Activity
From 12/02/2025 to 12/08/2025
Today
-
03:11 AM Ruby Revision a82aa08f (git): Make `ruby_reset_leap_second_info` internal
- It is exported only for the extension library to test, but the method
is no longer used since 29e31e72fb5a14194a78ec974c4ba56c33ad8d45. -
- It is used only in hash.c, when `ENV['TZ']` is set.
- 03:07 AM Ruby Revision fbc5bb91 (git): Update default gems list at 4655b174d5fa71b69781c56701be63 [ci skip]
-
03:06 AM Ruby Revision 4655b174 (git): [ruby/timeout] v0.5.0
- https://github.com/ruby/timeout/commit/837d5aac73
-
03:04 AM Ruby Feature #21767 (Open): Consider procs which `self` is Ractor-shareable as Ractor shareable
- I would like to allow procs which `self` is Ractor-shareable to be automatically eligible as shareable, without an explicit `Ractor.make_shareable` call.
```ruby
class C
PROC = proc { p ARRAY }.freeze
end
Ractor.new { C::PROC.... -
02:47 AM Ruby Revision be882278 (git): Move RBIMPL_ATTR_DEPRECATED_* macros to the appropriate header file
- Move these macros from include/ruby/backward.h to
include/ruby/internal/attr/deprecated.h, alongside the other similar
macros.
include/ruby/internal/intern/vm.h cannot currently use them because
include/ruby/backward.h is included too l... -
- The #ifdef is currently not taken because include/ruby/backward.h is
not included at this point. The attribute is unnecessary in an internal
header, so remove it. -
01:20 AM Ruby Feature #21766: Pathname + FileUtils making sweet music together
- I'm open to any and all implementation suggestions, happy to do whatever is required. I just want the feature!
12/07/2025
-
11:07 PM Ruby Misc #21154: Document or change Module#autoload?
- BTW, I do not know if you know it, but while I started working on Zeitwerk to fix autoloading in Rails, I changed my mind in the mid of it and decided to do something more generic that supports multiple autoloaders in the same process ma...
-
- Thanks for the example @mame.
In my view, this is more a problem of circularity than a problem of autoload vs require. When you are loading a file, a require is executed and loads a file that needs a constant that was still not defined.... -
09:57 PM Ruby Feature #21766: Pathname + FileUtils making sweet music together
- Remember that `FileUtils` is loaded dynamically, so it would be something like
```ruby
class Pathname
def mkdir_p(...) = _fileutils(:mkdir_p, ...)
def ln(...) = _fileutils(:ln, ...)
def ln_s(...) = _fileutils(:ln_s, ...)
... -
- I love Pathname. I love FileUtils. Let's bring these two classes EVEN CLOSER TOGETHER by adding some tragically missing helpers. Something like this, perhaps?
``` ruby
class Pathname
def mkdir_p(...) = FileUtils.mkdir_p(@path, ...... -
04:51 PM Ruby Revision 4f900c35 (git): Output ivar length for T_OBJECT in obj_info
- 04:35 PM Ruby Revision a4d14213 (git): [DOC] Update bundled gems list at 379d22ce8418448ade3d410e5c76dd
-
04:35 PM Ruby Revision 379d22ce (git): Bump RDoc version to 6.17.0 (#15439)
-
03:22 PM Ruby Misc #21689: DevMeeting-2025-12-11
- * [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? -
- PR to make all Regexps frozen and not just literals: https://github.com/ruby/ruby/pull/14547
Is it OK to merge it for Ruby 4.0? -
- Fix ruby/pathname#75.
https://github.com/ruby/pathname/commit/5ba967b274 -
- https://github.com/ruby/pathname/commit/a2edd25bc1
-
- https://github.com/ruby/pathname/commit/4580540a2b
-
- https://github.com/ruby/pathname/commit/d33d18e5e2
-
- Just clean the directory if it exists and is empty.
- 11:55 AM Ruby Revision eafaff94 (git): Re-introduce support for `io_close` hook. (#15434)
-
-
- Still development of the gem continues, sync as possible manually.
-
12/06/2025
-
11:22 PM Ruby Revision 33837abb (git): [ruby/stringio] [DOC] Tweaks for StringIO#pos=
- (https://github.com/ruby/stringio/pull/194)
https://github.com/ruby/stringio/commit/3cef1e0e5f -
- (https://github.com/ruby/stringio/pull/193)
https://github.com/ruby/stringio/commit/90728bbbca -
- (https://github.com/ruby/stringio/pull/192)
https://github.com/ruby/stringio/commit/8b1ee03cbe -
- (https://github.com/ruby/stringio/pull/191)
https://github.com/ruby/stringio/commit/f2a2a5a99e -
10:51 PM Ruby Misc #21689: DevMeeting-2025-12-11
- * [Feature #19979] Reconsider adding `&nil` to method declarations, to signal it won't accept a block
* Adding this feature should be possible now that the syntax moratorium is over -
- The id2ref table needs to be under a VM lock to ensure there are no race
conditions. The following script crashes:
o = Object.new
ObjectSpace._id2ref(o.object_id)
10.times.map do
Ractor.new do
10_000.times do... -
- https://github.com/ruby/stringio/commit/a7c118d786
- 05:26 PM Ruby Revision 47c0dae1 (git): [ruby/openssl] asn1integer_to_num: don't cast away const
- ASN1_ENUMERATED_to_BN() has been const-correct for a long time in all
supported libcrytos, so we can remove this workaround.
https://github.com/ruby/openssl/commit/d0f36a7c65 -
- It's possible that both half_days_in_day and day_in_nanoseconds are Ruby
objects, which means that creating day_in_nanoseconds may trigger GC.
Since half_days_in_day is not registered as a mark object until after
day_in_nanoseconds is al... - 04:50 PM Ruby Revision 8d3da814 (git): [ruby/openssl] Convert ossl_x509ext.c to opaque ASN1_STRING
- https://github.com/ruby/openssl/commit/a41cf28bab
- 04:50 PM Ruby Revision 38ad0806 (git): [ruby/openssl] Convert ossl_ts.c to opaque ASN1_STRING
- https://github.com/ruby/openssl/commit/8945f379b3
- 04:50 PM Ruby Revision a07997bf (git): [ruby/openssl] Convert ossl_ns_spki.c to opaque ASN1_STRING
- https://github.com/ruby/openssl/commit/0941ebbda5
- 04:50 PM Ruby Revision 98c151b0 (git): [ruby/openssl] Convert some of ossl_asn1.c to opaque ASN1_STRING
- This uses the normal accessors but leaves out BIT STRINGS, which will
need compat implementations for ASN1_BIT_STRING_get_length() and
ASN1_BIT_STRING_set1() for older libcryptos.
https://github.com/openssl/openssl/issues/29184
https://... - 04:50 PM Ruby Revision d3aa7b88 (git): [ruby/openssl] Convert ossl_ocsp.c to opaque ASN1_STRING
- OpenSSL plans to make asn1_string_st opaque, the struct underlying most
ASN.1 types such as ASN1_*STRING, ASN1_ENUMERATED, ASN1_INTEGER, etc.
Most of ruby/openssl's C code can be straigtforwardly converted to use
accessors available sinc... - 04:50 PM Ruby Revision 9dfb7bd7 (git): [ruby/openssl] const correct ossl_bin2hex()
- This helper only reads from its in parameter. Making that const
avoids a couple of casts in an upcoming change.
https://github.com/ruby/openssl/commit/970d5764e3 -
- Hmm, yes. I'll work on rebasing the PR.
-
08:12 AM Ruby Feature #21084: Declare objects have weak references
- @peterzhu2118 regardless of whether it should be public or not, given the performance benefits, should we merge your PR?
-
- The following warnings are emitted. We can use type punning to prevent
strict aliasing violations.
io_buffer.c:1935:23: warning: dereferencing type-punned pointer will break strict-aliasing rules [-Wstrict-aliasing]
1935 | r... -
- https://github.com/ruby/pathname/commit/60f5d58d73
-
- This method has been defined since 1.9, as the standard conversion
procedure.
https://github.com/ruby/pathname/commit/8f582dc65d -
- The constant `TO_PATH` was defined for 1.9 compatibility but the code
for it was dropped 10 years ago.
https://github.com/ruby/pathname/commit/95ad4ceb19 -
- Fix up ruby/ruby#14700.
-
- 11:11 AM Ruby Revision 0346206d (git): [ruby/prism] Avoid out-of-bounds reads
- Fixes https://github.com/ruby/prism/pull/3790.
https://github.com/ruby/prism/commit/173ccb84ad -
- - for non-String argument
- for NUL-contained argument
- for ASCII-incompatible argument -
- https://github.com/ruby/timeout/commit/4de4b4759c
- 08:44 AM Ruby Revision 42f5654b (git): Yield to scheduler if interrupts are pending. (#14700)
- 07:35 AM Ruby Revision 180020e1 (git): Fix `io_pwrite` fiber scheduler hook. (#15428)
- Fix io_pwrite fiber scheduler hook.
-
07:00 AM Ruby Revision 7259c18c (git): [DOC] Update NEWS about ZJIT (#15426)
-
-
- (https://github.com/ruby/stringio/pull/178)
https://github.com/ruby/stringio/commit/6449251678 -
-
-
-
-
-
- https://github.com/ruby/stringio/commit/dc93aa51d2
- 02:55 AM Ruby Revision a7dc53b9 (git): Add support for `u128`, `U128`, `s128` and `S128` integers to `IO::Buffer`. (#15399)
12/05/2025
-
- We don't set RSTRUCT_GEN_FIELDS when RCLASS_MAX_IV_COUNT(klass) != 0, so
we need to set RSTRUCT_SET_FIELDS_OBJ to 0 otherwise it may have an
invalid value and crash. -
- This reverts commit 228d13f6ed914d1e7f6bd2416e3f5be8283be865.
This commit makes default.c and mmtk.c depend on shape.h, which prevents
them from building independently. -
- * https://github.com/ruby/ruby-dev-builder/actions/runs/19973218359/job/57293388626
https://github.com/ruby/timeout/commit/45816b1b26 -
- ZJIT: Print local variable names GetLocal and SetLocal instructions
-
08:49 PM Ruby Revision d396a66a (git): ZJIT: Build by default when build environment allows
- "Default" means when `--enable-zjit` is absent from `./configure`
arguments. -
- Building ZJIT requires GNU make at the moment.
To get access to `$gnumake`, lift the `make` flavour detection up to the
environment section, before the JIT section runs. -
- This runs the detection, but does nothing with the result.
* Fixed version requirement in messages -- ZJIT requires >= 1.85 unlike
YJIT.
* New: Detect when rust 1.85 is available, and neither --enable-yjit nor
--enable-zjit is pas... -
-
-
-
- ~/zjit/zjit.mk:30: warning: overriding commands for target `~/build-default/'
~/yjit/yjit.mk:26: warning: ignoring old commands for target `~/build-default/'
~/zjit/zjit.mk:30: warning: overriding commands for target `~/build-def... -
-
-
- Its definition changes depending on e.g. whether there is YJIT in the
build. -
-
- Mostly YJIT. ZJIT already has the right bindings and this just tweaks
the CI configuration. -
-
-
- > insns.def:857:5: error: assigning to 'rb_zjit_func_t' (aka 'unsigned
> ...
> [-Werror,-Wpedantic] -
-
08:35 PM Ruby Revision 786f6739 (git): [ruby/prism] Correct constant pool bucket type logic
- When replacing an owned constant by a different type (constant or
shared) replace with the correct type instead of defaulting to
shared.
https://github.com/ruby/prism/commit/fbe9b131a1 - 08:21 PM Ruby Revision be12e198 (git): [ruby/prism] Avoid undefined int overflow behaviour
- Fixes https://github.com/ruby/prism/pull/3786.
https://github.com/ruby/prism/commit/b72b664675 - 08:20 PM Ruby Revision 834adc35 (git): [ruby/prism] Avoid out-of-bounds reads
- Fixes https://github.com/ruby/prism/pull/3784.
https://github.com/ruby/prism/commit/3fe862534b -
07:50 PM Ruby Feature #20750 (Closed): Allow rb_thread_call_with_gvl to work when thread already has GVL
- Applied in changeset commit:git|2b057859414d1ccfa4d88a898d27defc8c74dc1c.
----------
Allow rb_thread_call_with_gvl() to work when thread already has GVL
[Feature #20750]
Co-authored-by: Benoit Daloze <[email protected]> -
- [Feature #20750]
Co-authored-by: Benoit Daloze <[email protected]> -
- In a newly allocated OpenSSL X509 object, the notBefore and notAfter
fields contain an ASN1_STRING object with type V_ASN1_UNDEF rather than
an ASN1_TIME.
Commit https://github.com/ruby/openssl/commit/73484f67949a made asn1time_to_time(... -
- https://github.com/ruby/timeout/commit/281b2507e7
-
- * Fix indentation to stay a multiple of 2 spaces.
https://github.com/ruby/timeout/commit/a1d784cb66 -
- https://github.com/ruby/timeout/commit/82fb6f6925
-
- https://github.com/ruby/timeout/commit/daab9a2193
-
06:32 PM Ruby Revision ec28bd75 (git): [ruby/timeout] support Ractor
- 1. Introduce State to store all status.
2. Store State instance to the Ractor local storage if possible
3. Make `GET_TIME` (Method object) shareable if possible
3 is supporeted Ruby 4.0 and later, so the Rator support is works
only on R... -
05:14 PM Ruby Revision 65dbd571 (git): [ruby/psych] Use Node#to_ruby parse_symbols option
- https://github.com/ruby/psych/commit/907fd4fa97
-
02:21 PM Ruby Revision 930afa1c (git): Never shrink bignum on realloc
- As far as I can tell, this only ever shrinks by one, and it's really not
worth the expensive realloc for that. -
- Previously we only allocated bignums from the 40 byte sizepool, and
embedded bignum used a fixed size. -
- For the record: net/http now uses `TCPSocket.open(open_timeout:)` in place of `Timeout.timeout`.
https://github.com/ruby/net-http/pull/224 -
01:48 PM Ruby Feature #21552: allow String.strip and similar to take a parameter similar to String.delete
- Thank you, shugo.
"someone" he says is me. My use case is here.
I want to extract chunks from a file and pass them to a neural network model to detect the file type. The model requires two chunks: the `lstrip`ped beggining portion and ... -
08:36 AM Ruby Feature #21552: allow String.strip and similar to take a parameter similar to String.delete
- I'm not strongly opposed, but this kind of API that use a string to represent a collection of characters feel outdated. It is sometimes convenient, though.
-
02:26 AM Ruby Feature #21552: allow String.strip and similar to take a parameter similar to String.delete
- shugo (Shugo Maeda) wrote in #note-4:
> Unlike String#delete, the current implementation doesn't take multiple arguments.
> ...
I've noticed that String#count also take multiple selectors, so I've applied the same changes to String#str... -
- Earlopain (Earlopain _) wrote in #note-6:
> I have an implementation in https://github.com/ruby/prism/pull/3728 (maybe needs more tests)
It doesn't reject the following cases.
```ruby
x = begin
return
"NG"
end
# Expected:... -
- Earlopain (Earlopain _) wrote in #note-6:
> I have an implementation in [https://github.com/ruby/prism/pull/3728](https://github.com/ruby/prism/pull/3728) (maybe needs more tests)
Great, thanks!
> ...
Agreed. I think the evaluatio... -
- Sorry, I've missed this issue.
It is done to fix parse.y for 1 and 2.
I think 1 is OK, and probably no problem in practice.
But 2 is intentionally passed, IIRC, and bootraptest has such code.
```ruby
#1526 test_syntax.rb:572:
... - 12:34 PM Ruby Revision ea415e96 (git): [ruby/net-http] open: Never call Timeout.timeout in rescue clause
- The try-open_timeout-then-fallback-to-timeout introduced in
https://github.com/ruby/net-http/commit/1903cedd8cd0 works well, but when it errors
due to any reason in Rubies which do not support `open_timeout`, it
spits the rescued Argumen... -
- The current logic relies on sscanf() and error checks are almost
entirely missing. It also assumes that ASN1_STRING contents are NUL
terminated, which is undocumented and not guaranteed for all valid
ASN1_TIME objects.
Switch to using A... -
- Move variable declarations for OpenSSL::ASN1 classes to the top of the
file. asn1time_to_time() will need eASN1Error in the next patch.
https://github.com/ruby/openssl/commit/6c0ef87897 -
- https://github.com/ruby/timeout/commit/cd51eac3ca
-
- * [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. -
- https://github.com/ruby/ruby/pull/13232 had been already merged. Thanks!
-
- This variation is used when `-a` option is given.
-
- These variables are set by command line options, but it is deprecated
to assign them any value other than nil in ruby code. -
- Since c-ares has a compatibility issue on macOS, a full migration to it seems difficult. And DNS lookup is now interruptible by explicitly using pthread. Therefore, this is closed as not planned.
-
-
06:31 AM Ruby Revision 479185da (git): Add Set C API to news
- Also, don't use backticks around Set in the top level of the
Core classes updates section, as other classes/modules do not use
that format. -
05:20 AM Ruby Revision 5f6b31c2 (git): Correctly handle `Process.fork` with an active `Fiber.scheduler`. (#15385)
- In the child process, nullify the current fiber scheduler and set the current fiber to blocking.
-
- I extracted the relevant descriptions from the draft NEWS.md that
hsbt-san had Gemini generate by analyzing `git log`.
Co-authored-by: Hiroshi SHIBATA <[email protected]>
Co-authored-by: Alan Wu <[email protected]>
Co-autho... -
-
- `pr` should not change on this method.
12/04/2025
-
- https://github.com/ruby/rubygems/commit/3ddb740969
-
- I organized all examples the followings:
```
Total test time: 2468.41 seconds
Total files: 168
Group A: 42 files, 617.08 seconds
Group B: 42 files, 617.05 seconds
Group C: 42 files, 617.14 seconds
Group D: 42 files, 617.14 seconds
```
... -
- 09:51 PM Ruby Revision 8d8159e7 (git): Fix thread scheduler issue with thread_sched_wait_events (#15392)
- Fix race between timer thread dequeuing waiting thread and thread
skipping sleeping due to being dequeued. We now use `th->event_serial` which
is protected by `thread_sched_lock`. When a thread is put on timer thread's waiting
list, the ... -
-
- When defining a bmethod, we recorded the current Ractor's object in the
method. However that was never marked and so could be GC'd and reused by
a future Ractor. Instead we can use the Ractor's id, which we expect to
be unique forever.
... - 09:10 PM Ruby Revision de2c2bd6 (git): Take VM lock in `class_switch_superclass` (#15356)
- Safe multi-ractor subclass list mutation
We need to lock around mutation and accesses of a class's subclasses
list. Unfortunately we also need to do this when creating singleton
classes, as the singleton class does need to go into `supe... -
-
-
- Since around 2018, we have been using spaces for indentation for newly
added code[1]. The mixed use of tabs and spaces has repeatedly confused
new contributors who configured their editors to use a different tab
size than 8. Since git bl... -
- This reverts commit https://github.com/ruby/openssl/commit/830505172882.
The commit is part of the bigger effort to rewrite OpenSSL::ASN1 in
Ruby. OpenSSL::ASN1 is relatively isolated from the rest of ruby/openssl
and is not tightly bou... -
- with `RUBY_TYPED_FROZEN_SHAREABLE_NO_REC`,
if the receiver object is shareable on Method objects. -
- `T_DATA` has a flag `RUBY_TYPED_FROZEN_SHAREABLE` which means
if the `T_DATA` object is frozen, it can be sharable.
On the `Ractor.make_sharable(obj)`, rechable objects from the
`T_DATA` object will be apply `Ractor.make_shareable` recur... -
- ```ruby
$VERBOSE = true
Ractor.store_if_absent :key do
end #=> warning: the block passed to 'Ractor.store_if_absent' defined at <internal:ractor>:474 may be ignored
``` -
-
-
- Applied in changeset commit:git|cf12aff0593a0aa70d957b96f03df8288c025938.
----------
[Bug #21764] Propagate the encoding of ID to warning -
- The aliased name should be encoded properly.
```console
$ ruby -v -e 'def Warning.warn(s) = p(s, s.encoding); alias $分行 $/; $分行 = "\n"'
ruby 4.0.0dev (2025-11-27T12:05:45Z master 68a7edaa81) +PRISM [arm64-darwin25]
"-e:1: warning: ... -
03:32 PM Ruby Bug #21765: stop using the C runtime _read() on Windows
- I created [PR#15408](https://github.com/ruby/ruby/pull/15408)
For comparison, there is a patch [WIP] that utilizes encoding conversion.
https://github.com/YO4/ruby/commits/avoid-windows-crt-read/ -
- When creating an IO instance in Windows, the default data mode is text mode.
In reality, the IO encoding conversion mechanism is not used when encoding conversion is not performed. Instead, the CRLF conversion provided by the C runtime'... -
-
02:51 PM Ruby Bug #21758 (Closed): Prism fails to parse heredoc within %Q literal split by another heredoc
- Applied in changeset commit:git|465a86c3417d2936c311d9571aa9b6494a83eed8.
----------
[ruby/prism] Fix `%Q` with newline delimiter and heredoc interpolation
The lexer did not jump to the `heredoc_end`, causing the heredoc end delimiter
... -
- https://github.com/ruby/prism/pull/3780
-
- The lexer did not jump to the `heredoc_end`, causing the heredoc end delimiter
to be parsed twice.
Normally the heredocs get flushed when a newline is encountered. But because
the newline is part of the string delimiter, that codepath i... -
- I made PR [#15405](https://github.com/ruby/ruby/pull/15405)
-
- On Windows, output pipes have no crlf conversion applied by default except for stdout/stderr, also it ignore set_encoding(crlf_newline: true).
```
> ruby -e "IO.pipe { |r,w| w.set_encoding('ascii-8bit', crlf_newline: true); w.puts ?a; ... -
10:27 AM Ruby Feature #21722: Expose rb_gc_mark_weak API for use in extensions
- > @ivoanjo (Ivo Anjo) I don't know if this would work, but what about using rb_postponed_job_trigger to delay adding the object into a WeakMap? (there would also probably need to be a temporary buffer used to mark the object until it is ...
-
- > I'm fine with adding MOVEOBJ, it makes sense, but I don't think it's currently necessary. ObjectSpace.trace_object_allocations handles this by subscribing both to NEWOBJ and FREEOBJ and tracking the objects in its own hash table (subsc...
-
-
-
-
- ```ruby
$VERBOSE = true
Ractor.store_if_absent :key do
end #=> warning: the block passed to 'Ractor.store_if_absent' defined at <internal:ractor>:474 may be ignored
``` - 08:14 AM Ruby Revision 69173216 (git): Update default gems list at d343968ec3a899a29ff4e330dc914d [ci skip]
-
- https://github.com/ruby/json/commit/e5e4fd558e
-
- Fix: https://github.com/ruby/json/issues/912
In the case of surogate pairs we consume two backslashes, so
`json_next_backslash` need to ensure it's not sending us back in the
stream.
https://github.com/ruby/json/commit/0fce370c41 -
- I'd like to share a simplified example that demonstrates the autoload/require issue.
```ruby
# my_gem.rb
module MyGem
autoload :M, "my_gem/m" # (2)
end
# my_gem/m.rb
module MyGem
module M
autoload :X, "my_gem/m/x" # ... - 06:47 AM Ruby Revision 932762f2 (git): [ruby/rubygems] Increase connection pool to allow for up to 70% speed increase:
- - ### TL;DR
Bundler is heavily limited by the connection pool which manages a
single connection. By increasing the number of connection, we can
drastiscally speed up the installation process when many gems need
to be downloaded ... -
- Suggested by nobu, I've added documentation and tests for character selectors: https://github.com/ruby/ruby/pull/15400/commits/a9ad44007dbb0ea543ce1eb8748edd4213083c5f
Exmaples:
```
"012abc345".strip("0-9") # "abc"
"012abc345".st... -
- I just heard someone ask for a strip function that doesn't remove NUL characters.
Since Python's `str.strip` takes an optional argument, it might be a good idea to introduce a similar feature.
I've created a pull request at <https://... -
- - [Feature #21552] allow String.strip and similar to take a parameter similar to String.delete
- pull request: https://github.com/ruby/ruby/pull/15400
- Is it acceptable to take character selectors like String#delete?
- [Python's... -
02:27 AM Ruby Revision 0af85a1f (git): ZJIT: Optimize setivar with shape transition (#15375)
- Since we do a decent job of pre-sizing objects, don't handle the case where we would need to re-size an object. Also don't handle too-complex shapes.
lobsters stats before:
```
Top-20 calls to C functions from JIT code (79.4% of total ... -
- We generally know the receiver's class from profile info. I see 600k of these when running lobsters.
- 01:13 AM Ruby Revision 19f0df04 (git): ZJIT: Fix definite assignment to work with multiple entry blocks
- 01:13 AM Ruby Revision c764269f (git): ZJIT: Only use make_equal_to for instructions with output
- It's used as an alternative to find-and-replace, so we should have
nothing to replace. - 12:42 AM Ruby Revision 0e7e6858 (git): [ruby/psych] Add option to disable symbol parsing
- https://github.com/ruby/psych/commit/4e9d08c285
- 12:40 AM Ruby Revision 8f7d821d (git): [ruby/psych] Remove y Object extension in IRB
- Fixes: ruby#685
This feature can easily break how you use other gems like factory_bot or prawn.
https://github.com/ruby/psych/pull/747#issuecomment-3413139525
> But I kind of think we should leave `psych/y` around. If people really wan... -
12:39 AM Ruby Revision e96bbd71 (git): Remove needless parse.y `new_nil` macro
- In the past parse.y and ripper had different `new_nil` definition
so that `new_nil` returns `nil` for ripper.
```c
// parse.y
#define new_nil(loc) NEW_NIL(loc)
// ripper
#define new_nil(loc) Qnil
```
However Rearchitect Ripper (89cfc1... -
- In the past parse.y and ripper had different `value_expr` definition
so that `value_expr` does nothing for ripper.
```c
// parse.y
#define value_expr(node) value_expr_gen(p, (node))
// ripper
#define value_expr(node) ((void)(node))
```...
12/03/2025
- 09:59 PM Ruby Revision b43e66d3 (git): YJIT: Pass class and shape ID directly instead of object
- 09:59 PM Ruby Revision f1670733 (git): Move imemo fields check out of shape_get_next
- Not every caller (for example, YJIT) actually needs to pass the object.
YJIT (and, in the future, ZJIT) only need to pass the class. - 09:59 PM Ruby Revision 612a6680 (git): Remove spurious obj != klass check in shape_get_next
- This should never be true. I added an `rb_bug` in case it was and it
wasn't true in any of btest or test-all.
Co-authored-by: Aaron Patterson <[email protected]> -
- The reason I know of that it's unsafe to allocate (newobj or xmalloc) inside of the NEWOBJ hook is that a lot of code expects there to be a grace period between newobj and the next allocation where it can fill in the object without issui...
-
- Continuing from https://bugs.ruby-lang.org/issues/21710#note-17
ivoanjo (Ivo Anjo) wrote:
> Weakmap APIs are Ruby-level and can't be used from the NEWOBJ tracepoint.
Because that would recursively invoke the NEWOBJ tracepoint?
I though... -
09:37 PM Ruby Revision d7e55f84 (git): ZJIT: Use the custom iterator
- This commit uses the custom instruction iterator in arm64 / x86_64
instruction splitting. Once we introduce basic blocks to LIR, the
custom iterator will ensure that instructions are added to the correct
place. -
- This commit adds a specialized instruction iterator to the assembler
with a custom "peek" method. The reason is that we want to add basic
blocks to LIR. When we split instructions, we need to add any new
instructions to the correct bas... -
- I'm fine with adding `MOVEOBJ`, it makes sense, but I don't think it's currently necessary. `ObjectSpace.trace_object_allocations` handles this by subscribing both to `NEWOBJ` and `FREEOBJ` and tracking the objects in its own hash table ...
-
- +1 for `MOVEDOBJ` (or `MOVEOBJ` like `NEWOBJ/FREEOBJ`. It should be called just before/after moving) for object profiling purpose.
I'm not sure how to "simulate weak references"... -
- > That sounds like a use case for WeakRef or ObjectSpace::WeakMap.
> ...
In retrospect I was a bit confusing there -- The key missing bit is having something that's safe to use from `NEWOBJ` tracepoint, which is what `WeakRef` and `Weak... -
- ivoanjo (Ivo Anjo) wrote in #note-14:
> Maybe, yeah! Effectively the two requirements I believe are:
> ...
That sounds like a use case for `WeakRef` or `ObjectSpace::WeakMap`.
In #21722 you mention that would need the GVL, but don't you... -
- Having `MOVEDOBJ` would be very useful for being able to use the address as identity + I think it may come in handy for caches -- I think it would solve the "a) Be able to identify or reference object without preventing its GC, even if t...
-
- That makes me wonder if a `MOVEDOBJ` tracepoint would be a good solution. At least for now you can assume compaction almost never happens, and even when it happens, not so many objects would be moved.
So you'd get the benefit of using... -
- > rb_gc_mark_weak can be an option, another may be to have a mechanism to prevent compaction, so you can simply rely on addresses.
Maybe, yeah! Effectively the two requirements I believe are:
a) Be able to identify or reference obj... -
- Alright, I merged the fix, but as @jhawthorn mentioned, we should really figure out a way so you don't have to do this, because all we could think of when looking at the code was: "How has this ever worked?".
`rb_gc_mark_weak` can be ... -
- Applied in changeset commit:git|8c3909935e2ba9f79bf3492772c77c305a0d370b.
----------
Handle NEWOBJ tracepoints settings fields
[Bug #21710]
- struct.c: `struct_alloc`
It is possible for a `NEWOBJ` tracepoint call back to write fields... -
- Ah, I missed that `Enumerator.produce(...).size` returns `Infinity`.
I think that's the problem, it should return `nil` because it doesn't know if infinite or not, as you say in many cases it is not infinite.
If that's done, I think ... -
- I think having to provide a size for those cases to not raise is too inconvenient.
I agree with what has been said in #21654: we should not check raise for infinite loops, unless we are absolutely certain they are infinite.
A `size` of `... -
- I think what most people would agree here on is:
* If we know for sure a method will loop forever, then it seems helpful to raise.
* If we don't know for sure or have any doubt (e.g. with an Enumerator with `size`=`nil` which means u... -
-
-
- If GC trigger in the middle of `struct_alloc`, and the struct has more
than 3 elements, then `fields_obj` reference is garbage.
We must first check the shape to know if it was actually initialized. -
07:53 PM Ruby Misc #21762 (Open): Box: Inspect improvements
- 1\. Currently, `Ruby::Box` provides an `#inspect` method, and its output is reasonably short yet mysterious:
```ruby
Ruby::Box.new
#=> #<Namespace:3,user,optional>
```
First, it uses old `Namespace` name instead of `Ruby::Box`. Next... -
- Attempt to fix the following SEGV:
```
ruby(gc_mark) ../src/gc/default/default.c:4429
ruby(gc_mark_children+0x45) [0x560b380bf8b5] ../src/gc/default/default.c:4625
ruby(gc_mark_stacked_objects) ../src/gc/default/default.c:4647
ruby(gc_m... - 06:37 PM Ruby Revision fd02356e (git): ZJIT: Optimize NewArray to use rb_ec_ary_new_from_values (#15391)
-
- Not so sure how to trigger it but this is definitly more correct.
https://github.com/ruby/prism/commit/1bc8ec5e5d -
- See https://bugs.ruby-lang.org/issues/21756. Ripper fails to parse this,
but prism actually also doesn't handle it correctly.
When heredocs are used, even in lowercase percent arays there can be
multiple `STRING_CONTENT` tokens. We need... -
-
05:24 PM Ruby Misc #21689: DevMeeting-2025-12-11
- - [Feature #21701] Enumerator.produce accepts an optional `size` keyword argument
- Enumerator#to_set now rejects infinite sequences, and the current situation of Enumerator::Producer that its size is fixed to infinity creates an incon... -
- While profiling `Monitor#synchronize` and `Mutex#synchronize`
I noticed a fairly significant amount of time spent in
`rb_check_typeddata`.
By implementing a fast path that assumes the object is valid
and that can be inlined, it does mak... -
-
- https://github.com/ruby/prism/commit/c8e1b11120
-
04:22 PM Ruby Feature #21720: Add a native Binary Heap / Priority Queue to Ruby's Standard Library (heapify, heappush, heappop)
- ```ruby
Heap.heapify(array)
Heap.heappush(array, value)
Heap.heappop(array)
```
This looks like a very fragile API to me:
```ruby
Heap.heapify(array)
array << 'something random' # The heap is not encapsulated, so we can perform a... - 03:27 PM Ruby Revision 20fc8aff (git): Update default gems list at 94581b1ffde5e2afeba4631152955c [ci skip]
-
03:27 PM Ruby Feature #21005: Update the source location method to include line start/stop and column start/stop details
- bkuhlmann (Brooke Kuhlmann) wrote in #note-18:
> Any chance this could be fixed and included in time for the Ruby 4.0.0 release?
This is already implemented in Ruby 4.0.0-preview2 for Proc, Method, and UnboundMethod, but not Binding. -
- https://github.com/ruby/json/commit/4bdb2d14fe
-
- https://github.com/ruby/json/commit/ccca602274
-
- The NEWOBJ tracepoint can generate an object_id, that's alright,
what we don't want is actual instance variables. -
- The "EXIVAR" terminology has been replaced by "gen fields"
AKA "generic fields".
Exivar implies variable, but generic fields include more than
just variables, e.g. `object_id`. -
- > ConditionVariable and Mutex are too slow on CRuby.
So I looked at `Mutex` and `Monitor` for now.
I found an easy win for all `TypedData` which helps a bit: https://github.com/ruby/ruby/pull/15387
Before:
```
Mutex 13... -
- I ran the benchmark on TruffleRuby to get an idea how much of that slowdown is due to a slow ConditionVariable & other things, the results are interesting:
CRuby:
```
ruby 3.4.7 (2025-10-08 revision 7a5688e2a2) +YJIT +PRISM [x86_64-... -
- > To get an idea of the gain, how about benchmarking with a Thread::Queue for the AR::ConnectionPool?
Here's a benchmark of `Thread::Queue` vs Active Record's stack, vs the `TimedStack` in the popular `connection_pool` gem https://gis... -
- https://github.com/ruby/json/commit/7b62fac525
-
- https://github.com/ruby/json/commit/e0257b9f82
-
- When serializing an Array, and one of the elements of the Array requires
calling `to_json`, if the depth is changed, it will be used for the next
entries, which wasn't the case before
https://github.com/ruby/json/commit/5abd43490714, and... -
- Otherwise it's very error prone.
https://github.com/ruby/json/commit/c54de70f90 -
- It can be even simpler:
```rb
%Q
#{<<B}
B
```
Lex output similarly contains B twice. -
- https://github.com/ruby/rubygems/commit/90a0af8204
-
09:38 AM Ruby Bug #21761: NameError: uninitialized constant only when RUBY_BOX=1
- It does not reproduce without Rake.
```ruby
$ RUBY_BOX=1 ruby test_box.rb
ruby: warning: Ruby::Box is experimental, and the behavior may change in the future!
See doc/language/box.md for known issues, etc.
Constants defined:
R... -
- Here are original issue found.
# Steps to reproduce
```
git clone https://github.com/rails/rails
cd rails/activerecord
rm ../Gemfile.lock
bundle install
RUBY_BOX=1 bundle exec rake test
```
# Actual behavior
```
$ RUBY_BOX... -
- This issue is originally found when I tested Ruby::Box against Rails main branch, I think I have created a repro.
### Steps to reproduce
1. Install Ruby master branch
2. Follow these steps below.
```
git clone https://github.c... -
- For the prism part of the issue https://github.com/ruby/prism/pull/3778
-
- [Bug #21710]
- struct.c: `struct_alloc`
It is possible for a `NEWOBJ` tracepoint call back to write fields
into a newly allocated object before `struct_alloc` had the time
to set the `RSTRUCT_GEN_FIELDS` flags and such.
Hence we can't... -
- I am not sure whether all the problems have the same nature, but putting them together for now.
**1. Enabling box breaks `gem`+`require`**
Minimal reproducible example:
`test.rb`:
```ruby
gem 'faraday', '= 2.14.0'
require '... -
- I submitted a pull request to fix this: https://github.com/ruby/ruby/pull/15383
- 05:16 AM Ruby Revision 4762f429 (git): [DOC] typo fix in ruby/file.c
-
- The following failure of `make btest` is happend with `RUBY_MN_THREADS=1`.
```
#289 test_fork.rb:78:
def now = Process.clock_gettime(Process::CLOCK_MONOTONIC)
Thread.new do
loop { sleep 0.0001 }
end
... -
- 02:43 AM Ruby Revision 65cfd5e1 (git): Update default gems list at 1af7550114a0401229eda42de24a82 [ci skip]
-
- https://github.com/ruby/rubygems/commit/9d744beb56
-
- https://github.com/ruby/rubygems/commit/a55c485226
-
-
- 02:35 AM Ruby Revision f4466ec8 (git): Bump actions/checkout in /.github/actions/setup/directories
- Bumps [actions/checkout](https://github.com/actions/checkout) from 6.0.0 to 6.0.1.
- [Release notes](https://github.com/actions/checkout/releases)
- [Changelog](https://github.com/actions/checkout/blob/main/CHANGELOG.md)
- [Commits](http... - 02:31 AM Ruby Revision bf144d8c (git): Bump actions/checkout from 6.0.0 to 6.0.1
- Bumps [actions/checkout](https://github.com/actions/checkout) from 6.0.0 to 6.0.1.
- [Release notes](https://github.com/actions/checkout/releases)
- [Commits](https://github.com/actions/checkout/compare/v6...v6.0.1)
---
updated-dependen... -
- `ALWAYS_INLINE()` and `NOINLINE()` are defined with one argument.
https://github.com/ruby/json/commit/8fb727901e -
- This commit makes these codes to be invalid.
```ruby
case 0
in [a] | 1
end
case 0
in { a: b } | 1
end
case 0
in [{ a: [{ b: [{ c: }] }] }] | 1
end
```
12/02/2025
- 10:35 PM Ruby Revision a211abbc (git): Cache array length in `rb_ary_join` (#15362)
- When all elements are strings, we never have to recalculate the length
of the array because there are no conversion methods that are called, so
the length will never change. This speeds up the fast path by ~10%.
```ruby
a = ["1"*10, "2"... -
10:24 PM Ruby Revision a63147ee (git): [ruby/strscan] [DOC] Avoid being interpreted as a link
- (https://github.com/ruby/strscan/pull/180)
Since `[](n)` is being interpreted as a Markdown link, it cannot be
displayed as a method call.
I have corrected this by escaping the brackets so that they are
interpreted as strings rather tha... -
- https://github.com/ruby/prism/commit/cc0ca08757
-
- https://github.com/ruby/prism/commit/1988615ce1
-
- https://github.com/ruby/prism/commit/7ab6d9df47
-
- https://github.com/ruby/prism/commit/a20afe1674
-
- https://github.com/ruby/prism/commit/3e0b5c9eb7
-
- https://github.com/ruby/prism/commit/7eb169513a
-
- Eregon (Benoit Daloze) wrote in #note-10:
> FWIW it sounds quite bad contention-wise to use a global queue/stack (EDIT: and a stack is worse than a queue re contention) for connections, at least for the case `#connections >= #threads`.
... -
- To get an idea of the gain, how about benchmarking with a Thread::Queue for the AR::ConnectionPool?
For the micro benchmark it probably doesn't matter much semantically if LIFO or FIFO I guess since only 1 thread, but that way we can hav... -
- I'm in favor of a LIFO-alternative to `Thread::Queue` and `Thread::SizedQueue` in core. However, it should not have queue in the name, as queue implies FIFO behavior. It should have stack in the name, as stack implies LIFO behavior. `Thr...
-
- > I'm not sure how big of a bottleneck this is though, can you share the profile?
If you wish to investigate yourself:
- https://github.com/rails/rails/issues/55728
- https://github.com/rails/rails/pull/55736 -
- byroot (Jean Boussier) wrote:
> Define different classes for LIFO queues
> ...
Queue is actually defined under `Thread` so that's not that big of a concern.
If we add this I'd want a separate class for sure to not make many compromise... -
- > isn't that complicated
It isn't trivial either, but more importantly, it's not very performant. It used not to be important because checkout and checkin would only happen once per request. But async users demanded that we'd go throu... -
- Also I think one can't even do `BiasableQueue` (at least not that way) if that LIFO Queue was in core.
-
- Thanks for the link, that's helpful to discuss this.
> So I'd be happy to get rid of all that code.
https://github.com/rails/rails/blob/f5b981e6390a2574e63cf0889c421cdd4e446df0/activerecord/lib/active_record/connection_adapters/abs... -
- > How about making your own then with Array + Mutex + ConditionVariable?
As explained, that's what I currently have: https://github.com/rails/rails/blob/f5b981e6390a2574e63cf0889c421cdd4e446df0/activerecord/lib/active_record/connect... -
- byroot (Jean Boussier) wrote:
> I tried to use `git blame` to see if there was a justification for this
I think it's because many people AFAIK call FIFO a "Queue" and LIFO a "Stack".
From the description I get you're not simply using A... -
- Ah, yeah. It's not expecting disjointed strings in arrays where interpolation is not allowed. That doesn't hold with line continuations and heredocs. I'm working on a fix, should not be so difficult.
Edit: Sorry, I totally missed the di... -
05:15 PM Ruby Bug #21756: Ripper fails to parse pathological heredoc
- Looks like there is one more separate issue here. Prism and parse.y execution result differs.
~~~ruby
p <<-A, %w[a\
A
b c]
# Prism: ["a\n", "b", "c"]
# parse.y: ["a\nb", "c"]
~~~
Changing `%w` to `%W`, Prism and parse.y are the... -
- ```rb
# test.rb
pp <<-A, %w[j\
i
A
j]
```
```sh
$ RBENV_VERSION=ruby-dev ruby -rripper -ve "Ripper.lex(File.read('test.rb'), raise_errors: true)"
ruby 4.0.0dev (2025-11-28T10:49:46Z master dcb9e17f46) +PRISM [x86_64-linux]
/h... -
- Here is a patch: https://github.com/ruby/ruby/pull/15373
-
- * From https://github.com/truffleruby/truffleruby/commit/1f81db82d2969ff7c5de0dacdecb38252664f42c
https://github.com/ruby/psych/commit/dbabe7aac6 -
- * rb_struct_initialize() does not accept a Hash, and it's very brittle
to pass `[{...}]` and to rely on that C function using rb_keyword_given_p().
It basically worked accidentally, by having **members in the caller of the caller.
... -
- Thank you for the report. This is a bug in the allocationless anonymous splat optimization in Ruby 3.4, still present in the master branch. I'll work on a fix.
-
04:26 PM Ruby Bug #21757 (Open): Splatted args array is mutated when passing unexpected kwargs
- The bug is very simple to provoke.
``` ruby
proc = ->(*) { } # we do NOT expect any keyword arguments.
args = [1, 2]
proc.(*args, bug: true)
pp args #=> [1, 2, {bug: true}]
```
The `args` array now has a third element ... -
- Prism fails to parse this code
~~~ruby
<<A; %Q
A
#{<<B}
B
~~~
Parse error:
~~~
a.rb:4: syntax error found (SyntaxError)
2 | A
3 | #{<<B}
> 4 | B
| ^ unexpected constant, expecting end-of-input
~~~
`B` at line 4... -
- Mostly not having to list version-specific excludes when testing against ripper/parse.y
Also don't test new syntax additions against the parser gems. The version support
for them may (or may not) be expanded but we shouldn't bother whil... -
- Otherwise, it uses the latest prism version
https://github.com/ruby/prism/commit/86406f63aa -
- https://github.com/ruby/json/commit/28c57df8f7
- 02:49 PM Ruby Revision 9eafeaed (git): Box: Free rb_classext_t struct for a box when the box is GCed
- 02:49 PM Ruby Revision 84bc1f03 (git): Box: Mark boxes when a class/module is originally defined in it.
- When a class/module defined by extension libraries in a box, checking
types of instances of the class needs to access its data type (rb_data_type_t).
So if a class still exists (not GCed), the box must exist too (to be marked). - 02:49 PM Ruby Revision 75f8a116 (git): Box: Fix data type name
- 02:49 PM Ruby Revision 03ed220c (git): Box: load_wrapping is not needed now
- Top level constants are defined on the Object class's constant table,
and those constants can be referred as box::CONST_NAME from outside box.
So load_wrapping() is not needed now. -
01:39 PM Ruby Feature #21005: Update the source location method to include line start/stop and column start/stop details
- Any chance this could be fixed and included in time for the Ruby 4.0.0 release?
-
- I think this is fixed with https://github.com/ruby/resolv/pull/110
-
-
- It would be simpler to do `options[:lockfile] ||= ENV["BUNDLE_LOCKFILE"]`,
but that doesn't work as `options` is frozen.
Fixes https://github.com/ruby/rubygems/pull/9117
https://github.com/ruby/rubygems/commit/6e3603a0e9 -
07:00 AM Ruby Revision 0e22108d (git): [ruby/optparse] Remove `const_set` and instead use explicit assignments
- https://github.com/ruby/optparse/commit/6e2709a5fd
-
- Co-authored-by: Benoit Daloze <[email protected]>
https://github.com/ruby/rubygems/commit/463488b439
Co-authored-by: Patrik Ragnarsson <[email protected]> -
- Is there a way to resolve this? I just want to make Pathname (and Ruby) better :)
-
- Use the given `make-command` instead of the hard-coded `make` command.
TODO: Use it for `make up` as well, in the future. -
-
-
-
-
-
-
-
- - Suppress logos from sub makes.
- Set the prompt for `for` command when `echo` is on. -
- `rmdirs.bat` may require this option explicitly to remove parent
directories, in the future. -
- - intermediate source files
- timestamp files
- build directories -
-
- Try `rd` first for symlink to a directory; `del` attemps to remove all
files under the target directory, instead of the symlink itself. -
-
-
- `@` is not a command, and cannot be placed after `||`.
-
- * All Invariant::SingleRactorMode PatchPoint are replaced by
assume_single_ractor_mode() to fix https://github.com/Shopify/ruby/issues/875
for SingleRactorMode patchpoints.