountered this situation with a Happy Eyeballs implementation. There are two timeouts for getaddrinfo and connect. We need to wait the longer timeout because a timeout for one doesn't stop another. Also, we don't ignore results after a timeout as long as the algorithm waits for something. If "no timeout" is represented as nil, `ts = [t1, t2, t3, ...]; ts.include?(nil) ? nil : ts.max` is the maximum. If "no timeout" is represented as Float::INFINITY, we can compute the maximum more easily: `[t1, t2, t3, ...].max` It makes code simpler. However, Float::INFINITY must be converted to nil for IO.select. This proposal removes this conversion. Several Consideration: Consideration 1: Methods other than IO.select. Several methods take a timeout. An incomplete list of methods is as follows. (I searched rb_time_interval.) - IO.select(rs, ws, es, timeout) - sleep(secs) - TCPSocket.new(connect_timeout:) - io.wait_readable(timeout) - io.wait_writable(timeout) - io.wait_priority(timeout) - mutex.sleep(timeout) Do we want to modify them consistently to accept Float::INFINITY? Consideration 2: C-level API If we want to change the timeout of many methods, we would wish to new C-level API similar to rb_time_interval but can return NULL. Unfortunately, rb_time_interval cannot return NULL because the return type is struct timeval. Note that ext/io/wait/wait.c contains get_timeout function. It seems a good first step for such API. Consideration 3: IEEE 754 dependency. Minor platforms (such as VAX) use non-IEEE 754 floating point numbers without infinity. Note that NetBSD/vax still works. (And there is an emulator, simh). Consideration 4: It seems no major languages accept infinity as select's timeout. I found Perl, Python, and OCaml take a floating point number as a timeout of select function. But they don't accept infinity. -- https://bugs.ruby-lang.org/

3 7

[ruby-core:116085] [Ruby master Bug#20165] Ractors moving a Struct breaks beyond the first 3 fields
by codekitchen (Brian Palmer) 09 May '25

09 May '25

Issue #20165 has been reported by codekitchen (Brian Palmer). ---------------------------------------- Bug #20165: Ractors moving a Struct breaks beyond the first 3 fields https://bugs.ruby-lang.org/issues/20165 * Author: codekitchen (Brian Palmer) * Status: Open * Priority: Normal * ruby -v: ruby 3.3.0 (2023-12-25 revision 5124f9ac75) [arm64-darwin23] * Backport: 3.0: UNKNOWN, 3.1: UNKNOWN, 3.2: UNKNOWN, 3.3: UNKNOWN ---------------------------------------- Experimenting with Ractors on ruby 3.3.0, and I'm seeing a bug where if you move a struct between ractors, all but the first 3 fields are set to nil. ``` ruby Foo = Struct.new(:a,:b,:c,:d,:e,:f) r = Ractor.new { foo = Foo[0,0,0,0,0,0] p foo Ractor.yield(foo, move: true) } p r.take ``` ``` ❯ ruby -v ruby 3.3.0 (2023-12-25 revision 5124f9ac75) [arm64-darwin23] ❯ ruby rbug.rb rbug.rb:3: warning: Ractor is experimental, and the behavior may change in future versions of Ruby! Also there are many implementation issues. #<struct Foo a=0, b=0, c=0, d=0, e=0, f=0> #<struct Foo a=0, b=0, c=0, d=nil, e=nil, f=nil> ``` This seems specific to moving, if I set `move: false` the struct makes it across OK. -- https://bugs.ruby-lang.org/

3 2

[ruby-core:119941] [Ruby master Bug#20897] resolv: not usable across ractors
by chucke (Tiago Cardoso) 09 May '25

09 May '25

Issue #20897 has been reported by chucke (Tiago Cardoso). ---------------------------------------- Bug #20897: resolv: not usable across ractors https://bugs.ruby-lang.org/issues/20897 * Author: chucke (Tiago Cardoso) * Status: Open * Backport: 3.1: UNKNOWN, 3.2: UNKNOWN, 3.3: UNKNOWN ---------------------------------------- Using the `resolv` library in a non-main ractor is impossible, due to DNS resources being memoized in a mutable class variable. [This patch](https://github.com/ruby/resolv/pull/62) fixes it by storing resource information in static variables. -- https://bugs.ruby-lang.org/

3 2

[ruby-core:112091] [Ruby master Bug#19387] Issue with ObjectSpace.each_objects not returning IO objects after starting a ractor
by luke-gru (Luke Gruber) 09 May '25

09 May '25

Issue #19387 has been reported by luke-gru (Luke Gruber). ---------------------------------------- Bug #19387: Issue with ObjectSpace.each_objects not returning IO objects after starting a ractor https://bugs.ruby-lang.org/issues/19387 * Author: luke-gru (Luke Gruber) * Status: Open * Priority: Normal * Backport: 2.7: UNKNOWN, 3.0: UNKNOWN, 3.1: UNKNOWN, 3.2: UNKNOWN ---------------------------------------- ```ruby r = Ractor.new do receive # block, the problem is not the termination of the ractor but the starting end ObjectSpace.each_object(IO) { |io| p io # we get no objects } ``` -- https://bugs.ruby-lang.org/

6 8

[ruby-core:112207] [Ruby master Bug#19408] Object no longer frozen after moved from a ractor
by luke-gru (Luke Gruber) 09 May '25

09 May '25

Issue #19408 has been reported by luke-gru (Luke Gruber). ---------------------------------------- Bug #19408: Object no longer frozen after moved from a ractor https://bugs.ruby-lang.org/issues/19408 * Author: luke-gru (Luke Gruber) * Status: Open * Priority: Normal * Backport: 2.7: UNKNOWN, 3.0: UNKNOWN, 3.1: UNKNOWN, 3.2: UNKNOWN ---------------------------------------- I think frozen objects should still be frozen after a move. ```ruby r = Ractor.new do obj = receive p obj.frozen? # should be true but is false p obj end obj = [Object.new].freeze r.send(obj, move: true) r.take ``` -- https://bugs.ruby-lang.org/

3 3

[ruby-core:120257] [Ruby master Bug#20957] RangeError on Array#values_at with negative ranges
by kyanagi (Kouhei Yanagita) 08 May '25

08 May '25

Issue #20957 has been reported by kyanagi (Kouhei Yanagita). ---------------------------------------- Bug #20957: RangeError on Array#values_at with negative ranges https://bugs.ruby-lang.org/issues/20957 * Author: kyanagi (Kouhei Yanagita) * Status: Open * ruby -v: ruby 3.4.0dev (2024-12-13T10:19:31Z master 3cb79d4082) +PRISM [arm64-darwin22] * Backport: 3.1: UNKNOWN, 3.2: UNKNOWN, 3.3: UNKNOWN ---------------------------------------- ``` [0, 1, 2, 3].values_at(10) #=> [nil] [0, 1, 2, 3].values_at(10..10) #=> [nil] [0, 1, 2, 3].values_at(-10) #=> [nil] [0, 1, 2, 3].values_at(-10..-10) #=> 'Array#values_at': -10..-10 out of range (RangeError) ``` Is this the intended behavior? I am aware that this behavior was recently added to the documentation, but I would like to confirm whether it is the intended behavior. -- https://bugs.ruby-lang.org/

3 2

[ruby-core:114025] [Ruby master Feature#19744] Namespace on read
by tagomoris (Satoshi TAGOMORI) 06 May '25

06 May '25

Issue #19744 has been reported by tagomoris (Satoshi TAGOMORI). ---------------------------------------- Feature #19744: Namespace on read https://bugs.ruby-lang.org/issues/19744 * Author: tagomoris (Satoshi TAGOMORI) * Status: Open * Priority: Normal ---------------------------------------- # What is the "Namespace on read" This proposes a new feature to define virtual top-level namespaces in Ruby. Those namespaces can require/load libraries (either .rb or native extension) separately from the global namespace. Dependencies of required/loaded libraries are also required/loaded in the namespace. ### Motivation The "namespace on read" can solve the 2 problems below, and can make a path to solve another problem: The details of those motivations are described in the below section ("Motivation details"). #### Avoiding name conflicts between libraries Applications can require two different libraries safely which use the same module name. #### Avoiding unexpected globally shared modules/objects Applications can make an independent/unshared module instance. #### (In the future) Multiple versions of gems can be required Application developers will have fewer version conflicts between gem dependencies if rubygems/bundler will support the namespace on read. ### Example code with this feature ```ruby # your_module module YourModule end # my_module.rb require 'your_module' module MyModule end # example.rb namespace1 = NameSpace.new namespace1.require('my_module') #=> true namespace1::MyModule #=> #<Module:0x00000001027ea650>::MyModule (or #<NameSpace:0x00...>::MyModule ?) namespace1::YourModule # similar to the above MyModule # NameError YourModule # NameError namespace2 = NameSpace.new # Any number of namespaces can be defined namespace2.require('my_module') # Different library "instance" from namespace1 require 'my_module' # require in the global namespace MyModule.object_id != namespace1::MyModule.object_id #=> true namespace1::MyModule.object_id != namespace2::MyModule.object_id ``` The required/loaded libraries will define different "instances" of modules/classes in those namespaces (just like the "wrapper" 2nd argument of `Kernel.load`). This doesn't introduce compatibility problems if all libraries use relative name resolution (without forced top-level reference like `::Name`). # "On read": optional, user-driven feature "On read" is a key thing of this feature. That means: * No changes are required in existing/new libraries (except for limited cases, described below) * No changes are required in applications if it doesn't need namespaces * Users can enable/use namespaces just for limited code in the whole library/application Users can start using this feature step by step (if they want it) without any big jumps. ## Motivation details This feature can solve multiple problems I have in writing/executing Ruby code. Those are from the 3 problems I mentioned above: name conflicts, globally shared modules, and library version conflicts between dependencies. I'll describe 4 scenarios about those problems. ### Running multiple applications on a Ruby process Modern computers have many CPU cores and large memory spaces. We sometimes want to have many separate applications (either micro-service architecture or modular monolith). Currently, running those applications require different processes. It requires additional computation costs (especially in developing those applications). If we have isolated namespaces and can load applications in those namespaces, we'll be able to run apps on a process, with less overhead. (I want to run many AWS Lambda applications on a process in isolated namespaces.) ### Running tests in isolated namespaces Tests that require external libraries need many hacks to: * require a library multiple times * require many different 3rd party libraries into isolated spaces (those may conflict with each other) Software with plugin systems (for example, Fluentd) will get benefit from namespaces. In addition to it, application tests can avoid unexpected side effects if tests are executed in isolated namespaces. ### Safely isolated library instances Libraries may have globally shared states. For example, [Oj](https://github.com/ohler55/oj) has a global `Obj.default_options` object to change the library behavior. Those options may be changed by any dependency libraries or applications, and it changes the behavior of `Oj` globally, unexpectedly. For such libraries, we'll be able to instantiate a safe library instance in an isolated namespace. ### Avoiding dependency hells Modern applications use many libraries, and those libraries require much more dependencies. Those dependencies will cause version conflicts very often. In such cases, application developers should resolve those by updating each libraries, or should just wait for the new release of libraries to conflict those libraries. Sometimes, library maintainers don't release updated versions, and application developers can do nothing. If namespaces can require/load a library multiple times, it also enables to require/load different versions of a library in a process. It requires the support of rubygems, but namespaces should be a good fundamental of it. ## Expected problems ### Use of top-level references In my expectation, `::Name` should refer the top-level `Name` in the global namespace. I expect that `::ENV` should contain the environment variables. But it may cause compatibility problems if library code uses `::MyLibrary` to refer themselves in their deeply nested library code. ### Additional memory consumption An extension library (dynamically linked library) may be loaded multiple times (by `dlopen` for temporarily copied dll files) to load isolated library "instances" if different namespaces require the same extension library. That consumes additional memory. In my opinion, additional memory consumption is a minimum cost to realize loading extension libraries multiple times without compatibility issues. This occurs only when programmers use namespaces. And it's only about libraries that are used in 2 or more namespaces. ### The change of `dlopen` flag about extension libraries To load an extension library multiple times without conflicting symbols, all extensions should stop sharing symbols globally. Libraries referring symbols from other extension libraries will have to change code & dependencies. (About the things about extension libraries, [Naruse also wrote an entry](https://naruse.hateblo.jp/entry/2023/05/22/193411).) # Misc The proof-of-concept branch is here: https://github.com/tagomoris/ruby/pull/1 It's still work-in-progress branch, especially for extension libraries. -- https://bugs.ruby-lang.org/

17 70

[ruby-core:116370] [Ruby master Bug#20203] `TestEnumerable` test failures with GCC 14
by vo.x (Vit Ondruch) 28 Apr '25

28 Apr '25

Issue #20203 has been reported by vo.x (Vit Ondruch). ---------------------------------------- Bug #20203: `TestEnumerable` test failures with GCC 14 https://bugs.ruby-lang.org/issues/20203 * Author: vo.x (Vit Ondruch) * Status: Open * Priority: Normal * ruby -v: ruby 3.3.0 (2023-12-25 revision 5124f9ac75) [x86_64-linux] * Backport: 3.0: UNKNOWN, 3.1: UNKNOWN, 3.2: UNKNOWN, 3.3: UNKNOWN ---------------------------------------- There is ongoing mass rebuild in Fedora and that is first time GCC 14 is used and we observe test failures in `TestEnumerable`. Here are a few examples: ~~~ [ 3000/26419] TestEnumerable#test_transient_heap_sort_bymalloc_consolidate(): unaligned fastbin chunk detected ~~~ ~~~ [ 2455/26535] TestEnumerable#test_transient_heap_sort_bycorrupted size vs. prev_size in fastbins ~~~ ~~~ [ 9716/26532] TestEnumerable#test_any_with_unused_blockdouble free or corruption (fasttop) ~~~ The full logs are accessible [here](https://koji.fedoraproject.org/koji/taskinfo?taskID=112176941). Please drill through `Descendants` and `build.log` -- https://bugs.ruby-lang.org/

5 12

[ruby-core:111526] [Ruby master Bug#19288] Ractor JSON parsing significantly slower than linear parsing
by maciej.mensfeld (Maciej Mensfeld) 18 Apr '25

18 Apr '25

Issue #19288 has been reported by maciej.mensfeld (Maciej Mensfeld). ---------------------------------------- Bug #19288: Ractor JSON parsing significantly slower than linear parsing https://bugs.ruby-lang.org/issues/19288 * Author: maciej.mensfeld (Maciej Mensfeld) * Status: Open * Priority: Normal * ruby -v: ruby 3.2.0 (2022-12-25 revision a528908271) [x86_64-linux] * Backport: 2.7: UNKNOWN, 3.0: UNKNOWN, 3.1: UNKNOWN, 3.2: UNKNOWN ---------------------------------------- a simple benchmark: ```ruby require 'json' require 'benchmark' CONCURRENT = 5 RACTORS = true ELEMENTS = 100_000 data = CONCURRENT.times.map do ELEMENTS.times.map do { rand => rand, rand => rand, rand => rand, rand => rand }.to_json end end ractors = CONCURRENT.times.map do Ractor.new do Ractor.receive.each { JSON.parse(_1) } end end result = Benchmark.measure do if RACTORS CONCURRENT.times do |i| ractors[i].send(data[i], move: false) end ractors.each(&:take) else # Linear without any threads data.each do |piece| piece.each { JSON.parse(_1) } end end end puts result ``` Gives following results on my 8 core machine: ```shell # without ractors: 2.731748 0.003993 2.735741 ( 2.736349) # with ractors 12.580452 5.089802 17.670254 ( 5.209755) ``` I would expect Ractors not to be two times slower on the CPU intense work. -- https://bugs.ruby-lang.org/

9 21