Activity
From 05/15/2026 to 05/21/2026
Today
- 02:55 PM Revision 945eab1d (git): ZJIT: Print module name in HIR Type display (#17039)
- Go from `ModuleExact[VALUE(0x1008)]` to `ModuleExact[Enumerable@0x1008]` if the module's name isn't `nil`.
-
12:47 PM Bug #22077: Endless methods in 3.3.5 allows for endless rescues. 4.0 shows different behaviour
- Seems caused by http://github.com/ruby/prism/commit/07202005cbd01e860edb7aa7c93e8fe7d74efcf5
-
09:44 AM Bug #22077 (Open): Endless methods in 3.3.5 allows for endless rescues. 4.0 shows different behaviour
ruby 3.3.5 (2024-09-03 revision ef084cc8f4) [x86_64-linux]
```
irb(main):001> def endless = nil * 1000 rescue raise('surprise') rescue 'please stop'
=> :endless
irb(main):002> endless
=> "please stop"
irb(main):003>
```
...-
11:26 AM Revision 81c68150 (git): [ruby/strscan] Fix call-seq return values
- (https://github.com/ruby/strscan/pull/207)
Some `call-seq` comments did not match the actual return values
(documentation only, no behavior change):
* `match?` returns the match size, not a position
* `skip` was missing the `->` arrow
... -
09:04 AM Bug #22076 (Closed): defined? returns nil for protected methods defined in a module even when callable
- Applied in changeset commit:git|d1391a000359655f948e16d7f2b0e85890082b5a.
----------
Fix defined? for protected methods defined in a module
me->defined_class is 0 for methods stored on a module, so the
protected visibility check in vm_... -
08:53 AM Bug #22076: defined? returns nil for protected methods defined in a module even when callable
- The expectation and the fix seem reasonable both.
-
- Filing this to confirm whether defined? should return "method" here (matching the call check), or whether the current nil behavior is intentional.
### Summary
`defined?` returns `nil` for a protected method defined in a module, eve... -
- me->defined_class is 0 for methods stored on a module, so the
protected visibility check in vm_defined always failed and defined?
returned nil even when the call would succeed.
Use rb_callable_method_entry_with_refinements (where define... -
07:19 AM Revision bf920ef6 (git): test_thread_join_hang: join leaked inner sleeper (#17068)
- SleepingUnblockScheduler#unblock deliberately breaks the join path,
so the inner Thread.new{sleep(0.01)} created inside Fiber.schedule
is abandoned mid-sleep when the outer thread exits. It eventually
wakes and lingers in Thread.list, br... -
06:53 AM Revision f562993d (git): Trim tarball-* reusable inputs
- Backport of ruby/ruby#17063 to ruby_4_0. All four reusables and the
make-snapshot composite are now blob-identical to master, including
tarball-windows since ruby_4_0 already runs the modern windows-2022 +
windows-2025-vs2026 matrix. tar... -
- The composite is blob-identical across branches after PR #17063 lands
everywhere, so pointing maintenance branches at master removes the
need to re-cherry-pick future fixes. The local copy stays in place
because ruby/actions still calls ... -
- Backport of ruby/ruby#17063 to ruby_3_4. The reusables and the
make-snapshot composite are now blob-identical to master except for
tarball-windows, which keeps the windows-2022 + vs / vcvars matrix and
the manual vcvars64.bat setup env s... -
- The composite is blob-identical across branches after PR #17063 lands
everywhere, so pointing maintenance branches at master removes the
need to re-cherry-pick future fixes. The local copy stays in place
because ruby/actions still calls ... -
- Backport of ruby/ruby#17063 to ruby_3_3. The reusables and the
make-snapshot composite are now blob-identical to master except for
tarball-windows, which keeps the windows-2022 + vs / vcvars matrix and
the manual vcvars64.bat setup env s... -
-
-
- The bin/gem and bin/bundle scripts in the source tree do not run
cleanly without an install step on Windows. Keep ruby -v only. -
- Without the workaround the snapshot-tarball run leaves
~/.gnupg/{crls.d,private-keys-v1.d,public-keys.d} behind after
test_sync_default_gems, which trips Diff stats of HOME. GPG 2.x's
agent processes ignore GNUPGHOME for these directorie... -
- The inputs context is only populated for workflow_dispatch and
workflow_call, so referencing it on push, pull_request, or merge_group
relies on undefined behaviour. Short-circuit on github.event_name so
non-dispatch runs always pass fals... -
- Lets callers opt out of the Packages and Info uploads. tarball-test
keeps the default true so per-OS reusables can download them, while
upload-tarball in ruby/actions can pass false once it runs the same
composite in a matrix across four... -
- SIMPLER_ALERTS_URL always fires on failure so developer-facing CI
(PR, push, merge_group, manual dispatch) surfaces breakage in the
ruby-core channel. notify-release-channel additionally routes to
SNAPSHOT_SLACK_WEBHOOK_URL for the daily... -
- Match the smoke test added to tarball-ubuntu. macos checks the
installed /usr/local/bin/{ruby,gem,bundle}; windows runs the
build-tree ruby against bin/{gem,bundle} from the extracted tarball
because the workflow has no install step. -
- ruby_3_3 was the only caller that disabled the extra mkdir step.
Always create $HOME/.local/share and $HOME/.ssh in the main Fixed
world writable dirs step. mkdir -p $HOME/.local/share also covers
$HOME/.local, so the bare mkdir for it g... -
- Confirms the snapshot tarball ships a working RubyGems and Bundler in
addition to the ruby binary. -
- Hardcode the post-install smoke test to /usr/local/bin/ruby, matching
the configure default prefix used on master/4_0/3_4. ruby_3_3's `ruby`
override falls away when it aligns with the same install layout. -
- The three apt-mode variants existed to selectively uninstall git (and
sometimes ruby) per test_task. ruby-and-git on ruby_3_3 also pinned an
apt-installed host ruby. With ruby/setup-ruby covering the host ruby
unconditionally and git alw... -
- The Homebrew Ruby uninstall/reinstall dance was only enabled on
ruby_3_3 to work around macos-15-intel quirks, and the reinstall step
never fired because macos-15-intel only runs the check task. Drop both
steps and the input. -
- No caller passes patch-url: tarball-test.yml never set it, and
ruby/actions draft-release.yml stopped forwarding it once it started
calling these workflows by ref. Remove the input and the Apply patch
step from the ubuntu, macos, and win... -
- The step force-removed $HOME/.gnupg between Tests and Diff stats of
HOME, but the original cause for the directory persisting was never
identified. Remove it and observe whether the HOME diff still passes. -
- power_assert was allowed to fail on master/4_0 by passing it through
this input. The allow-list belongs in tool/test-bundled-gems.rb now
that the tool ships in the same repo as the workflow. -
- Always install Ruby 3.2 via ruby/setup-ruby for test-bundled-gems.
ruby_3_3 will follow once it stops relying on the apt-installed ruby
for its host.
05/20/2026
-
11:48 PM Revision eeed9ea3 (git): Move rb_gc_update_set_refs into gc.c
- The code is in gc/gc.h but it is not used outside of gc.c, so we can move
it there. -
10:20 PM Revision 63c35b40 (git): [DOC] Doc for Pathname#descend
-
-
-
09:34 PM Bug #22074: YJIT misaligns locals when there are > 256 local variables
- Link to the PR: https://github.com/ruby/ruby/pull/17043
-
- That's great!
-
- I migrated snapshot package tests from ruby/actions to ruby/ruby repository.
Now we test `make install` every pull reqeust with Linux, macOS and Windows.
https://github.com/ruby/ruby/blob/master/.github/workflows/tarball-ubuntu.yml... -
- Concretely allowing Strings means extra checks for unknown or missing members, with only Symbols a `kwargs.size == CLASS_MEMBERS.size` check would be enough.
It also leads to ambiguous cases like:
```ruby
MyData = Data.define(:value)
MyD... -
09:14 PM Revision 3002cea4 (git): Fix GC Bug in imemo cdhash creation
- It's possible for st_init_existing_table_with_size to trigger GC. If
that happens we need to ensure that the table entries count doesn't
contain garbage data, or we'll try and mark random stuff -
-
- Snapshotting at start of marking lets sweep-time frees count against the
next epoch, which roughly halves GC frequency on alloc-heavy workloads.
Move the snapshot to end of sweep so the next epoch starts from a clean
baseline. -
-
-
- Several SETs in rb_gc_impl_stat may allocate a T_BIGNUM RVALUE when
the value exceeds FIXNUM_MAX
This is invisible on LP64 but trips on LLP64 Windows and ILP32 Linux
where FIXNUM_MAX ~= 1.07GB.
If those allocations happen *after* setti... -
-
- Replace the single
objspace->malloc_counters.{increase,oldmalloc_increase} size_t fields
with pairs of monotonically-increasing counters. Snapshots of these
counters are taken at each GC, so that the live malloc_increase is computed as
... -
05:20 PM Bug #22075: heap-use-after-free in `rb_vm_ci_lookup` under parallel Ractors
- 4.0 backport: https://github.com/ruby/ruby/pull/17055
-
- Applied in changeset commit:git|cb9f2ba8046b10de6b4fa0ff364a8e9d09235d00.
----------
Use atomics for kwargs reference count
Fixes [Bug #22075] -
- Thank you for the reliable reproduction. I believe the issue is that `super` calls are duplicating callinfos, but the reference count for the kwargs array isn't using atomics.
```ruby
class Base
def foo(a:, b:, c:) = a
end
cla... -
- ## Environment
- `ruby 4.0.5 (2026-05-20 revision 64336ffd0e) +PRISM [aarch64-linux]`
- Also reproduced on `ruby 4.0.4 ... [x86_64-linux]`.
## Summary
Under a parallel Ractor pool, a GC sweep on one Ractor frees a `T_DATA` obje... -
- Fixes [Bug #22075]
- 04:25 PM Revision c124d555 (git): ZJIT: Profile all args on stack, including blockarg (#17038)
- No need to special-case this in the HIR builder: just profile arguments as they actually come in on the stack.
Don't special case the send opcode with `stack_offset` and stuff. We can handle this in the profiler. -
- We made a mistake calculating slot sizes during the heap slot sizes
refactor. Previously BASE_SLOT_SIZE included RVALUE_OVERHEAD, this was
lost during the refactor to use the SLOT macro.
The result of this was that when Ruby was compile... -
- zizmor flags `cd "${{ inputs.archname }}/"` inside `run:` blocks as
code injection via template expansion: GitHub Actions substitutes the
input value into the shell script verbatim, so a caller passing shell
metacharacters could execute ... -
- The per-OS tarball-test workflows currently hard-code `cd snapshot-*/`
to enter the extracted tarball directory, which prevents callers from
producing tarballs under any other prefix (for example `ruby-3.4.0-rc1`
emitted by `tool/make-sn... -
- zizmor flags `cd "${{ inputs.archname }}/"` inside `run:` blocks as
code injection via template expansion: GitHub Actions substitutes the
input value into the shell script verbatim, so a caller passing shell
metacharacters could execute ... -
- The per-OS tarball-test workflows currently hard-code `cd snapshot-*/`
to enter the extracted tarball directory, which prevents callers from
producing tarballs under any other prefix (for example `ruby-3.4.0-rc1`
emitted by `tool/make-sn... -
- zizmor flags `cd "${{ inputs.archname }}/"` inside `run:` blocks as
code injection via template expansion: GitHub Actions substitutes the
input value into the shell script verbatim, so a caller passing shell
metacharacters could execute ... -
- The per-OS tarball-test workflows currently hard-code `cd snapshot-*/`
to enter the extracted tarball directory, which prevents callers from
producing tarballs under any other prefix (for example `ruby-3.4.0-rc1`
emitted by `tool/make-sn... -
- zizmor flags `cd "${{ inputs.archname }}/"` inside `run:` blocks as
code injection via template expansion: GitHub Actions substitutes the
input value into the shell script verbatim, so a caller passing shell
metacharacters could execute ... -
- The per-OS tarball-test workflows currently hard-code `cd snapshot-*/`
to enter the extracted tarball directory, which prevents callers from
producing tarballs under any other prefix (for example `ruby-3.4.0-rc1`
emitted by `tool/make-sn... -
07:13 AM Revision 27186cc2 (git): [DOC] Add missing colon in IO#seek doc
-
- The legacy mode (vs2022 + vcvars 14.2) is no longer exercised, so
collapse the conditional matrix into the modern 2022 / 2025-vs2026
pair and remove the now-dead setup-env (legacy) step. -
-
- ruby_4_0 only builds with vssetup.cmd against windows-2022 and
windows-2025-vs2026, so the mode input and the legacy branch are
dead code on this branch. Drop the input, hardcode the modern
matrix and setup-env step, and stop passing mod... -
- GitHub Actions schedule triggers only fire on the default branch, so
the daily cron in tarball-test.yml never ran on ruby_4_0. Switch the
trigger to workflow_dispatch so master's tarball-test-schedule.yml
can dispatch it via gh workflow ... -
- ruby_3_4 only builds with the vs2022 + vcvars 14.2 toolchain on this
branch because win32/vssetup.cmd has not been backported, so the mode
input and the modern branch are dead code here. Drop the input,
hardcode the legacy matrix and set... -
- GitHub Actions schedule triggers only fire on the default branch, so
the daily cron in tarball-test.yml never ran on ruby_3_4. Switch the
trigger to workflow_dispatch so master's tarball-test-schedule.yml
can dispatch it via gh workflow ... -
05:33 AM Revision c13fe5aa (git): Remove unused generate-tar-bz2
- Release packages stopped to include `*.tar.bz2` since 3.0.0.
https://bugs.ruby-lang.org/issues/16483 -
- ruby_3_3 only builds with the vs2022 + vcvars 14.2 toolchain, so the
mode input and the modern branch are dead code on this branch. Drop
the input, hardcode the legacy matrix and setup-env step, and stop
passing mode: legacy from tarball... -
- Release packages stopped to include `*.tar.bz2` since 3.0.0.
https://bugs.ruby-lang.org/issues/16483 -
- GitHub Actions schedule triggers only fire on the default branch, so
the daily cron in tarball-test.yml never ran on ruby_3_3. Switch the
trigger to workflow_dispatch so master's tarball-test-schedule.yml
can dispatch it via gh workflow ... -
05:10 AM Bug #22058: {Method,InstanceMethod}#super_method doesn't work correctly for refined method with refinements for method active in the caller's scope
- jeremyevans0 (Jeremy Evans) wrote in #note-13:
> ```ruby
> ...
Initially I thought the pre-2.7 behavior was more reasonable, but I now agree that the 2.7+ behavior is more consistent.
I'd like to confirm this with Matz at the next dev... -
12:19 AM Bug #22058: {Method,InstanceMethod}#super_method doesn't work correctly for refined method with refinements for method active in the caller's scope
- I merged commit:cb0cd76a086d2d9423adaa4c5032a5fa149e7ef1 and commit:8f8dd01a93b247f607d927770a2ad09e1e0a1647 to handle address the issues identified in this bug report. I didn't have them auto-close this report, because there is still a ...
-
- GitHub Actions schedule triggers only fire from the default branch, so
the daily run defined in tarball-test.yml never executed on ruby_3_3,
ruby_3_4, or ruby_4_0. Move the cron into a new dispatcher workflow
that calls workflow_dispatch... -
03:49 AM Revision bc048b50 (git): configure: check some warning options only with clang
- GCC doesn't emit any warning when the -Wno- form is used, so we can't
check reliably if the warning option is available:
When an unrecognized warning option is requested (e.g.,
-Wunknown-warning), GCC gives an error stating that the opt... - 03:32 AM Revision 4e3e8062 (git): [DOC] Update bundled gems list at 7f76cbe86597280313551bfa9609a8
-
- This reverts commit 3d73cd752f5dbdff4c77c57d011ca7f8d783c8ba, that is
no longer needed since ruby/rubygems#8735. -
- gc/mmtk/mmtk.c:480:1: warning: function 'rb_mmtk_gc_thread_bug' could be declared with attribute 'noreturn' [-Wmissing-noreturn]
480 | {
| ^
gc/mmtk/mmtk.c:510:1: warning: function 'rb_mmtk_mutator_thread_panic_handle... -
12:13 AM Revision 8601f5a4 (git): Fix reading freed memory in rb_getaddrinfo
-
12:06 AM Revision 8f8dd01a (git): Correctly handle refinements in Method#super_method for bmethods
- Bmethods need to use a different approach to find the cref to use
in order to determine the refinements in effect. -
- A `Method#super_method` chain should return the methods that `super`
would call if the method was called normally. Previously, there were
multiple problems:
* There was an infinite `super_method` loop for refined methods.
* Lookup cons...
05/19/2026
-
11:45 PM Bug #22074 (Open): YJIT misaligns locals when there are > 256 local variables
- Hi! We ran into this issue in production with a large Slim template in a Rails app, but we found that it's reproducible in any function with a lot of local variables. It seems like YJIT tracks local variables for register allocation in a...
-
11:23 PM Revision c7a7bc98 (git): Merge branch 'ruby_4_0' into v4.0.5
-
-
-
-
-
09:57 PM Revision 3373fcc2 (git): Fix UAF in IO::Buffer#~ when self is an invalidated slice
- `io_buffer_not` accessed `buffer->base` directly without validating that the buffer was still live. A slice whose parent had been freed retained its stale base pointer, so calling `~` on it caused a UAF.
Use `io_buffer_get_bytes_for_rea... -
- This sounds like a serious eCryptfs bug to me, could you report it there? (or is there already an open issue about that? If so please link it here).
(IMO eCryptfs is kind of a hack leaking many implementation details, nevertheless it'... -
- They are implementation details:
* https://docs.ruby-lang.org/en/4.0/Enumerator/Yielder.html
* https://docs.ruby-lang.org/en/4.0/Enumerator/Producer.html
* https://docs.ruby-lang.org/en/4.0/Enumerator/Generator.html -
08:26 PM Revision 8267e508 (git): ZJIT: x64: Prefer 7-byte sign extending `mov` over 10-byte `movabs`
- Relevant for small negative immediates. Previously:
# Insn: v16 SetLocal l1, EP@3, v10
mov rsi, qword ptr [r13 + 0x20]
mov rsi, qword ptr [rsi - 8]
and rsi, 0xfffffffffffffffc
# call rb_vm_env_write
push rdi
push rdi
mov... -
- Check that these two methods opposite always.
-
- 10:27 AM Revision 6c140117 (git): [ppc64le] Add ELFv2 global entry prologue to coroutine_transfer
- The hand-written `coroutine_transfer` in `coroutine/ppc64le/Context.S`
lacks the ELFv2 ABI global entry prologue. When the function is reached
via the PLT (lazy resolution path), r2 (TOC pointer) is left pointing
at whatever the resolver... -
- `rb_load_entrypoint` is used only by `rb_box_load`; it checks that the
argument has two-elements (implicitly assuming it is an array),
extracts its contents, and then simply discards the argument. -
- 09:40 AM Revision 0f7f4278 (git): Make sync script work correctly with Windows-style newlines
- I'm almost certain nobody is actually running this script on Windows,
but the tests for it do run during `nmake check`, and they fail at least
on my git configuration.
The $ anchor doesn't match \r\n with git's -E regex matching, so we ... -
- Port the daily snapshot tarball pipeline from ruby/actions into
ruby/ruby so the tarball build and per-OS tests run on every push and
pull request. The make-snapshot composite action gains a srcdir input so
the same logic can either clon... -
- Port the daily snapshot tarball pipeline from ruby/actions into
ruby/ruby so the tarball build and per-OS tests run on every push and
pull request. The make-snapshot composite action gains a srcdir input so
the same logic can either clon... -
- Port the daily snapshot tarball pipeline from ruby/actions into
ruby/ruby so the tarball build and per-OS tests run on every push and
pull request. The make-snapshot composite action gains a srcdir input so
the same logic can either clon... -
- Promote test-bundler-parallel from a macos-14 include entry into the
base test_task list so macos-15 picks it up too, and limit
macos-15-intel to a single check job via include. The Intel runner is
slower and noisier; running only check ... -
- Cherry-picking tarball-test.yml to maintenance branches previously
required editing six places (four archname literals, branch-label, and
the Materialize step). Route everything through env.BRANCH on the
tarball job and expose it via out... -
- The composite action previously read $archname from the caller's job
env, leaving the dependency implicit. Declare archname as a required
input and pass it via with: so the action is self-contained and the
linkage is visible at the call ... -
- Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
-
- Add a daily 18:30 UTC schedule trigger to match the existing ruby/actions
snapshot cron, and limit the Slack failure notifications in the reusable
workflows to schedule runs. PRs and merge_group runs are still gated by
the workflow's CI ... -
- zizmor flags inline `${{ inputs.X }}` expansions inside run scripts as
template-injection errors. Reusable workflow inputs come from the
in-repo caller and are trusted, but routing them through env avoids the
class of mistake entirely an... -
- tool/make-snapshot reads HEAD via tool/lib/vcs.rb#_get_revisions and,
on detached checkouts (pull_request, merge_group), parses
"(HEAD detached at <ref>)" into branch name "pull/N/merge" which then
breaks `git clone -b` in the export ste... -
- Port the daily snapshot tarball pipeline from ruby/actions into
ruby/ruby so the tarball build and per-OS tests run on every push and
pull request. The make-snapshot composite action gains a srcdir input so
the same logic can either clon... -
-
-
- Stop using uutils-coreutils that fails to install on the arm64-windows
runner. - 07:39 AM Revision 399727de (git): Fix duplicate assertion in test_metaclass
- The final assertion of test_metaclass was identical to the one on
line 161. Correcting to metametaclass_method_b. -
- (https://github.com/ruby/strscan/pull/206)
Came here to fix a broken link (second pattern generated a bogus link);
stayed to add examples and generally brighten.
https://github.com/ruby/strscan/commit/32cc23934c -
-
- Fix for GH-15898.
-
- `io_buffer_xor` accessed `buffer->base` and `mask_buffer->base` directly without validating that the buffers were still live. A slice whose parent had been freed retained its stale base pointer, so calling `^` on it caused a UAF.
Use `i... -
- This reverts commit 50c5b9a19f4c2851d1ae766b6728ca1ab52d6701.
Commit miss. -
05/18/2026
-
- Thanks for reporting back!
-
08:36 PM Bug #22073: Ruby 3.4.9 segfaults when loading native default gems compiled on Ubuntu 26.04 (GCC 15.2); Ruby 4.0.3 unaffected on same system
- Update / Resolution:
After further investigation, I was able to identify the root cause. The issue was not with Ruby 3.4.9, GCC 15.2, or Ubuntu 26.04 itself.
My home directory is mounted with eCryptfs (Ubuntu's home directory encry... -
- Seems the Ubuntu build system is broken after upgrading. I tried to compile Neovim from sources and it segfaults too. Testing on the fresh VM, it works.
-
- I had an idea here and installed a fresh Ubuntu 26.04 copy on a new VM, and it worked!
```
taq@taq2604:~$ rvm use 3.4.9
Using /home/taq/.rvm/gems/ruby-3.4.9
taq@taq2604:~$ ruby -v
ruby 3.4.9 (2026-03-11 revision 76cca827ab) +PRI... -
- Check this out:
`❯ rvm use ruby-4.0.4
Using /home/taq/.rvm/gems/ruby-4.0.4
❯ irb
4.0.4 :001 > exit
❯ gem install bundler
Fetching bundler-4.0.11.gem
Successfully installed bundler-4.0.11
Parsing documentation for bundler-4.... -
- Uh-oh ... tried to install 4.0.4 and ... same behaviour.
-
- Hello, I'm reporting a reproducible issue on Ruby 3.4.9 on a freshly upgraded Ubuntu 26.04 LTS. Please let me know if you'd prefer this as a comment on #21655 instead.
On Ubuntu 26.04 LTS, Ruby 3.4.9 produces unusable binaries when co... -
- There is no such thing:
```rb
irb(main):001> sprintf("%F", 123)
(irb):1:in 'Kernel#sprintf': malformed format string - %F (ArgumentError)
``` -
- @nobu Was it intentional to keep accepting `String` and `to_str` kwargs in addition to Symbols?
From what matz said above I would expect only Symbols are accepted, which would be valuable to optimize `Data#initialize` better. - 05:01 PM Revision 4351ee96 (git): ZJIT: StringExact/NilClass for GetSpecialSymbol and GetSpecialNumber (#17012)
- According to `rb_reg_nth_match`, `rb_reg_match_pre(post/last)`,
`GetSpecialSymbol` and `GetSpecialNumber` always returns `StringExact` or `NilClass`.
Fix `infer_type` to `StringExact|NilClass` instead of `BasicObject`. -
-
- When a single #[test] function holds multiple assert_snapshot! calls, only
the first one is updated per cargo-insta run, so updating all of them takes
multiple zjit-test -> zjit-test-update cycles. Splitting into one snapshot
per test le... -
-
- SIZED_REALLOC_N already assigns to its first argument, and the double
assignment causes two writes to the same struct without a sequence
point, which is undefined behavior.
gc.c: In function 'rb_gc_register_address':
gc.c:3758:3... -
- fafb55877aaf34592278eb3ef9ba3f61473d0a56 fixed issues with
refinement zsuper methods by converting them to a cfunc that calls
super. This changes Method creation to recognize the cfunc used
and skip those methods just as regular zsuper m... -
- Followup: https://github.com/ruby/ruby/pull/17016
`cdhash` is no longer a RHash, also its values are no
longer boxed. -
- Now that they're no longer a RHash instance, we don't
have to box the offsets, we can directly srore the raw
values, and stop marking them. -
- RHash isn't a good fit for storing `cdhash` as this force to allow
arbitrary hash types into RHash, which doesn't work with AR tables.
It also cause the cdhash to be larger than needed. -
-
- https://github.com/ruby/ruby/pull/16907 has been merged.
-
- `io_buffer_or` accessed `buffer->base` and `mask_buffer->base` directly without validating that the buffers were still live. A slice whose parent had been freed retained its stale base pointer, so calling `|` on it caused a UAF.
Use `io... -
- Switch examples to use `>=2` instead of `~>2.0.0` (which is really
terrible as you are limited to 2.0.x versions).
For the help output, show a simpler example with a single version
restriction, before an example with multiple version re...
05/17/2026
-
- Same optimization that was applied to `set_table` in
85c52079aa35a1d2e063a5b40eebe91701c8cb9e
This saves one pointer per `st_table`.
Now that GC support 64B slots, if we can save another 8B
either in `st_table` or in `struct RHash`, has... -
- https://github.com/ruby/mmtk/commit/40053354bd
- 06:09 PM Revision 84df2383 (git): * remove trailing spaces. [ci skip]
-
-
- Instead of a linked list. Should be both more compact and
faster to iterate.
Also fix reporting its memory size. -
- That API made sense before b00f280d4b9569e7153365d7e1c522b3d6b3c6cf
as finding a module name was costly, but it now is a very cheap
operation, and that function is now just an outdated alias. -
- Applied in changeset commit:git|6bc1aa31ccf39837bb0678da093b82f86a6534eb.
----------
[Bug #22072] Fix rb_cvar_set for multi-Ractor
rb_cvar_set breaks the RCLASS_CVC_TBL because it creates a new cvc_tbl but
forgets to copy the contents ... -
- Thank you for this bug report. Using the seed in the crashing runs you have linked, I was able to reliably reproduce the issue. I was able to [create a minimal reproduction and have a fix in this PR](https://github.com/ruby/ruby/pull/170...
-
- rb_cvar_set breaks the RCLASS_CVC_TBL because it creates a new cvc_tbl but
forgets to copy the contents over. This causes the following script to
crash:
r = Ractor.new {}
module Foo
def self.foo = @@foo
end
Foo.c... -
11:32 AM Revision c17d5b05 (git): Extract bundled-gem C extensions before configuring exts (#16973)
-
05:05 AM Feature #21857: Introduce to_proc pattern
- I understand the motivation, but I'm still not fully convinced that this particular use case is enough to justify introducing dedicated syntax.
There are similar requests as well, such as wanting to match based on `respond_to?`. It migh... -
-
03:39 AM Revision 716be73e (git): Add source_ref_or_sha
-
-
- Stop using uutils-coreutils that fails to install on the arm64-windows
runner. -
- - Skip Windows 11 ARM CI for now
commit:ba43e07c12093273e46630c55e93c7491dbbd372.
- Unlock uutils-coreutils version
commit:ded07a7d929dea4457e1c5c1466f34e7f5a8cd46.
05/16/2026
-
-
-
- Support for this was removed.
-
- It's failing with:
Get-Command: C:\Users\runneradmin\scoop\apps\scoop\current\lib\install.ps1:191
Line |
191 | $bin = (Get-Command $target).Source
| ~~~~~~~~~~~~~~~~~~~
| The... -
- Version 0.5.0 is failing on CI with the following error. Maybe unlocking
the version will fix it?
Get-Command: C:\Users\runneradmin\scoop\apps\scoop\current\lib\install.ps1:191
Line |
191 | $bin = (Get-Command $... -
- I believe this can be closed now that https://bugs.ruby-lang.org/issues/20064 was implemented
-
-
-
-
-
-
-
- Applied in changeset commit:git|9736b7ab56c87b589b8c8a638a019b6bbb19b127.
----------
Prohibit super in refined module method
Before, this super could result in a call to a method in BasicObject
(or a module included in or prepended to ... -
- shugo (Shugo Maeda) wrote in #note-4:
> It seems reasonable except that `RCLASS_INCLUDER(me->defined_class)` is better than `((struct RClass_and_rb_classext_t*)me->defined_class)->classext.as.iclass.includer`.
Thank you. I updated th... -
- jeremyevans0 (Jeremy Evans) wrote in #note-3:
> ```diff
> ...
Thank you.
It seems reasonable except that `RCLASS_INCLUDER(me->defined_class)` is better than `((struct RClass_and_rb_classext_t*)me->defined_class)->classext.as.iclass.in... -
- Before, this super could result in a call to a method in BasicObject
(or a module included in or prepended to BasicObject), or a call
to method_missing for the receiver's class.
This explicitly disallows the use of super in such a case,... -
- 02:09 PM Revision 4c0b5717 (git): ZJIT: Remove dead SideExitReasons
-
-
-
- It took a lot of work, but I was able to fix `Method#super_method` to return the same method that `super` would call when refinements are used: https://github.com/ruby/ruby/pull/16997
I'm not positive it handles every case correctly, ... -
- jeremyevans0 (Jeremy Evans) wrote in #note-10:
> @shugo Thank you, I think I finally understand. After the first refinement method is called, the caller's scope no longer matters, because each `super` call site considers the refinements... -
- Follow up of commit d5c5fcb80a432e2078139c460230dc7952216c35.
`target` may be the basename only, skip all rules under bundled gem. -
-
02:53 AM Bug #21789 (Closed): IO#flush doesn't invoke fiber scheduler io_write hook
- The linked PR was merged.
-
-
-
-
-
- This reverts commit bf01f6ae89a95d8f5572e050facfe311c8c28aaf, that was
an accidental commit by mistake. -
- https://github.com/ruby/mmtk/commit/9772a1d30a
-
01:25 AM Bug #22070: `Thread.each_caller_location(1, 1)` segfaults when called from a cfunc
- Thank you for creating PRs. Merged into `ruby_3_4` at https://github.com/ruby/ruby/commit/0dcf36db7b22b0eac26281cb7b9b0f1f1b85f374
- 01:25 AM Revision 0dcf36db (git): [Bug #22070] Fix segfault in `Thread.each_caller_location`
-
- Thank you for creating PR. merged into r`ruby_3_4` at https://github.com/ruby/ruby/commit/b1420a81f92197844ae91fb7e85f1e0fce6f6433
-
01:23 AM Revision b1420a81 (git): Merge openssl-3.3.3
- The changes can be found at:
https://github.com/ruby/openssl/compare/v3.3.2...v3.3.3 -
- The changes can be found at:
https://github.com/ruby/openssl/compare/v3.3.1...v3.3.2 -
- This removes the discussion of included modules in refinements,
since support for that was removed in Ruby 3.2.
This switches the method lookup description to be ancestor-based
instead of class-based, so it better reflects how the proce...
05/15/2026
-
- 09:10 PM Revision 9bd9f262 (git): [DOC] Fix typos and wording in Time RDoc (#16986)
-
- @shugo Thank you, I think I finally understand. After the first refinement method is called, the caller's scope no longer matters, because each `super` call site considers the refinements in effect at that `super` call site. You cannot u...
-
- jeremyevans0 (Jeremy Evans) wrote in #note-8:
> > > I think option 2 makes more sense, but it requires that Method/InstanceMethod objects for refined methods keep a reference to the scope in which they were created
> ...
`Kernel#method... -
- * ZJIT: Share a single JITFrame across all C method frames
Replace the per-call JITFrame::new_cfunc allocation in gen_push_frame
with a sentinel value (ZJIT_JIT_RETURN_C_FRAME) stored in cfp->jit_return.
CFP_ZJIT_FRAME now returns a poi... -
- shebang that includes "ruby"
(https://github.com/ruby/prism/pull/4110)
* Clarify the documentation of -x, it always searches a shebang that includes "ruby"
* Update one more occurence
---------
https://github.com/ruby/prism/commit/0c... -
- shugo (Shugo Maeda) wrote in #note-1:
> Refinements for modules were initially prohibited, and were allowed in [Feature #12534] while `super` was still prohibited. Supporting `super` in a refinement for a module would go beyond a bug fi... -
-
- Refinements for modules were initially prohibited, and were allowed in [Feature #12534] while `super` was still prohibited. Supporting `super` in a refinement for a module would go beyond a bug fix, so I think it should be discussed as a...
-
05:26 PM Feature #22068: Adding post-quantum cryptography (PQC) support across Ruby standard libraries
- hsbt (Hiroshi SHIBATA) wrote in #note-4:
> Thanks for your detailed summary. I will handle rubygems and bundler.
Thanks! -
- > As I heard from my colleagues working for OpenSSL, that SLH-DSA is not ready to be used, we can ignore SLH-DSA for now.
Sorry, the above was wrong or outdated information.
SLH-DSA is ready to use for signature use case (but not for... -
- Thanks for your detailed summary. I will handle rubygems and bundler.
-
04:58 PM Bug #22059: make -j problem in dtrace build
- I've committed the attached patch to pkgsrc, which so far seems to fix the problem for me.
-
02:22 PM Bug #22072 (Closed): [BUG] should have cvar cache entry
Hi! I have a CI build for my project that runs on Ruby 4.0.4, and it has recently started _sometimes_ segfaulting with this error: "".
I have seen failures in two different spots:
- In the middle of tests:
```
/home/runne...-
- Followup for https://bugs.ruby-lang.org/issues/20409
-
- The implementation is still the same as of today:
https://github.com/ruby/ruby/blob/4d87d43b01dbb312eb1ff5fbbc6c9f33218d91a2/thread.c#L2100-L2115
Meanwhile, `ruby_thread_has_gvl_p()` became a public API since #20877 ([commit](https:/... -
- This is currently blocking Ruby's doc generation:
https://github.com/ruby/actions/actions/runs/25895616458/job/76107951475
We'll need to make rbs work for the process so we can adopt the rbs-integrated
RDoc. But for now let's drop it to... -
- test/ruby/test_box.rb: extend timeout for slow Windows CI
The default 10-second assert_separately timeout is too tight for
test_calling_root_box_methods_does_not_change_user_boxes_newly_created
on Windows CI, where the subprocess comple... - 02:27 AM Revision b5038ac1 (git): Bump taiki-e/install-action
- Bumps the github-actions group with 1 update in the / directory: [taiki-e/install-action](https://github.com/taiki-e/install-action).
Updates `taiki-e/install-action` from 2.77.7 to 2.78.0
- [Release notes](https://github.com/taiki-e/i... -
-