This patch introduce M:N thread scheduler for Ractor system.
In general, M:N thread scheduler employs N native threads (OS threads)
to manage M user-level threads (Ruby threads in this case).
On the Ruby interpreter, 1 native thread is provided for 1 Ractor
and all Ruby threads are managed by the native thread.
From Ruby 1.9, the interpreter uses 1:1 thread scheduler which means
1 Ruby thread has 1 native thread. M:N scheduler change this strategy.
Because of compatibility issue (and stableness issue of the implementation)
main Ractor doesn't use M:N scheduler on default. On the other words,
threads on the main Ractor will be managed with 1:1 thread scheduler.
There are additional settings by environment variables:
`RUBY_MN_THREADS=1` enables M:N thread scheduler on the main ractor.
Note that non-main ractors use the M:N scheduler without this
configuration. With this configuration, single ractor applications
run threads on M:1 thread scheduler (green threads, user-level threads).
`RUBY_MAX_CPU=n` specifies maximum number of native threads for
M:N scheduler (default: 8).
This patch will be reverted soon if non-easy issues are found.
[Bug #19842]
When using a Ruby preview the require path of `bundler/setup` is
similar to `-r/opt/ruby3.3.0-preview2/lib/ruby/3.3.0+0/bundler/setup`.
The special character `+` in the string makes the Regexp fail,
leading to multiple addition of the same require statement each time
`set_rubyopt` is called (e.g. server reloading).
Escaping the characters in the string esure a correct match with all
the different Ruby versions.
https://github.com/rubygems/rubygems/commit/dd43dfa709
As discussed in https://github.com/rubygems/rubygems/issues/6273#issuecomment-1449176658
The `gem` method behaves awkwardly in standalone mode. Assuming bundler
isn't loaded at all, a call to gem might activate a gem that is not part
of the bundle (because it's the gem method defined in
lib/rubygems/core_ext/kernel_gem.rb and not
lib/bundler/rubygems_integration.rb). And when running with
`--disable-gems`, the gem method won't be defined at all so we'll get a
NoMethodError.
Calls to `gem` can appear in dependencies outside an application's
control. To work around this at GitHub we defined our own `Kernel#gem`
that no-ops.
I agree with https://github.com/rubygems/rubygems/issues/6273#issuecomment-1440755882
> people using standalone mode don't want to activate gems like Kernel.gem
This commit redefines `Kernel#gem` in the standalone script to no-op.
https://github.com/rubygems/rubygems/commit/bea17b55f1
We were setting the wrong `extension_dir` for git specs stubs
Additionally, the call to `self.extension_dir` was loading the
remote spec, which was avoidable since the stub had an extension dir
(and in fact its #gem_build_complete_path does exactly what we want
anyway)
Finally, now set the base_dir when loading the remote_spec from a
stub specification, since the git source sets the base dir for stubs
based on where the spec _will_ be installed to, and we want to preserve
that so the base_dir for the loaded spec & the stub are the same
https://github.com/rubygems/rubygems/commit/a94acb465b
Previously, Kernel#lambda returned a non-lambda proc when given a
non-literal block and issued a warning under the `:deprecated` category.
With this change, Kernel#lambda will always return a lambda proc, if it
returns without raising.
Due to interactions with block passing optimizations, we previously had
two separate code paths for detecting whether Kernel#lambda got a
literal block. This change allows us to remove one path, the hack done
with rb_control_frame_t::block_code introduced in 85a337f for supporting
situations where Kernel#lambda returned a non-lambda proc.
[Feature #19777]
Co-authored-by: Takashi Kokubun <takashikkbn@gmail.com>
[Bug #19012]
man recvmsg(2) states:
> Return Value
> These calls return the number of bytes received, or -1 if an error occurred.
> The return value will be 0 when the peer has performed an orderly shutdown.
Not too sure how one is supposed to make the difference between a packet of
size 0 and a closed connection.
If a legacy multi remote Gemfile depends transitively on a default gem,
then in standalone mode we'd fail to fetch the proper version from the
source that includes it, since we were adding it to `specs` (instead of
`remote_specs`), which was already including the default version of the
gem, and thus preventing the remote version from "overwriting that" and
being added to the index. We should add it to the `remote_specs` index
directly instead.
https://github.com/rubygems/rubygems/commit/05f4f9dfc0
On legacy Gemfiles with multiple remote sources, where all of them
support the compact index API, we were still falling back to full
indexes.
Fixing this also allows to simplifying the code.
https://github.com/rubygems/rubygems/commit/b1357c8e72
https://github.com/ruby/ruby/actions/runs/5956398507/job/16157091112
This has been extremely flaky on macOS GitHub Actions.
Benoit suggested to quarantine it if it's too problematic (it is) and
there's no reasonable fix in a short time (it already took too long).
So this commit follows the suggestion.
We should remove revert this once rb_cloexec_open() is fixed.
Koichi Sasada
Ellen Marie Dash
Martin Emde
Jacopo
Nobuyoshi Nakada
Daniel Colson
David Rodriguez
Peter Boling
Harshal Bhakta
Graham Marlow
krororo
Samuel Giddins
Alan Wu
Benoit Daloze
Hiroshi SHIBATA
Jean Boussier
Imir Kiyamov
Samuel Williams
Takashi Kokubun
Gaurav Khanna
Ngan Pham