Activity
From 10/16/2022 to 10/22/2022
10/22/2022
-
10:10 PM Feature #19076: Improve performance Proc#parameters
- If I'm reading the benchmark results correctly, this is slower in the case that the keyword argument is not provided. Is that correct? If so, I don't think we should do this, because it is far more common to not provide a keyword argum...
-
02:44 PM Feature #19076 (Closed): Improve performance Proc#parameters
- Improve performance `Proc#parameters` method(write in Ruby)
benchmark:
```yml
prelude: |
prc = proc{|x, y=42, *other|}
lam = lambda{|x, y=42, *other|}
benchmark:
proc_no_kwarg: |
prc.parameters
proc_given_kwarg: ... -
09:56 PM Feature #19078 (Closed): Introduce `Fiber#storage` for inheritable fiber-scoped variables.
- Pull Request: https://github.com/ruby/ruby/pull/6612
This is an evolution of the previous ideas:
- https://bugs.ruby-lang.org/issues/19058
- https://bugs.ruby-lang.org/issues/19062
This PR introduces fiber scoped variables, and... -
- Pull Request: <https://github.com/ruby/ruby/pull/6613>.
-
- The performance cost and complexity of this proposal is too high.
-
01:08 PM Feature #19071: Add Time#deconstruct, #deconstruct_keys, and #to_h
- @sawa With all due respect, I fail to see your point.
For all I know:
* most modern common serialization formats are using specifically "ISO8601 date and time" (which I specifically pointed at, NOT "everything ISO8601 is able to sta... - 10:34 AM Revision 8dfe3bdf (git): Improved formatting of `hash_foreach_iter` functions. (#6615)
-
08:55 AM Feature #19075: Binary searching for the last element
- I was only thinking of adding a new method, but an optional argument for `bsearch` also seems to be good.
The name of the argument would be as follows?
```ruby
bsearch(target: :first)
bsearch(target: :last)
```
-
02:15 AM Feature #19075: Binary searching for the last element
- Looks reasonable.
An alternative could be to add an optional (keyword) argument to `bsearch` and `bsearch_index`.
Do you consider the separated methods will be better?
About find-any mode, the last option (3) sounds something bet... -
08:08 AM Revision 2c1b1eae (git): merge revision(s) f88bff770578583a708093f4a0d8b1483a1d2039:
- [ruby/net-http] Revert "Replace Timeout.timeout in Net:HTTP#connect"
This reverts commit https://github.com/ruby/net-http/commit/753cae3bbccc.
https://github.com/ruby/net-http/commit/98caa38204
---
lib/... -
- [ruby/net-http] Rename `D` to `debug` in `Net::HTTP`
Renames `D` to `debug` in `Net::HTTP` and introduces an alias for
backwards compatibility. This was done for readability reasons, in that
`D` did not clearly r... -
08:01 AM Revision 420bdba1 (git): [rubygems/rubygems] Allow upcoming JRuby to pass keywords for Kernel#warn
- jruby-head (which will be JRuby 9.4.0.0) can now properly process
the keywords to Kernel#warn. I cannot think of any capability based
test for this so I constrained it using a version guard. Only JRuby
will ever hit the version guard.
... -
07:46 AM Revision 0d697e34 (git): Add [Feature #19013] to NEWS [ci skip]
-
01:46 AM Feature #19000: Data: Add "Copy with changes method" [Follow-on to #16122 Data: simple immutable value object]
- ufuk (Ufuk Kayserilioglu) wrote in #note-4:
> I think allowing the `dup` method on `Data` instances to take optional keyword arguments would be the ideal API for this
I had a strong visceral "no" reaction to this. To a certain extent...
10/21/2022
-
- My latest proposal is https://bugs.ruby-lang.org/issues/19075#note-6.
I will leave the initial proposal below.
---
PR: https://github.com/ruby/ruby/pull/6611
(I'm going to talk about `Array` here, but the same argument can be... -
- ufuk (Ufuk Kayserilioglu) wrote in #note-4:
> I think allowing the `dup` method on `Data` instances to take optional keyword arguments would be the ideal API for this:
I agree. This is the pattern that Sequel uses for datasets (which... -
09:35 PM Feature #19000: Data: Add "Copy with changes method" [Follow-on to #16122 Data: simple immutable value object]
- Since Ruby already has the [`Object#dup` method](https://ruby-doc.org/core-3.1.0/Object.html#dup-method), I think we don't need to create a new method to duplicate a `Data` instance with potentially slightly different values.
I think al... -
05:22 PM Feature #19000: Data: Add "Copy with changes method" [Follow-on to #16122 Data: simple immutable value object]
- Example from [value_semantics](https://github.com/tomdalling/value_semantics) gem which uses `with` as well:
```ruby
Point = ValueSemantics::Struct.new do
x Numeric
y Numeric
end
# An immutable instance
Origin = Point.new(x:... -
03:05 PM Feature #19000: Data: Add "Copy with changes method" [Follow-on to #16122 Data: simple immutable value object]
- Example from Sorbet which uses `with` as well:
``` ruby
class Point < T::Struct
const :x, Numeric
const :y, Numeric
end
# An immutable instance
Origin = Point.new(x: 0, y: 0)
right = Origin.with(x: 1.0)
up = Origin.with(... -
09:57 PM Revision a11952da (git): Rename `iv_count` on shapes to `next_iv_index`
- `iv_count` is a misleading name because when IVs are unset, the new
shape doesn't decrement this value. `next_iv_count` is an accurate, and
more descriptive name. -
- Before object shapes, we were using class serial to invalidate
inline caches. Now that we use shape_id for inline cache keys,
the class serial is unnecessary.
Co-Authored-By: Aaron Patterson <tenderlove@ruby-lang.org> -
09:06 PM Revision 87bb0bee (git): YJIT: Fix page rounding for icache busting
- Previously, we found the current page by rounding the current pointer to
the closest smaller page size. This is incorrect because pages are
relative to the start of the address we reserve. For example, if the
starting address is 12KiB mo... -
- > I think it is impossible to implement the method even in the future because a Fiber cannot be duplicated.
Of course, it is impossible directly.
I just might imagine that _if_ it would be a common stumbling question for consuming e... -
01:21 PM Feature #19061: Proposal: make a concept of "consuming enumerator" explicit
- zverok (Victor Shepelev) wrote in #note-7:
> I think it is totally Ok for the first implementation
Not only "the first implementation". I think it is impossible to implement the method even in the future because a Fiber cannot be dup... -
- @knu
Re:
> "But I'm skeptical about the usefulness of the `consuming?` flag" (from dev.log)
I believe it is extremely useful for introspection. For example the method like shown in the original ticket:
```ruby
def parse(enumera... -
- @mame
> To be honest, I'm not sure which ones are consuming and which ones are not.
Which is one of the points of this ticket! The distinction is internally present (as displayed in original code samples) but never spelled out and... -
- This patch makes sure that we're not accidentally reading rb_num_t
instruction arguments as VALUE and accidentally baking them into
code and marking them. Some of these are simply moving the cast earlier,
but some of these avoid potentia... -
07:42 PM Revision c7754a4d (git): [DOC] Add link targets (#6602)
-
06:44 PM Feature #19071: Add Time#deconstruct, #deconstruct_keys, and #to_h
- @duerst (Martin Dürst), @zverok (Victor Shepelev)
In short, I think you two misinterpreted my comment. I never wrote that the order of the units is wrong. I am talking about the combination of the units (thus, I wrote *arbitrarily select... -
- @matz Understood, thanks, I'll adjust the PR.
@sawa, for the sake of the argument
> The sequence `[year, month, mday, hour, min, sec, subsec]` is culturally biased and is arbitrarily selected.
I don't think this statement is fair.... -
08:39 AM Feature #19071: Add Time#deconstruct, #deconstruct_keys, and #to_h
- sawa (Tsuyoshi Sawada) wrote in #note-3:
> I have another reason against `deconstruct`. The sequence `[year, month, mday, hour, min, sec, subsec]` is culturally biased and is arbitrarily selected.
There's a bit of cultural bias, but ... -
- I have another reason against `deconstruct`. The sequence `[year, month, mday, hour, min, sec, subsec]` is culturally biased and is arbitrarily selected. For example, following the fact that month is a highly irregular concept, some peop...
-
- GC uses shapes to determine IV buffer width. Since allocation can
trigger GC, we need to ensure we only set the shape once we've fully
allocated new memory for the IV buffer, otherwise the GC can end up
trying to mark invalid memory. -
03:24 PM Bug #19053: Build errors in the WebAssembly case
- > The pre-built SDK is built on Ubuntu, and it expects the build machine to have libtinfo.so.6 with proper .gnu.version_d section.
I was told in the Fedora project that this issue is a known issue on the Fedora ncurses RPM package inc... -
- This reverts commit 97e50a056e795f6551cae2f07a9996f4adf73a1f.
-
- (https://github.com/ruby/stringio/pull/36)
Treats:
- #each_codepoint
- #gets
- #readline (shows up in doc for module IO::generic_readable, not class
StringIO)
- #each_line
https://github.com/ruby/stringio/commit/659aca7fe5 -
- Disable msystem: MINGW64 job on GitHub Actions (#6545)
---
.github/workflows/mingw.yml | 5 +----
1 file changed, 1 insertion(+), 4 deletions(-) -
- Skip running a DRb test on MinGW
It hangs even after a retry
https://github.com/ruby/ruby/runs/7966439530?check_suite_focus=true
We contacted GitHub Suppport about this before, and we concluded that
the ... -
- Skip TupleSpaceProxyTest on MINGW64 too
It wasn't specific to UCRT64
https://github.com/ruby/ruby/runs/7062992464
---
test/rinda/test_rinda.rb | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-) -
- Fix Rinda test teardown for omitted tests
New test failures on MINGW appeared after
c2e37c8ff7da395f33fae546d9ae9e2408fc9236.
1) Error:
Rinda::TupleSpaceProxyTest#test_00_template:
... -
- Try skipping TupleSpaceProxyTest on MinGW UCRT64
As you all know, MinGW UCRT64 CI has randomly got stuck despite its
"Finished tests" output.
Looking at the logs closely, it seems like all of the recent such
... -
- Bump actions/checkout from 2 to 3
Bumps [actions/checkout](https://github.com/actions/checkout) from 2 to 3.
- [Release notes](https://github.com/actions/checkout/releases)
- [Changelog](https://github.com/action... -
- Bump actions/cache from 2 to 3
Bumps [actions/cache](https://github.com/actions/cache) from 2 to 3.
- [Release notes](https://github.com/actions/cache/releases)
- [Commits](https://github.com/actions/cache/compar... -
- [MSWin] Cache installed vcpkg packages
---
.github/workflows/windows.yml | 7 +++++++
1 file changed, 7 insertions(+) -
- Fold command line items
---
.appveyor.yml | 28 +++++++++++++++++++++++++---
.github/workflows/windows.yml | 7 +++++--
2 files changed, 30 insertions(+), 5 deletions(-) -
- Show tool versions
---
.github/workflows/mingw.yml | 12 ++++++++++++
1 file changed, 12 insertions(+) -
-
12:39 PM Revision f88bff77 (git): [ruby/net-http] Revert "Replace Timeout.timeout in Net:HTTP#connect"
- This reverts commit https://github.com/ruby/net-http/commit/753cae3bbccc.
https://github.com/ruby/net-http/commit/98caa38204 -
-
- # The next dev meeting
**Date: 2022/11/17 13:00-17:00** (JST)
Log: https://github.com/ruby/dev-meeting-log/blob/master/DevMeeting-2022-12-01.md
- Dev meeting *IS NOT* a decision-making place. All decisions should be done at the bu... -
-
- done.
-
- Extract bundled gems by BASERUBY
---
common.mk | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-) -
- Applied in changeset commit:git|e026368061c56bc925aff58910a4b02f18b78c70.
----------
Range#size returns nil for (.."a") and (nil..)
Fixes [Bug #18983] -
- Oops, I missed the attached patch. Sorry
-
- https://github.com/ruby/ruby/pull/6604
-
- Bundled gems test needs `prepare-gems`
`prepare-gems` downloads and extracts the bundled gems, and these gems
are built by `build-exts` now.
---
.github/workflows/bundled_gems.yml | 5 +++++
.git... -
- Fixes [Bug #18983]
-
- Move to tool/lib/bundled_gem.rb
---
common.mk | 6 +++---
defs/gmake.mk | 4 ++--
tool/gem-unpack.rb | 53 -----------------------------------------------
tool/lib/... -
- Remove github and git related files from extracted bundled gems
---
common.mk | 3 ++-
defs/gmake.mk | 1 +
2 files changed, 3 insertions(+), 1 deletion(-) -
- Also the tests should use the configured pkg-config
---
test/mkmf/test_pkg_config.rb | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-) -
- Fix parallel build on MINGW
When the build is running with a base ruby then generating `x64-ucrt-ruby320.rc`
could fail due to a missing dependency to `x64-mingw-ucrt-fake.rb`.
This commit adds this dependency.
... -
- Exclude LIBPATHENV wrapper from PREP
---
common.mk | 4 ++--
configure.ac | 2 +-
template/Makefile.in | 2 ++
3 files changed, 5 insertions(+), 3 deletions(-) -
- Move duplicate dependencies
---
common.mk | 2 ++
template/Makefile.in | 5 -----
win32/Makefile.sub | 2 --
3 files changed, 2 insertions(+), 7 deletions(-) -
- Run find in PATH [ci skip]
---
template/Makefile.in | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-) -
- Clean intermediate files for update-deps [ci skip]
---
template/Makefile.in | 1 +
1 file changed, 1 insertion(+) -
- Load fake.rb at `BTESTRUBY`
So that `mkmf` checks work from `make run`, and also remove
duplicate `$(MINIRUBYOPT)` which is used in `$(MINIRUBY)`.
---
common.mk | 8 ++++----
template/... -
- Quiet if the target is already linked the same source
---
ext/extmk.rb | 2 +-
tool/ln_sr.rb | 6 ++++++
2 files changed, 7 insertions(+), 1 deletion(-) -
- Ignore EACCES on Windows
---
tool/ln_sr.rb | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-) -
- Create temporary file exclusively and clean
---
ext/extmk.rb | 9 +++++++--
1 file changed, 7 insertions(+), 2 deletions(-) -
- Make extensions under `Gem.extension_api_version` directory
---
ext/extmk.rb | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-) -
- Fix conversion from absolute path to relative path
---
tool/ln_sr.rb | 25 +++++++++++++++----------
1 file changed, 15 insertions(+), 10 deletions(-) -
- Avoid to symlink under symlink
---
ext/extmk.rb | 2 +-
tool/ln_sr.rb | 21 ++++++++++++++-------
2 files changed, 15 insertions(+), 8 deletions(-) -
- Move copying/linking extra files to Makefile so removed by `clean`
---
ext/extmk.rb | 81 ++++++++++++++++++++++++-----------------
tool/ln_sr.rb | 113 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++... -
- Ensure symlinks to bundled gem with exts have parent dir
When configuring with `--disable-rpath` and `--static-linked-ext` (e.g.
building for WASI), `extmk.rb` doesn't build exts under bundled gems,
and `.bundle/... -
- Move timestamps directory for bundled gems
---
ext/extmk.rb | 5 ++++-
1 file changed, 4 insertions(+), 1 deletion(-) -
- I think annotation per Fiber will cause conflict in usage between libraries. For example, consider the following library code.
```ruby
def fetch_info_from_url(uri)
Fiber.new do
Fiber.annotate "fetching info from #{ uri }"
... -
- Use built bundled gems in test-bundled-gems
---
tool/test-bundled-gems.rb | 49 ++++++++++++++++++++++++-----------------------
1 file changed, 25 insertions(+), 24 deletions(-) -
- Kill bundled gem tests when interrupted
---
common.mk | 2 +-
tool/test-bundled-gems.rb | 12 ++++++++++--
2 files changed, 11 insertions(+), 3 deletions(-) -
- Applied in changeset commit:git|5d4bfaccabe66d89460739ee682f1f78698c93a3.
----------
merge revision(s) ab3cb29bd9bff9c16cfb9d19cc02026998282c12:
Avoid defining the same test class in multiple files
Should fix issues wi... -
- Fallback PKG_CONFIG to the configured pkg-config always
---
lib/mkmf.rb | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-) -
- mkmf: Add a configure option to set verbose mode (V=1 or 0) in
mkmf.rb.
Note this change is only for `configure.ac`, not for Windows using
`win32/configure.bat`.
```
$ ./configure --help | grep ... -
- Use compiled instead of complied
Capitalize creates
---
lib/mkmf.rb | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-) -
- [DOC] make internally used classes/methods nodoc
Empty class documents are generated even with `:stopdoc:`.
---
lib/mkmf.rb | 18 ++++++------------
1 file changed, 6 insertions(+), 12 deletions(-) -
- mkmf: unify duplicate code in pkg_config
---
lib/mkmf.rb | 11 +++--------
1 file changed, 3 insertions(+), 8 deletions(-) -
- Avoid defining the same test class in multiple files
Should fix issues with parallel testing sometimes not running all
tests.
This should be viewed skipping whitespace changes.
Fixes [Bug #18731]
... -
- Update the test for [Feature #18615]
---
test/mkmf/test_flags.rb | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-) -
- Make implicit function declaration error [Feature #18615]
Enable `-Werror=implicit-function-declaration` by default for
building C extensions for early failures.
---
lib/mkmf.rb | 2 +-
1 file ch... -
- mkmf: pkg_config accepts multiple options
---
lib/mkmf.rb | 48 +++++++++++++++++++++++---------------------
test/mkmf/test_pkg_config.rb | 7 +++++++
2 files changed, 32 insertions(+),... -
- test: backfill coverage for MakeMakefile.pkg_config
---
test/mkmf/test_pkg_config.rb | 61 ++++++++++++++++++++++++++++++++++++++++++++
1 file changed, 61 insertions(+)
create mode 100644 test/mkmf/test... -
- Document MakeMakefile#append_cflags
This method is at least 7 years old and is widely used in the wild.
Since we need to support it, let's document it to make it discoverable.
Add docs and move it out of the `# :... -
- The commit 575ae50d16a03ed23357ec4ea0dbf7167fc26c8c was for debugging
the failure triggered by f55212bce939f736559709a8cd16c409772389c8, and
it was fixed at the commit 39f7eddec4c55711d56f05b085992a83bf23159e.
10/20/2022
-
10:42 PM Feature #19071: Add Time#deconstruct, #deconstruct_keys, and #to_h
- I agree with adding `deconstruct_keys` but not others (`deconstruct` and `to_h`).
* positional deconstruction force users to remember the meaning of order of elements (deconstruct)
* even though it's the same behavior to `deconstruct... -
- The expandarray instruction interpreters its arguments as rb_num_t.
YJIT was treating the num argument as a VALUE previously and when
it has a certain bit pattern, it can look like a GC pointer. The
argument is not a pointer, so YJIT cra... -
09:35 PM Bug #19043: Segfault on macOS 11.7 while using StringScanner in multiple threads
- eightbitraptor (Matthew Valentine-House) wrote in #note-2:
> Have you tried replicating with `GC.auto_compact=true` and `GC.stress=true`?
Thanks for the response. We tried to replicate the crash with `GC.stress=true` but were unabl... -
-
-
-
- Co-Authored-By: Aaron Patterson <tenderlove@ruby-lang.org>
-
- > For Invalid configuration wasm32-unknown-wasi': system wasi' not recognized error, you need to download the latest config.guess following the README.md build steps.
I was able to run the `configure` after running the `ruby tool/down... -
- > Therefore, you need to build the SDK by yourself on Fedora, then you will be able to build for WASI target.
I built the SDK from the source and installed it on the Fedora, and I was able to build the Ruby by following the [instructi... -
- > I'm not familiar with Fedora environment, but it looks like the pre-built wasi-sdk distributed in GitHub Releases is incompatible with Fedora.
> ...
Thanks for checking it! Let me check the things you mentioned step by step.
First,... -
- Thanks for the discussion, guys.
So, I think that we need to fix the Regexp timeout's code examples at least in the next Ruby (Ruby 3.2.0 preview3 or official 3.2.0) release note. The comment "# never interrupted" below in the example... -
- Discussed at the dev meeting.
> Is it though? It's basically breaking the intent of whoever set the global Regexp.timeout.
@matz agreed with this. He decided *not* to allow users to specify "force no timeout per instance, with ignoring... -
- Applied in changeset commit:git|0378e2f4a8319440dd65c82b16f189161472d237.
----------
Add Class#attached_object
Implements [Feature #12084]
Returns the object for which the receiver is the singleton class, or
raises TypeError if the re... -
- `attached_object` looks good to me. Accepted.
Matz. -
- Implements [Feature #12084]
Returns the object for which the receiver is the singleton class, or
raises TypeError if the receiver is not a singleton class. -
- Fix missing replacement in 1f91dcdab3b
---
template/Makefile.in | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-) -
01:10 PM Feature #19066: Enable Scorecard Github Action
- Sure, feel free to explore it. It is very simple to enable, to be honest. If you need some help, feel free to reach out to me.
-
- Hi, @joycebrum, Thanks to introduce Scorecards.
I try to setup scorecards action on ruby/ruby repository at first. -
-
-
-
- Copy from bundled gem source for test
---
common.mk | 13 +++++++----
defs/gmake.mk | 4 ++++
tool/gem-unpack.rb | 63 +++++++++++++++++++++++++++++++++++-------------------
3 file... -
12:30 PM Bug #19073: openssl gem 3.0.0 problem
- rhenium (Kazuki Yamaguchi) wrote in #note-3:
> Yes, openssl issues should go to the GitHub.
> ...
I just managed to try the other two combinations of ruby and gem versions, and a bit unexpected, it seems it actually depends on the ruby... -
12:09 PM Bug #19073 (Third Party's Issue): openssl gem 3.0.0 problem
- Yes, openssl issues should go to the GitHub.
But we need a way to reproduce the problem, or at least some clues, to investigate further. As a starter, you can try other combinations, such as Ruby 3.0.x + openssl 3.0.x or Ruby 3.1.x + ... -
- retro (Josef Šimánek) wrote in #note-1:
> Can you share more details to reproduce like Dockerfile? Also if I understand it well, Ruby OpenSSL issues are tracked at https://github.com/ruby/openssl/issues now.
Sorry, no. Ok, I wasn't a... -
11:18 AM Bug #19073: openssl gem 3.0.0 problem
- Can you share more details to reproduce like Dockerfile? Also if I understand it well, Ruby OpenSSL issues are tracked at https://github.com/ruby/openssl/issues now.
-
- We have a server using openssl 3.0.3 (i.e. written i C).
This has worked with all kinds of different clients (ruby, C, java...) without any problem... except now, with the latest ruby gem, openssl 3.0.0.
The behavior is odd: It suc... -
- Use configured GIT
---
common.mk | 4 ++--
defs/gmake.mk | 68 +++++++++++++++++++++++++++++------------------------------
2 files changed, 36 insertions(+), 36 deletions(-) -
- Get rid of building main again when test-bundled-gems
---
.github/workflows/ubuntu.yml | 1 +
.github/workflows/yjit-ubuntu.yml | 1 +
common.mk | 3 ++-
defs/gmake.m... -
- Run the prerequisites of test-bundled-gems in order
---
common.mk | 6 +++++-
defs/gmake.mk | 7 ++++++-
2 files changed, 11 insertions(+), 2 deletions(-) -
- After many discussion, I think it's fair to say, `dup` by default makes the most sense, and with that model, it could be considered more local. So I'll open a new proposal with the refined design.
-
- At least it shouldn't be named `locals` since it's not **local** to a fiber.
Maybe `storage`?
Matz.
-
- Followed up CVE-2022-39253 for bundler examples
---
spec/bundler/cache/git_spec.rb | 3 +++
spec/bundler/install/gemfile/git_spec.rb | 9 +++++++++
spec/bundler/update/git_spec.rb | 3 ... -
- Bypass git submodule add/update with git config
protocol.file.allow=always option.
Co-authored-by: Nobuyoshi Nakada <nobu@ruby-lang.org>
---
test/rubygems/test_gem_source_git.rb | 5 +++++
1 fil... -
10:20 AM Feature #18559 (Closed): Allocation tracing: Objects created by the parser are attributed to Kernel.require
- This was implemented by @ko1 in e35c528d721d209ed8531b10b46c2ac725ea7bf5
I tested it on our system it has the desired effect. Thank you! -
- The development version of Ruby 3.2.0 added ABI version suffix like `+2`. It will remove at the final release of Ruby 3.2.0.
@nobu said it's already implemented on our package script(?). But We should check it before the final release. -
09:25 AM Feature #19063: Hash.new with non-value objects should be less confusing
- austin (Austin Ziegler) wrote in #note-4:
> This mistake is much more common than I would have thought, and it appears in some fairly large projects (not everyone may have access to this result because cs.github.com is still apparently ... -
- Summary so far
### Is there any example of `Hash.new([])` being used intentionally?
Yes. [Shugo's example](https://github.com/shugo/textbringer/blob/65dff909295f14f466f7b27e824ea36fe1f9ddd7/bin/merge_mazegaki_dic#L5) is a valid usa... -
- austin (Austin Ziegler) wrote in #note-8:
> Edit on a couple of these. The GitLab behaviour, and @shugo’s code both do `hash[key] |= value_list`, so they *do* effectively do `hash[key] = value_list`, which makes them safer to use *in the... -
08:38 AM Revision e35c528d (git): push dummy frame for loading process
- This patch pushes dummy frames when loading code for the
profiling purpose.
The following methods push a dummy frame:
* `Kernel#require`
* `Kernel#load`
* `RubyVM::InstructionSequence.compile_file`
* `RubyVM::InstructionSequence.load_fr... -
-
- `load "foo.rb"` should work (similar enough) with or without wrapping module. In that sense, replacing `self` is unacceptable.
Probably you want to get the "target class/module" which is the target of constant/method definitions. I have... -
08:11 AM Feature #18462 (Closed): Proposal to merge WASI based WebAssembly support
- Updating the ticket status to say that WebAssembly/WASI support has been upstreamed by https://github.com/ruby/ruby/pull/5407
Some examples can be found here: https://github.com/ruby/ruby.wasm
-
- Unfortunately, we lost subversion server on svn.ruby-lang.org when I failed to migrate its server.
I will do the following works:
* Redirect `svn.ruby-lang.org` to `git.ruby-lang.org`.
* Backup and cleanup subversion data.
* Rem... -
- Applied in changeset commit:git|7563604fb868d87057733f52d780d841fc1ab6bb.
----------
[Bug #18998] Honor `#to_str` next to `#to_int` in `Kernel#Integer` -
-
- For the record, I write down the concerns:
* I don't see the convenience of `Fiber.annotate`. Convince me.
* Users need to understand how to use `Fiber#annotation` wisely. Need documents.
* The state should be stacked, not overridde... -
-
05:30 AM Bug #19033: One-liner pattern match as Boolean arg syntax error
- mame (Yusuke Endoh) wrote in #note-6:
> Note that `foo(1, 2, 3 in ary)` could be inherently ambiguous. It can be interpreted as `foo((1, 2, 3 in ary))` or `foo(1, 2, (3 in ary))`.
Agreed, these are difficult edge cases, and trying to... -
- Note that `foo(1, 2, 3 in ary)` could be inherently ambiguous. It can be interpreted as `foo((1, 2, 3 in ary))` or `foo(1, 2, (3 in ary))`.
-
- I can understand the feeling the `a in b` should be accepted as an argument expression. But in reality, it needs huge effort to modify the parser.
I don't think it worth both cost and complexity. If someone comes up with the pull-reques... -
- I agree with inconsistency; if edges were numeric, return the count (or Infinity), otherwise it should return nil.
`nil .. nil` can be considered as numeric-range both edge omitted, but at the moment I feel it should return nil as well.... -
- Upper bits than the least significant 4 bits need not be 0.
-
12:47 AM Revision b963fb25 (git): YJIT: Respect writable_addrs on --yjit-dump-iseq-disasm as well (#6596)
- YJIT: Respect writable_addrs on --yjit-dump-iseq-disasm
as well -
- 12:38 AM Revision 7f175e56 (git): Avoid missed wakeup with fiber scheduler and Fiber.blocking. (#6588)
- * Ensure that blocked fibers don't prevent valid wakeups.
-
- This solves multiple problems.
First, RB_VM_LOCK_ENTER/LEAVE is a barrier. We could at least use the
_NO_BARRIER variant.
Second, this doesn't need to interfere with GC or other GVL users when
multiple Ractors are used. This needs to b...
10/19/2022
-
- My first reaction seeing the example usage: they seem more for `Date` than `Time`, but I can see this being useful for both :)
-
- I believe that `Time` being suitable for pattern-matching is a reasonable feature with many possible usages, which will increase usability of `Time` and would be a good show case for pattern matching.
**Implementation decisions**
`... -
- YJIT: Skip dumping code for the other cb
on --yjit-dump-disasm -
-
- Co-Authored-By: Aaron Patterson <tenderlove@ruby-lang.org>
-
- Previously, enabling only "disasm" didn't actually build. Since these
two features are closely related and we don't really use one without the
other, let's simplify and merge the two features together. -
- I think this is a bug. Since `(0.51..5.quo(2)).count` raises a `TypeError` and `(0.51..5.quo(2)).size` is supposed to follow that, it should raise a `TypeError` as well.
-
-
- (https://github.com/ruby/stringio/pull/35)
Treated:
- #getc
- #getbyte
- #ungetc
- #ungetbyte
- #readchar
- #readbyte
- #each_char
https://github.com/ruby/stringio/commit/6400af8d9f -
02:54 PM Revision eeea633e (git): Stop zeroing memory on allocation / copy
- Shapes gives us an almost exact count of instance variables on an
object. Since we know the number of instance variables that have been
set, we will never access slots that haven't been initialized with an
IV. -
-
- Define BOOTSTRAPRUBY from HAVE_BASERUBY
---
common.mk | 1 +
configure.ac | 8 --------
template/Makefile.in | 7 +++++--
win32/Makefile.sub | 6 ++++--
4 files chang... -
-
- >fgrep: warning: fgrep is obsolescent; using ggrep -F
-
11:51 AM Feature #19063: Hash.new with non-value objects should be less confusing
- baweaver (Brandon Weaver) wrote in #note-29:
> I would also be very positive of this change, potentially for Array as well, to make this visible to folks working in the language.
For Array it doesn't make sense since the default valu... -
- etienne (Étienne Barrié) wrote in #note-24:
> While I like the idea of checking the arity of the block, maybe the core issue isn't entirely with the API for `Hash.new`, but the fact that default value/default proc are not visible. From ... -
- austin (Austin Ziegler) wrote in #note-4:
> **Note**: At least two of the examples later use `hash[key] |= values`, which is assignment, not resulting in the possible reuse of the default value in general. Others may result in the same.... -
- Adding a new element in the default proc will add by just referencing.
I'm not for doing it implicitly. -
- Doesn't `Ripper.lex` already provide this information?
-
- # Background
Implementation for Language Server Protocol (LSP) sometimes needs token information. For example both `m(1)` and `m(1, )` has same AST structure other than node locations then it's impossible to check the existence of `,`... -
- Co-authored-by: Nobuyoshi Nakada <nobu@ruby-lang.org>
- 10:49 AM Revision 567725ed (git): Fix and improve coroutines for Darwin (macOS) ppc/ppc64. (#5975)
- 05:53 AM Revision fc3137ef (git): Add support for anonymous shared IO buffers. (#6580)
-
- @nobu (Nobuyoshi Nakada) I interpret that your concern is that the hash will grow endlessly by random referencing (unintentionally or by an attack). I have some possible alternative modifications to the proposal for that.
1. Warn when a... -
- Adding a new element in the default proc will add by just referencing.
I'm not for doing it implicitly, and ditto for #19063. -
- > And if we use its naming convention for the 4 colours, they'll be:
> ...
Sounds good. I think it'd be also fair to make the APIs a bit more consistent by using `selection_foreground_color` and `selection_background_color`, but then yo... -
- zverok (Victor Shepelev) wrote in #note-3:
> `File#each_line` is consuming by design, if I understand correctly.
Well, I guess so. To be honest, I'm not sure which ones are consuming and which ones are not.
> ...
I think this is a... -
-
- (https://github.com/ruby/stringio/pull/33)
Treated:
- ::new
- ::open
- #string
- #string=
- #close
- #close_read
- #close_write
- #closed?
- #closed_read?
- #closed_write?
- #eof?
https://github.com/ruby/stringio/commit/be9b64d739 -
10/18/2022
-
-
09:33 PM Bug #19041: Weakref is still alive after major garbage collection
- tenderlovemaking (Aaron Patterson) wrote in #note-7:
> That method may not be putting the object in a register. Something else may have put it in a register or in the stack, and it just happens that no other machine code has overwritte... - 08:03 PM Revision 12c30f21 (git): Fixes remaining issues and turns back on dead code and unused (#6584)
-
- Treated:
- #lineno
- #lineno=
- #binmode
- #reopen
- #pos
- #pos=
- #rewind
- #seek
- #sync
- #each_byte - 06:19 PM Revision b2ba71df (git): Code clean around unused code for some architectures or features (#6581)
-
- mame (Yusuke Endoh) wrote in #note-2:
> But it introduces yet another new type of Enumerator (unrewindable Enumerator?), which is very complicated.
It's more complicated, but unrewindable enumerators already exist in practice (as sho... -
- For problem 1 you can check if an IO is seekable, and this would tell you whether you could restart from the beginning.
-
- > Here is my understanding
This is correct.
> ...
That's an interesting problem indeed! I'll look deeper into it.
But for now, I consider it an edge case that can be, in the worst case, just covered by docs. E.g. something like ... -
- Here is my understanding:
```ruby
[1, 2, 3].each.consuming? #=> false
$stdin.each_line.consuming? #=> true
# A user must guarantee whether it is consuming or not.
Enumerator.new {}.consuming? #=> false
Enumerator.new... -
-
03:56 PM Feature #18996: Proposal: Introduce new APIs to reline for changing dialog UI colours
- If we want to establish naming convention based on terminal apps, I think I'd pick `iTerm2` as the reference because 1) its very popular and 2) it allows more customisation than `Terminal.app`:
, 1 deletion(-) -
- builtin.c includes mini_builtin.c when cross-compiling
---
common.mk | 3 +++
1 file changed, 3 insertions(+) -
- Use cross compiling settings as fake.rb for snapshots
---
tool/make-snapshot | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-) -
- Ignore fake.rb for snapshot
---
common.mk | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-) -
- fake.rb needs id.h
---
common.mk | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-) -
- Now fake.rb is needed to prepare sources
---
tool/make-snapshot | 1 +
1 file changed, 1 insertion(+) -
- Prefer `File` class methods to write a string
---
tool/make-snapshot | 23 ++++++++++-------------
1 file changed, 10 insertions(+), 13 deletions(-) -
- BOOTSTRAPRUBY needs fake.rb when cross-compiling
---
configure.ac | 1 +
1 file changed, 1 insertion(+) -
- Fix potential target type confliction [ci skip]
---
common.mk | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-) -
- Make sources by BASERUBY if available instead of miniruby
---
common.mk | 25 +++++++++++++++----------
configure.ac | 6 +++++-
cygwin/GNUmakefile.in | 4 ++--
template/Ma... -
- Check if MSys shell can run a command with a drive letter
---
configure.ac | 10 ++++++++--
1 file changed, 8 insertions(+), 2 deletions(-) -
- > One of the problems I'm trying to solve is behavioural changes due to the internal fiber of enumerator. Can you propose some other way to fix this?
Maybe Enumerator-created Fiber should always use the Fiber locals of the caller Fibe... -
- > So in the context of creating a fiber, it's 30% slower. That's quite significant.
It's less than that on my machine, more like 20%, also it would be sligthly faster than that since done from C, so no method lookup etc.
```
Warmi... -
- ```
Warming up --------------------------------------
baseline 287.720k i/100ms
locals-dup 229.210k i/100ms
Calculating -------------------------------------
baseline 2.922M (± 0.7%) i/s - ... -
- Sorry but that's not remotely a good benchmark. We're talking about a potential overhead, so it has to be considered relative to the exiting work.
```ruby
require 'bundler/inline'
gemfile do
source 'https://rubygems.org'
gem... -
- If we were concerned about having a separate but similarly inheritable scope for Threads, maybe we should have this model:
```
Thread.locals -> child threads will inherit
Fiber.locals -> initially blank and only propagated to other fibe... -
- > It's extremely negligible.
Okay, good to know.
```ruby
require 'benchmark'
locals = {a: 1, b: 2, c: 3}
Benchmark.bm do |b|
b.report("dup") {1_000_000.times{x = locals.dup}}
b.report("assign") {1_000_000.times{x = loc... -
- > It's not thread unsafe to update the locals in Ruby because of the GVL.
We should not design general semantics based on a CRuby-only implementation detail.
It is thread unsafe and could crash other Ruby implementations for instance... -
- > I'm concerned about the memory and performance cost of calling dup per fiber.
It's extremely negligible.
> ...
I really don't see how.
> It also prevents shared optimisations which is something I want to use for my connection ... -
- Thanks for your examples. I agree there are different ways to solve this problem.
The key model I'm trying to introduce is a shared state "per request". It's actually similar to <https://github.com/steveklabnik/request_store> in a lot... -
- > this is more logical?
What is logical to me is to behave like POSIX environment variables. When you spawn a new fiber/thread it is initialized with a *shallow copy* of it's parent "locals".
What makes sense to me is:
```ruby
... -
- Fix paths of exts.mk to clean
exts.mk files are one level under the top of extension directories.
---
template/exts.mk.tmpl | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-) -
- exts.mk.tmpl: propagate MINIRUBY to enc.mk even though invoking from
exts.mk
This is another attempt to fix out-of-src build with
--with-static-linked-ext. The first attempt was
4f1888bda70981d9f5b1bf55a... -
12:05 PM Bug #19067: Module methods not usable at toplevel under wrapped script
- This is actually even confusing prior to the recent change:
> If the optional wrap parameter is true, the loaded script will be executed under an anonymous module, protecting the calling program's global namespace.
```ruby
# foo.r... -
- > This is a bit incompatible, but I wonder if it is better than dealing with `Module#using`, `Module#private_method`, etc. individually.
Yes that's a fair point, and to be honest, I didn't realize how much failed to work as "expected" u... -
- I think we can say the following by the same reasoning:
```ruby
# foo.rb
p self
```
```ruby
module MyModule; end
load "./foo.rb", MyModule #=> current: self, expected?: MyModule
```
because:
```ruby
module MyModule
... -
- > As an author of the target file I no longer control the nesting, which is fundamental to know the constant references I am writing are the ones I actually mean.
You cannot be sure the constant you are referencing is the one you mean,... -
- > The documentation is unclear about what does "being executed under" mean. Without that definition, we cannot know if this is a bug.
This is true. I am labelling it as a "bug" but it should be understood as "inconsistency". -
09:48 AM Bug #19067: Module methods not usable at toplevel under wrapped script
- I have always been skeptical about the ability to pass a module to `Kernel#load`.
As an author of the target file I no longer control the nesting, which is fundamental to know the constant references I am writing are the ones I actual... -
- ```ruby
#foo.rb
module Foo
end
private_constant :Foo
```
```ruby
module MyModule; end
load "./foo.rb", MyModule
# undefined method `private_constant' for main:Object (NoMethodError)
# ... -
- I opened a new issue #19069 regarding my proposal https://bugs.ruby-lang.org/issues/19063#note-15. If you have comments related to that idea, please continue on #19069.
-
- I'm also of the stance that this shouldn't surprise someone familiar with how Ruby is evaluated, but it is true that people have to start from somewhere and it's always good to try to help them get that understanding.
The problem with... -
09:57 AM Feature #19063: Hash.new with non-value objects should be less confusing
- While I like the idea of checking the arity of the block, maybe the core issue isn't entirely with the API for `Hash.new`, but the fact that default value/default proc are not visible. From the examples above, when we print the value of ...
-
- This is a spin-out from #19063, and is a recapture of my comment https://bugs.ruby-lang.org/issues/19063#note-15.
I propose to change the behavior of `Hash.new` when it takes a block with its parameter signature absent. In such case, ... -
- - [Feature #19066] Enable Scorecard Github Action
- I don't know the details, but at first sight, it seems worth doing -
- * [Feature #19068] Upgrades required Bison version for development
* Change set is only improvement of yydebug option output, however this upgrades required Bison version. Then want to confirm it's acceptable.
* [Feature #19070] Enhanc... -
- - [Bug #19067] Module methods not usable at toplevel under wrapped script
- Similar to #18960
- Based on my understanding of the `wrap` module, this should work, but `top_self` currently extends the `wrap` module.
- Up to 3.2, `wra... -
- # Background
Printing node type, e.g. `NODE_LIT`, for `yydebug` option is helpful when developing parse.y.
Before:
```
Reducing stack by rule 639 (line 5062):
$1 = token "integer literal" (1.0-1.1: 1)
-> $$ = nterm simple_... -
-
- Line with substituted issue references with URLs are often very long.
Although Git (and GitHub) recommends folding subject lines less than
50 columns, but many commits ignore this, so fold at 68 columns for
now. - 08:22 AM Revision 93d73cfe (git): Update default gems list at dffca50bb609661f04660221834c48 [ci skip]
-
- (https://github.com/ruby/fiddle/pull/109)
These structs don't need to be freed as part of finalization, so lets
free them immediately.
https://github.com/ruby/fiddle/commit/8a10ec1152 - 08:21 AM Revision e84ea4af (git): [ruby/fiddle] Add support for linker script on Linux
- GitHub: fix https://github.com/ruby/fiddle/pull/107
Reported by nicholas a. evans. Thanks!!!
https://github.com/ruby/fiddle/commit/49ea1490df -
08:21 AM Revision 93da67d4 (git): [ruby/fiddle] Fix filenames for glibc SO files on alpha and ia64
- (https://github.com/ruby/fiddle/pull/105)
Fixes [Bug #18645]
https://github.com/ruby/fiddle/commit/9a5a1dab1d - 08:21 AM Revision 08ec6562 (git): [ruby/fiddle] Bump version
-
- Fiddle is a gem and has the external upstream which supports older
versions of Ruby. -
- > there's a lot of sharp edges
I consider the first one a bug, thanks for identifying it. I've filed an issue for it: https://bugs.ruby-lang.org/issues/19067
The others I'm not quite clear on, but if the current `wrap` parameter is... - 07:54 AM Revision 007946c2 (git): [rubygems/rubygems] Bundler: github DSL has used https protocol over git
- This behavior change was done in Bundler 2.2.0.
https://github.com/rubygems/bundler/pull/7142
Signed-off-by: Takuya Noguchi <takninnovationresearch@gmail.com>
https://github.com/rubygems/rubygems/commit/9510190be1 -
- https://github.com/rubygems/rubygems/commit/6214d00b2315ed37c76b1fbc1c72f61f92ba5a65
-
07:33 AM Revision 2ab7bb89 (git): [rubygems/rubygems] TestGemSecuritySigner#test_sign: Correct the expectation for new cert.
- https://github.com/rubygems/rubygems/commit/bed360410d
-
- https://github.com/rubygems/rubygems/commit/e2d533591c
-
- https://github.com/rubygems/rubygems/commit/3a607f43d1
-
07:33 AM Revision cb63a3f4 (git): [rubygems/rubygems] Materialize platforms strictly on Windows too
- https://github.com/rubygems/rubygems/commit/ab11545f80
-
- https://github.com/rubygems/rubygems/commit/ba9d9b1890
-
- I didn't realize how the `Bundler::GemHelpers.generic` method works when
I added this. It already matches this and other java platforms properly.
https://github.com/rubygems/rubygems/commit/5f0f0c678c -
- https://github.com/rubygems/rubygems/commit/06faad1e05
-
- https://github.com/rubygems/rubygems/commit/1c0eb63c6a
-
- https://github.com/rubygems/rubygems/commit/62c2edd255
-
- https://github.com/rubygems/rubygems/commit/788e46e152
-
- https://github.com/rubygems/rubygems/commit/e60459d6b6
-
- https://github.com/rubygems/rubygems/commit/b31308fb4c
-
- https://github.com/rubygems/rubygems/commit/996fd81871
-
- We just call it once for resolution, so we can simplify things.
https://github.com/rubygems/rubygems/commit/99c144fbe3 -
- https://github.com/rubygems/rubygems/commit/33769ddb07
-
- https://github.com/rubygems/rubygems/commit/b4a0fcd2d2
-
- https://github.com/rubygems/rubygems/commit/426748ed06
-
- https://github.com/rubygems/rubygems/commit/665051d085
-
- https://github.com/rubygems/rubygems/commit/999b644708
-
- Removing /\A.+\n/ could have an unexpected impact, depending on how
RubyLex.generate_local_variables_assign_code is implemented. It feels
like a too much assumption and the intention isn't immediately clear,
so I added these changes.
ht... -
05:44 AM Revision a09f764c (git): [ruby/irb] Always use local variables in current context to parse code (https://github.com/ruby/irb/pull/397)
- * Use local_variables for colorize, code_block_open check, nesting_level and assignment_expression check
* Check if expression is an assignment BEFORE evaluating it. evaluate might define new localvars and change result of assignment_ex... -
- * Fix backtick method def method call handled as backtick open
* Fix handling heredoc in check_string_literal
* Sort result of lexer.parse by pos in ruby<2.7. It's not sorted when the given code includes heredoc.
* Update lib/irb/ruby... -
- to ruby/git.ruby-lang.org
-
- - Substitute `GH-xxxx` which does not contain `#` too.
- Split each substitutions. -
02:25 AM Bug #19042 (Closed): Bug: Dir.glob ignores subdirectories in alternation when alternation is preceded by recursive directory pattern
- Applied in changeset commit:git|329d5424a479bb08e75bd750c51a5382e382731c.
----------
[Bug #19042] Fix Dir.glob brace with '/'
Dir.glob brace pattern with '/' after '**' does not match
paths in recursive expansion process.
We expand bra... -
- Dir.glob brace pattern with '/' after '**' does not match
paths in recursive expansion process.
We expand braces with '/' before expanding a recursive.
Co-authored-by: Nobuyoshi Nakada <nobu@ruby-lang.org>
10/17/2022
-
- to possibly deal with failures like https://ci.appveyor.com/project/ruby/ruby/builds/45097615.
I'm not sure how vcpkg retries things, but at least aria2c seems to attempt downloads 5 times by default.
https://aria2.github.io/manual/en/h... -
- > Even if the "locals hash" is copied when a new fiber is created, you can still have the behavior your want by storing a mutable object.
And that's exactly my point, so why bother trying to prevent mutability, especially if there is ... -
- > Here is a short example of what I want:
That example makes very little sense to me:
- How is this a "local" if it's accessible by other fibers?
- I could understand inheriting (a copy of) the locals of the parent fibers, but... -
- Here is a short example of what I want:
```ruby
Fiber[:count] = 0
enumerator = Enumerator.new do |y|
10.times do |i|
Fiber[:count] += 1
y << i
end
end
enumerator.next
p Fiber[:count] # 1
enumerator.to_a
p Fiber[:count] # ... -
- The only thing I really care about is an efficient way to implicitly share specific state within an execution context.
```
let(x = 10) do
Thread.new do
get(x) # => 10
end
Enumerator.new do
get(x) # => 10
end
end
```
Li... -
- ioquatix (Samuel Williams) wrote in #note-17:
> > It would be a semantic bug if adding a new local in a child fiber affects a parent fiber as @byroot (Jean Boussier) said.
> ...
I would think all other languages with similar functionalit... -
- > ActiveRecord connections are thread safe
Again, "thread safe" is a very loaded term. It just means that things won't be utterly broken if you use that API in a concurrent scenario, it doesn't mean that you want to share that connect... -
- ActiveRecord connections are thread safe, and it makes total sense to me that an enumerator might like to run in the same transaction as its surrounding context. In fact I've read many times people complaining about this exact problem.
... -
- > No, it's not bad to share a DB connection between threads, if the DB connection itself is thread safe
"Thread safe" can mean a lot of things. Generally speaking, the vast majority of protocols don't support concurrent access, so mos... -
- > It would be a semantic bug if adding a new local in a child fiber affects a parent fiber as @byroot (Jean Boussier) said.
I don't agree this is a semantic bug. Ruby doesn't have any existing data structures which behave like this, and... -
- For sure we need the `.dup` on the internal hash used for storage when inheriting it to a new Fiber.
It would be a semantic bug if adding a new local in a child fiber affects a parent fiber as @byroot said.
Or we'd need some kind of pe... -
- Okay, with this small change, we can propagate locals across threads: https://github.com/ruby/ruby/pull/6566/commits/ac09cff4c36ca9535b0439e3d599ac7ef9899c97
Small test program:
```ruby
Fiber[:x] = 10
thread = Thread.new do
pp Fibe... -
- Just for my own note:
```ruby
struct rb_execution_context_struct {
/* storage (ec (fiber) local) */
struct rb_id_table *local_storage;
}
```
Fiber locals already exist on `rb_execution_context_struct`. -
- > Well same than for your Fiber example, but with threads.
I don't think we should introduce `Thread#locals` because then we end up with two interfaces for execution context state. Fiber locals are introducing a limited form of "dynamic... -
- > I think of that more of a design choice rather than an outright issue.
Of course. But it is generally considered good design to help avoid shared mutable state. But again, I can live with the current proposal.
> ...
Well same tha... -
- > If you simply inherit this hash, a child fiber changing Something.current would change it in its parent, sibling and child fibers too, which is super unlikely to be what you want.
I think of that more of a design choice rather than an... -
- > Can you give an example of the problem caused by mutability?
Sure, you can look at how `ActiveSupport::IsolatedExecutionState` is used today. For instance [`ActiveSupport::CurrentAttributes`](https://github.com/rails/rails/blob/caf9... -
- > The problem however is the semantic. If all the values in Fiber.locals are immutable, it's quite painless, but if they are mutable what do you do?
Can you give an example of the problem caused by mutability?
> ...
The simplest wa... -
- So that is indeed a need that arise when you start using fibers or even threads. Existing Ruby code rely a lot on Fiber / Thread global state.
The problem however is the semantic. If all the values in `Fiber.locals` are immutable, it'... -
- Seems pretty different from the thread local storage.
Intentional? -
-
- https://github.com/ruby/ruby/commit/0d7292ec734d49def118959d6a80588e7c0f7a39
-
- https://github.com/ruby/irb/commit/ddd7dbe2c5
-
- Hi, I am Joyce and I'm working on behalf of Google and the [Open Source Security Foundation](https://openssf.org/) to help essential open-source projects improve their supply-chain security.
Would you consider adopting an OpenSSF tool... -
- @baweaver (Brandon Weaver) wrote:
> Should we open a new ticket for this shorthand Hash.new { [] } syntax?
Either way is okay with me. -
- I would agree that `Hash.new { [] }` would be very nice, and as mentioned has a dual in `Array.new(5) { [] }`.
While I do not think that it is a bug (and my apologies for earlier bad wording) to have `Hash.new([])` do what it currentl... -
- ioquatix (Samuel Williams) wrote in #note-16:
> The question is, are there valid use cases for `Hash.new(default_value)` and/or `Array.new(count, default_value)` where the user expects mutable values to be shared?
Sure:
```ruby
c... -
- Dan0042 (Daniel DeLorme) wrote in #note-19:
> The main downside is incompatibility with previous ruby versions. [...] Note that I haven't found that many gems that use `Hash.new{value}` syntax, and most of them would work fine with this ... -
- sawa (Tsuyoshi Sawada) wrote in #note-15:
> ```ruby
> ...
I *love* this idea! It makes the common case so much easier and uncluttered. It's easy to bypass if you need to. It works great in combination with a frozen default value; when ... -
- This is a good idea. Even though the current behavior is "logical", and experienced rubyists have gotten so used to this that it's second nature, there's an opportunity to make ruby more friendly for beginners by removing a well-known go...
-
- In support of `Hash.new{default_value}`, `Array.new(count){default_value}` works in the same way.
-
- @sawa I appreciate your technical knowledge and insight.
However... the practical reality is the vast majority of the users are using this interface incorrectly. This might imply that the theoretical basis for how `Hash.new(default_valu... -
- I can imagine why you particularly picked up (or people are more troubled with) `Hash.new` rather than `Array.new`. Probably, you are feeling that `Hash.new {|h, k| h[k] = []}` is too long to write. So I have an even better idea. If poss...
-
- This issue is essentially the same issue as people mistakenly writing `Array.new(5, [])` with the intention to actually do `Array.new(5){[]}`. And the same issue can appear elsewhere.
Users need to have better understanding of how argum... -
- duerst (Martin Dürst) wrote in #note-11:
> `Hash.new(my_method_returning_an_empty_array)` would behave exactly the same as `Hash.new([])` (assuming `my_method_returning_an_empty_array`) really did what its name said). So compositionalit... -
- ioquatix (Samuel Williams) wrote in #note-10:
> The current behaviour is ambiguous at best and as @austin has shown, buggy and incorrect at worst / in practice.
> ...
There is definitely a chance of breaking compatibility. It would be ... -
- sawa (Tsuyoshi Sawada) wrote in #note-9:
> 2. Your proposal 1 would break one of the very basic principles of Ruby, or perhaps of any computer language, or even any natural language. That is, the arguments passed to a method are evalu... -
- The current behaviour is ambiguous at best and as @austin has shown, buggy and incorrect at worst / in practice.
I prefer option 1 but it's true it can break compatibility with this poorly understood behaviour.
I agree in principle... -
- baweaver (Brandon Weaver) wrote in #note-5:
> Then what would qualify as a proof?
I don't think it is provable. That is my point. You made a statement that cannot be assured.
I know that a lot of beginners make the kind of mistake you ... -
03:25 AM Feature #19063: Hash.new with non-value objects should be less confusing
- Edit on a couple of these. The GitLab behaviour, and @shugo’s code both do `hash[key] |= value_list`, so they *do* effectively do `hash[key] = value_list`, which makes them safer to use *in these particular cases*. I don’t think that thi...
-
- sawa (Tsuyoshi Sawada) wrote in #note-2:
> > When people use Hash.new([]) they mean Hash.new { |h, k| h[k] = [] }.
> ...
It partially comes from not necessarily understanding `default_value` in the Hash documentation. The documentation... -
- > austin (Austin Ziegler) wrote in #note-4:
> ...
I would agree that option 2 is the most viable without serious breaking behavior, and option 3 would be better long-term. Option 1 was an attempt at a compromise, but would have breaking... -
- > For note, coming up with a lot of examples pointing towards that is not a proof.
Then what would qualify as a proof? I am confused by this as it asserts that examples of people being confused are not proof, whenever by definition it... -
- This mistake is much more common than I would have thought, and it appears in some fairly large projects (not everyone may have access to this result because `cs.github.com` is still apparently in preview):
https://cs.github.com/?scopeN... -
- sawa (Tsuyoshi Sawada) wrote in #note-2:
> > When people use Hash.new([]) they mean Hash.new { |h, k| h[k] = [] }.
> ...
Do you have examples of this? I have had several cases across multiple large Ruby companies where this behavior had ... -
- > When people use Hash.new([]) they mean Hash.new { |h, k| h[k] = [] }.
I don't think so. Can you prove it? For note, coming up with a lot of examples pointing towards that is not a proof.
> ...
This contradicts with what you have admi... -
- @matz what do you think?
-
05:49 PM Bug #18605: Fails to run on (newer) 32bit Windows with ucrt
- This (unsurprisingly) also breaks ruby on arm64 and is blocking our arm64 porting effort: https://github.com/msys2/MINGW-packages/pull/13115
-
- * YJIT: Allow --yjit-dump-disasm to dump into a file
* YJIT: Move IO implementation to disasm.rs
* YJIT: More consistent naming -
- Co-authored-by: Alan Wu <alansi.xingwu@shopify.com>
Co-authored-by: Maxime Chevalier-Boisvert <maxime.chevalierboisvert@shopify.com> -
- parker (Parker Finch) wrote in #note-6:
> I'm still curious _why_ calling `#weakref_alive?` on the `WeakRef` seems to put the underlying `Object` (that the `WeakRef` delegates to) in a register or on the stack. But the fact that this is... -
- * Make mjit_cont sharable with YJIT
* Update dependencies
* Update YJIT binding - 04:26 PM Revision 07a93b1e (git): YJIT: Do not call `mprotect` when `mem_size` is zero (#6563)
- This allows x86_64 based YJIT to run on Docker Desktop on Apple silicon (arm64)
Mac because it will avoid a subtle behavior difference in `mprotect` system call
between the Linux kernel and `qemu-x86_64` user space emulator. -
04:12 PM Feature #17576: Partial Functions (procs, lambdas)
- Isn't the topic starter talking about pattern arguments for function? This is something I'm personally looking froward to. It is similar to Elixir's functions pattern matching and increases the expressiveness of the language by order of ...
-
- The commented out instance of free_block() is left over from the port.
The addition in gen_single_block() was a place we missed. The new block
is allocated in the same function and could have invariants associated
with it even though the... -
-
- Applied in changeset commit:git|60defe0a68a40d1b3225cf6b971ea195e19ae2e2.
----------
thread_sync.c: Clarify and document the behavior of timeout == 0
[Feature #18982]
Instead of introducing an `exception: false` argument to have `non_... -
- [Feature #18982]
Instead of introducing an `exception: false` argument to have `non_block`
return nil rather than raise, we can clearly document that a timeout of 0
immediately returns.
The code is refactored a bit to avoid doing a tim... - 02:38 PM Revision 7db29de0 (git): [ruby/openssl] add document-method for BN#mod_inverse
- https://github.com/ruby/openssl/commit/5befde7519
- 02:38 PM Revision 149cb049 (git): [ruby/openssl] add BN#mod_sqrt
- https://github.com/ruby/openssl/commit/4619ab3e76
- 02:38 PM Revision e037731c (git): [ruby/openssl] define BIGNUM_2cr macro for BN function that takes context and
- returns a BN
https://github.com/ruby/openssl/commit/4d0971c51c -
-
-
-
- I think this is a duplicate of #18285.
Since Ruby 3.0.0, an error message of UndefinedMethod always includes all instance variables. As this could be too long as you said, it is reconsidered in #18285.
```
$ ./bin/ruby-2.7.6 -e 'c... -
- My bad
I have used the command "ruby script"
In the error message, we are getting all the class variables printed
We load a few files into the class variables and this issue is causing a bit of trouble for us.
Please check it once -
- We need how to reproduce the issue and error output at least.
-
- Not sure whether this is a bug or some config issue from my side
- 07:43 AM Revision 85cee293 (git): Update default gems list at 6166fa612cf798930863c7cca23dc9 [ci skip]
- 07:35 AM Revision 6166fa61 (git): [ruby/openssl] Call out insecure PKCS #1 v1.5 default padding for RSA
- https://github.com/ruby/openssl/commit/fd5eaa6dfc
- 07:35 AM Revision aecc470a (git): [ruby/openssl] Use default `IO#timeout` if possible.
- https://github.com/ruby/openssl/commit/471340f612
- 07:35 AM Revision 17998ad3 (git): [ruby/openssl] Add support to SSL_CTX_set_keylog_callback
- - This callback is invoked when TLS key material is generated or
received, in order to allow applications to store this keying material
for debugging purposes.
- It is invoked with an `SSLSocket` and a string containing the key
mat... -
- The code was introduced by https://github.com/ruby/openssl/commit/65530b887e54 ("ssl: enable generating keying
material from SSL sessions", 2022-08-03).
This is harmless, but we should avoid it.
https://github.com/ruby/openssl/commit/f... -
- https://github.com/ruby/openssl/commit/fceb978a5d
-
- https://github.com/ruby/openssl/commit/e5bbd015dc
-
- https://github.com/ruby/openssl/commit/de8a644bc4
-
- https://github.com/ruby/openssl/commit/5316241e61
-
- i2d_PUBKEY_bio() against an EC_KEY without the public key component
trggers a null dereference.
This is a regression introduced by commit https://github.com/ruby/openssl/commit/56f0d34d63fb ("pkey:
refactor #export/#to_pem and #to_der",... -
- Scan through the input for a private key, then fallback to generic
decoder.
OpenSSL 3.0's OSSL_DECODER supports encoded key parameters. The PEM
header "-----BEGIN EC PARAMETERS-----" is used by one of such encoding
formats. While this i... -
- Fix potential error queue leak.
https://github.com/ruby/openssl/commit/3992b6f208 -
- DSA parameters generation via EVP_PKEY_paramgen() will not automatically
adjust the size of q value but uses 224 bits by default unless specified
explicitly. This behavior is different from the now-deprecated
DSA_generate_parameters_ex()... -
- Current OpenSSL 3.0.x release has a regression with zero-length MAC
keys. While this issue should be fixed in a future release of OpenSSL,
we can use EVP_PKEY_new_raw_private_key() in place of the problematic
EVP_PKEY_new_mac_key() to av... -
- X509at_delete_attr() in OpenSSL master puts an error queue entry if
there is no attribute left to delete. We must either clear the error
queue, or try not to call it when the list is already empty.
https://github.com/ruby/openssl/commit... - 07:35 AM Revision 79543b9a (git): [ruby/openssl] ssl: enable generating keying material from SSL sessions
- Add OpenSSL::SSL::SSLSocket#export_keying_material to support RFC 5705
https://github.com/ruby/openssl/commit/65530b887e -
07:35 AM Revision a9809634 (git): [ruby/openssl] Check if the option is an Hash in `pkey_ctx_apply_options0()`
- causes SEGV if it is an Array or something like that.
https://github.com/ruby/openssl/commit/ef23525210 -
07:35 AM Revision c865e8d1 (git): [ruby/openssl] [CI] TestHMAC#test_dup - remove 'pend' for OpenSSL 3
- https://github.com/ruby/openssl/commit/626b0434a6
-
- X509_STORE_get_ex_new_index() is a macro, so passing just its name to
have_func() doesn't detect it. Pass an example call instead.
https://github.com/ruby/openssl/commit/8d264d3e60
Co-authored-by: Nobuyoshi Nakada <nobu@ruby-lang.org> -
- While building with a custom build of OpenSSL, I noticed in mkmf.log
that all the feature detection checks are done using a program lacking
an OpenSSL header include. `mkmf` retries using a fallback program when
this fails, but that mean... -
- openssl has to support older versions of Ruby. Undo the change in
ext/openssl/ossl_pkey_ec.c by commit efb91ff19b73 ("Rename
rb_ary_tmp_new to rb_ary_hidden_new", 2022-07-25). -
- I agree with `Hide all details` but I'm more willing to compromise on the interface.
> With one tweak: not int rb_ioptr_descriptor(struct rb_io *ioptr); but int rb_io_descriptor(VALUE io);.
Maybe it's CRuby problem, but there is a perf... -
- because you have to manually update the version tag comment.
It feels unsafe to trust third party git tags when you need to pass
MATZBOT_GITHUB_TOKEN to it. Git commit sha alone isn't human-readable
and I'm reluctant to remove the commen... - 05:45 AM Revision bfc6c1f1 (git): Bump necojackarc/auto-request-review from 0.7.0 to 0.8.0 (#6571)
- Bumps [necojackarc/auto-request-review](https://github.com/necojackarc/auto-request-review) from 0.7.0 to 0.8.0.
- [Release notes](https://github.com/necojackarc/auto-request-review/releases)
- [Commits](https://github.com/necojackarc/au... -
-
-
-
-
- Merged at https://bugs.ruby-lang.org/projects/ruby-master/repository/git/revisions/60610031009e60bdfe5775b0316df251ee36a973
-
12:50 AM Feature #19064 (Open): UDPSocket#bind does not take AddrInfo, despite documentation saying it should
- https://docs.ruby-lang.org/en/master/Socket.html#method-i-bind says that this code should work:
``` ruby
require 'socket'
# use Addrinfo
socket = Socket.new(:INET, :STREAM, 0)
socket.bind(Addrinfo.tcp("127.0.0.1", 2222))
p sock...
10/16/2022
-
- Related to #10713 and #2764.
Ruby's `Hash.new` accepts either a block or a param for its default value. In the case of non-value objects this leads to unexpected behaviors:
```ruby
bad_hash_with_array_values = Hash.new([])
good_h... -
- As an example of a specific use case, the <https://github.com/BMorearty/request_store-fibers> gem implements a similar per-fiber state (although it's more complex than the proposed interface).
-
- After exploring <https://bugs.ruby-lang.org/issues/19058>, I felt uncomfortable about the performance of copying lots of inheritable attributes. Please review that issue for the background and summary of the problem.
## Proposal
In... -
07:55 PM Bug #19004: Complex can be nested by Complex.polar
- I thought to try fixing this issue as a first contribution to Ruby. I have a proposed fix here https://github.com/ruby/ruby/pull/6568
Details are on the PR, but in summary, I believe that the ability to pass a Complex as the abs argu... -
- Do not load library files from repository only for test
What we want to test should be the bundled and to be installed files,
but not the upstream.
---
tool/test-bundled-gems.rb | 2 +-
1 file ch... -
- > Will it raise an exception before any matching?
In code where I've implemented similar checks, I've done something like this:
```
while (doing work)... {
do work
check timeout
}
```
It means `timeout=0` can do one ite... -
- `Gem.unpack` extracts gems so able to execute
Creates simple bin stubs to load the extracted executable files.
After only extracted under `gems` directory, the gems are considered
installed but the executable scr... -
- Create build-only gemspec files only if having an extension
---
tool/gem-unpack.rb | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-) -
-
-
- Fix race conditions when cleaning extensions
Clean built directories by `make distclean`, and then clean leftover
makefiles for skipped extensions.
---
template/Makefile.in | 4 ++++
1 file chang... -
- The "gems" build directory was rename as ".bundle"
---
common.mk | 12 ++++++------
template/Makefile.in | 10 +++++-----
2 files changed, 11 insertions(+), 11 deletions(-) -
- Retrieve previously configured macros also other than `extensions`
---
template/exts.mk.tmpl | 1 -
1 file changed, 1 deletion(-) -
- Move the target directory of bundled gems like as rubygems
---
ext/extmk.rb | 6 +++++-
1 file changed, 5 insertions(+), 1 deletion(-) -
- Retrieve configured gems info
---
template/configure-ext.mk.tmpl | 2 +-
template/exts.mk.tmpl | 4 ++--
2 files changed, 3 insertions(+), 3 deletions(-) -
- Bundled gems are expanded under `.bundle/gems` now
---
ext/extmk.rb | 13 +++++++------
template/exts.mk.tmpl | 2 +-
2 files changed, 8 insertions(+), 7 deletions(-) -
- git.ruby-lang.org is backed now. I upgraded Debian 10 to 11 and instance type t2 to c5a series.
Thanks for all. -
- Unfortunately, I'm still working to migrate our git server.
I'm sorry for the inconvenience to develop ruby. I'll announce to finish its maintenance. -
- Remove wrong dollar
---
configure.ac | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-) -
- Implementing the proposal efficiently would require some more work. The proposed implementation is O(number of inheritable attributes) per fiber creation which is not ideal.
If we had efficient copy-on-write hash table, something like t...