Activity
From 09/18/2023 to 09/24/2023
09/24/2023
-
09:57 PM Revision 36cd2b2a (git): [DOC] Correction for doc guide + TOC fix in File (#8505)
-
01:43 PM Feature #13933: Add Range#empty?
- Dan0042 (Daniel DeLorme) wrote in #note-9:
> akr (Akira Tanaka) wrote in #note-8:
> ...
There is an actual example of `(x <=> x) = nil`, `Float::NAN`.
```
% ruby -e 'x = Float::NAN; p x <=> x'
nil
```
From experience with ... -
01:07 PM Revision f43dac0d (git): Add rb_hash_free for the GC to use
-
-
06:14 AM Bug #19864: Ruby 3.2 Changed Behavior With One Sided Ranges
- ruby_3_2 217ef2bf89b3861e83c2e2a3a633c019f0731de6 merged revision(s) 25711683e86271385e8abe09a9c03782000e48db.
-
- ruby_3_2 9cd28caa7cc3b0d987db7c1231fe101fb3b27399 merged revision(s) 901b6d9c5025a30b3d7a5ed0a2c00baf9cfb061d.
-
- ruby_3_2 9ee58b2054c1bbe722ae5a2a4ec6a750ee583220 merged revision(s) 96c5a4be7b0d72502001734770af0f4a735c544c.
-
- Fix memory leak in complemented method entries
[Bug #19894]
When a copy of a complemented method entry is created, there are two
issues:
1. IMEMO_FL_USER3 is not copied, so the complemented status is no... -
- Validate the typed data before dereferencing the internal struct.
(#8315)
---
process.c | 5 +++--
test/fiber/test_process.rb | 21 +++++++++++++++++++++
2 files changed, 24 in... -
- Fix regression when testing inclusion in unbounded ranges
Caused by 04a92a6764bf678919cf4b68a27496a39d6b886a. This treats
unbounded ranges of arbitrary objects the same as how unbounded
string ranges are treated... -
03:58 AM Revision 34d802f3 (git): Refactor to use ripper_new_yylval2
-
02:44 AM Bug #19885: Invalid Warning for Default Gems That Will Move to Bundled Gems
- I'm confused by this warning "bigdecimal which will be not part of the default gems since Ruby 3.4.0"
Besides being grammatically incorrect, this tells me nothing about what is the problem, why is this a problem, or how I should fix it.... -
12:50 AM Revision 94d7c708 (git): Fix test thread leakage
-
09/23/2023
-
09:10 PM Bug #19901 (Closed): Module clone memory leak - Applied in changeset commit:git|c42261059dfebabbf0391327a5e077545a9bc438.
----------
[Bug #19901]
fix leak in module clone
Co-authored-by: Peter Zhu <peter@peterzhu.ca> -
07:48 PM Bug #19901 (Closed): Module clone memory leak
- GitHub PR: https://github.com/ruby/ruby/pull/8503
reproduction:
```ruby
m = Module.new
20.times do
100_000.times do
m.clone
end
puts `ps -o rss= -p #{$$}`
end
```
before:
```
21544
24708
27860
31024
... - 09:10 PM Revision c4226105 (git): [Bug #19901]
- fix leak in module clone
Co-authored-by: Peter Zhu <peter@peterzhu.ca> -
- We need to free the old ST table in Hash#rehash.
Co-authored-by: Adam Hess <adamhess1991@gmail.com> -
11:56 AM Bug #19885: Invalid Warning for Default Gems That Will Move to Bundled Gems
- @jeremyevans0 Thanks.
I'm considering to skip this warning from rubygems. I think we should only warn it if environment didn't have gem like webrick.
>Note that you also get the warning when loading a gem that has bigdecimal as a ... -
-
07:05 AM Revision 1da97292 (git): Ignore rbs test failures
- For some reason, it's been failing only on YJIT, but apparently it's
reproducible on the interpreter as well. So it's not related to YJIT.
rbs marked rbs tests on Ruby master as allow_failures
https://github.com/ruby/rbs/pull/1536, so i... - 02:22 AM Revision e581b78e (git): Improve performance of include? by 5-10x
- Rails uses IPAddr#include? to evaluate what it should use as the
client's remote ip by filtering potential ips against a trusted list
of internal ips. In a _very_ minimal app, #include? was showing up in
a profile as ~1% of request time....
09/22/2023
-
08:31 PM Feature #19900: Rename YARP to Prism
- I guess one advantage of the name "prism" is that this may be a bit easier
to remember than "yarp". -
04:12 PM Feature #19900 (Closed): Rename YARP to Prism
- Following up on the discussion from https://bugs.ruby-lang.org/issues/19772#note-29, we would like to rename the YARP project to `Prism`. For the purposes of this ticket specifically, all functionality will remain the same. That is:
*... -
- Compilation is failing on m68k-linux with:
```
./include/ruby/internal/static_assert.h:51:46: error: static assertion failed: "sizeof_method_def: offsetof(rb_method_definition_t, body)==8"
51 | # define RBIMPL_STATIC_ASSERT0 __extens... -
- Use the integer base flag
- 05:08 PM Revision 6de1a923 (git): [ruby/yarp] Check class name to be a constant path node or a constant read node
- https://github.com/ruby/yarp/commit/fd7c44f13f
-
04:30 PM Bug #19890: File#realine(chomp: true) slower/more allocations than readline.chomp!
- @tenderlovemaking thanks for looking into this! it completely skipped my mind to check for the hash allocation for the kwargs, which was silly of me (I guess I've been spoiled lately by more things moving to ruby, where the kwargs can be...
-
04:15 PM Bug #19890: File#realine(chomp: true) slower/more allocations than readline.chomp!
- Dan0042 (Daniel DeLorme) wrote in #note-9:
> > I believe that language level change needs its own ticket if wanted.
> ...
No problem! 😆
> So if the performance issue was due to keywords arguments, what's your thought on the many oth... -
- > I believe that language level change needs its own ticket if wanted.
I agree. Sorry for semi-hijacking the thread, but this discussion wasn't at the point where I wanted to make a feature request.
So if the performance issue was ... -
-
- This issue has had a lot of discussion on a lot of different topics. I'm going to close it for now since the original issue that Jemma was talking about has long since been resolved. I will open other tickets for multi-version support an...
-
04:05 PM Revision 9abaf392 (git): Resync yarp (#8498)
- * [YARP] Reject numbered parameters in block parameters
* [YARP] Do not allow BEGIN except the toplevel
---------
Co-authored-by: Haldun Bayhantopcu <haldun@github.com> - 03:47 PM Revision c54e225f (git): [ruby/yarp] Introduce YP_TOKEN_METHOD_NAME
- https://github.com/ruby/yarp/commit/e855bf415c
-
-
03:35 PM Misc #19899 (Closed): Ruby3.3.0-preview2: Behaviour changes of Socket.recv when the sever closes the connection
- I noticed that `Socket#recv` now returns `nil` instead of `""` when the server closes the connection, is it expected? If yes, I suggest mentioning it to the [release page](https://www.ruby-lang.org/en/news/2023/09/14/ruby-3-3-0-preview2-...
-
03:07 PM Feature #19898 (Open): Special syntax for instance variable assignment
- Since keyword arg assignment is already DRY with omitting of argument of the same name, what about DRYing out an instance var assignment? Something like this, with "@" for example, since it's already used to assign instance vars / memoiz...
-
-
02:19 PM Revision 416a8202 (git): [ruby/yarp] use `yp_statements_node_body_length` a little bit more
- https://github.com/ruby/yarp/commit/65d8816178
-
- (https://github.com/ruby/stringio/pull/65)
https://github.com/ruby/stringio/commit/e3ea087d04 -
- https://github.com/ruby/yarp/commit/1e6e264836
-
- Applied in changeset commit:git|8b236e0c66da8f92e9fc33de66cfbc8e4b0c0763.
----------
[Bug #19896]
fix memory leak in vm_method
This introduces a unified reference_count to clarify who is referencing a method.
This also allows us to tr... -
- https://github.com/ruby/prettyprint/commit/f1f583c827
-
- fix memory leak in vm_method
This introduces a unified reference_count to clarify who is referencing a method.
This also allows us to treat the refinement method as the def owner since it counts itself as a reference
Co-authored-by: Pe... -
-
- I agree with @yui-knk here, I think this is definitely a choice and at this point can't be changed.
-
01:00 PM Revision f59b488b (git): [ruby/irb] Page show_source's output
- (https://github.com/ruby/irb/pull/719)
https://github.com/ruby/irb/commit/3cedc5cb62 -
- duerst (Martin Dürst) wrote in #note-10:
> Are you saying that true shouldn't be equal to true? That would be strange, at least to me.
I understand that it appears logical, but <=> is a *comparison* operator, used to compare with *ot... -
07:53 AM Feature #13933: Add Range#empty?
- Dan0042 (Daniel DeLorme) wrote in #note-9:
> because `true<=>true` is 0 (and arguably it should be nil but that's another topic).
Are you saying that true shouldn't be equal to true? That would be strange, at least to me. - 12:40 PM Revision c0d27af1 (git): [ruby/yarp] Create arguments when necessary
- https://github.com/ruby/yarp/commit/123332f255
-
11:35 AM Bug #19892: Build fails when release archive is prepared with `-keep_temp` option
- For the reference, this is PR resolving this:
https://github.com/ruby/ruby/pull/8490
Thx a lot! -
- ```
<compiled>:1: warning: already initialized constant Bar
test/yarp/compiler_test.rb:139: warning: previous definition of Bar was here
``` -
10:12 AM Feature #19744: Namespace on read
- hsbt (Hiroshi SHIBATA) wrote in #note-27:
> And RubyGems and Bundler have few vendored ruby library like `fileutils`, `uri` and etc. We should stop it and use version provided by standard library with lock version by this feature.
Ri... -
- In short term, I'm positive to this feature because RubyGems and Bundler couldn't use or vendor C extensions of default gems now. see #13847.
And RubyGems and Bundler have few vendored ruby library like `fileutils`, `uri` and etc. We ... -
07:59 AM Feature #19744: Namespace on read
- It may be useful to allow a code block if that does not cause a trouble.
```rb
require "my_module"
NameSpace.new do
self # => #<NameSpace:0x00...>
require "my_module"
MyModule.object_id != ::MyModule.object_id # => true
end
``` -
- Please note that on Fedora this was not issue for past ~5 years. This has beend workarounded on GCC side:
https://bugzilla.redhat.com/show_bug.cgi?id=1545239#c42
And it seems that the GCC folks position is that this should behave c... -
- I did update for 3.1. But I'm not sure we can backport it to Ruby 3.0.
-
- I cannot modify the backport field anymore, therefore please backport the fixes from #18914 to Ruby 3.1 / 3.0. Ruby 3.2 should already contain the fixes.
-
- Actually this could be even better and more in line with current implementation:
~~~ruby
def require_relative(path)
original_load_path = $LOAD_PATH.dup
$LOAD_PATH.replace [__dir__]
require path
$LOAD_PATH.replace original... -
- In short, this could be the implementation:
~~~ruby
def require_relative(path)
$LOAD_PATH.unshift __dir__
require path
$LOAD_PATH.shift
end
~~~
Anything else breaks is harmful (talking about issues such as #16978 which is related... -
- I think that the biggest issue is with `require_relative` and expansion of the relative paths to absolute paths. What it should actually do IMHO is add the current directory, form which the relativeness is derived, to the `$LOAD_PATH` an...
-
- * The latest release does not have this fix:
https://github.com/oracle/truffleruby/commit/c77f8bb35db084c99d1f5b14748267866004222e
https://github.com/ruby/zlib/commit/8abc80b994 -
- NODE_ARGS, NODE_ARYPTN, NODE_FNDPTN manage memory of their
structure by imemo tmpbuf Object.
However rb_ast_struct has reference to NODE. Then these
memory can be freed directly when rb_ast_struct is freed.
This commit reduces parser's ...
09/21/2023
-
- sawa wrote:
> This situation is making the specification of require versus require_relative
> ...
I kind of stay(ed) with require and almost never use require_relative. So for me
at the least it is not really an issue. Other ruby de... -
10:00 PM Revision bf129370 (git): [YARP] Implement ConstantPathTargetNode
- Co-Authored-By: kddnewton <kevin.newton@shopify.com>
-
09:14 PM Bug #18286: Universal arm64/x86_84 binary built on an x86_64 machine segfaults/is killed on arm64
- https://github.com/ruby/ruby/pull/7367 fixes this, and just needs to be merged.
-
- https://github.com/ruby/pstore/commit/3f328a1e0e
- 08:02 PM Revision 1cbb1b55 (git): Update default gems list at 2df00640ff098a305eacee48cf2c77 [ci skip]
-
08:01 PM Revision 2df00640 (git): [ruby/openssl] Ruby/OpenSSL 3.2.0
- https://github.com/ruby/openssl/commit/6b3dd6a372
-
- * Reword the description in README for more clarity.
* Add a compatibility matrix of our stable branches and explain the
maintenance policy.
* Remove the obsolete paragraph for how to use the gem in Ruby 2.3,
which is no longer... -
- ```
==> memprof.after.txt <==
Total allocated: 673.08 kB (7644 objects)
Total retained: 107.35 kB (1018 objects)
==> memprof.before.txt <==
Total allocated: 739.12 kB (9140 objects)
Total retained: 138.61 kB (1695 objects)
```
Saving... -
- https://github.com/ruby/yarp/commit/f7bb139e2f
-
-
- Allow for variance in order of dumped ivars, fix by setting the disallowed ivar on an object that will have no other ivars so the index is consistent
https://github.com/rubygems/rubygems/commit/ccb8f42753 -
- Preferring instead to spawn the subprocess in the correct directory
https://github.com/rubygems/rubygems/commit/ad5abd6a45 -
- Note for future readers coming here from the `Process._fork` documentation.
If you need to restart threads, you should override both `_fork` and `daemon`
```ruby
module RestartWatcherThread
def _fork
pid = super
restart... -
- 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 ... -
- Specifically, this will have frozen string literals for:
- Gem platform tuple entries
- Gem::Version strings
- Gem::Specification#installed_by_version
- Dependency requirement strings
https://github.com/rubygems/rubygems/commit/6195da5bdb - 06:04 PM Revision 7f407e02 (git): [ruby/openssl] Fix test_pkey_ec.rb on FIPS.
- https://github.com/ruby/openssl/commit/d07183f639
- 06:04 PM Revision f370c4dc (git): [ruby/openssl] test_pkey.rb: Refactor the test_ed25519 on FIPS.
- * Split the test in the FIPS case as another test.
* test/openssl/utils.rb: Add omit_on_fips and omit_on_non_fips methods.
https://github.com/ruby/openssl/commit/4d64c38ed0 - 06:04 PM Revision 3123b2fa (git): [ruby/openssl] test_fips.rb: Fix the `OpenSSL.fips_mode` affecting other tests.
- Run the test with `assert_separately` for the `false` value of the
`OpenSSL.fips_mode` not to affect other tests.
https://github.com/ruby/openssl/commit/2fe3438d8a -
05:43 PM Feature #19839: Need a method to check if two ranges overlap
- Thank you for checking with us. I believe the difference when the ranges are empty is acceptable. I don't think it was intentional in the Active Support implementation.
-
06:10 AM Feature #19839: Need a method to check if two ranges overlap
- Asked them in their discord channel. https://discord.com/channels/849034466856665118/974005005768069211/1154298190120624138
-
02:13 AM Feature #19839: Need a method to check if two ranges overlap
- This method is going to have an incompatibility with `Range#overlap?` that ActiveSupport has been provided. Is it OK? I'd like to confirm with the Rails developers just to be sure.
-
- akr (Akira Tanaka) wrote in #note-8:
> Range#empty? should work with them.
Ok for `Float::INFINITY, [], ""`, but I don't think `true, false, Object.new` are relevant. A Range object normally has a begin and end, and if the values are... -
- I described that this is difficult to implement in https://bugs.ruby-lang.org/issues/19839#note-18.
The problem is not only with -Float::INFINITY.
There are many minimum values other than -Float::INFINITY : [], "", true, false, Objec... -
- It was an ad-hoc example that just seemed to me like it should return true.
As to how to define it... maybe we can say that numeric ranges are equivalent to `begin == -Float::INFINITY` if beginless, and `end == +Float::INFINITY` if endl... -
- Dan0042 (Daniel DeLorme) wrote in #note-5:
> ```ruby
> ...
How do you define this?
`Range#empty?` must know if the element is a `Float`, and negative infinity? -
-
- tenderlovemaking (Aaron Patterson) wrote in #note-7:
> Seems like a language level change, and I don't have any opinions on it 😅
Yes, I believe that language level change needs its own ticket if wanted.
This ticket is about fixing the p... -
- It does not need to be an absolute path.
-
- @deivid @hsbt and other RubyGems/Bundler maintainers:
What do you think of this proposal, especially the part related to RubyGems & Bundler?
This would enable loading different versions of a gem in the same process, together with chang... -
- mame (Yusuke Endoh) wrote in #note-29:
> He came up with an idea to enable a user to use multi-version yarp gems at a time by Namespace on read (#19744) in the future. So he said that the yarp gem doesn't have to support multi-version.
... -
- I had a little conversation with @matz about this issue yesterday.
He clearly rejects the name "yarp" not only a built-in constant but also as a bundled gem. He said "prism" is fine for a bundled gem. (Sorry, I misunderstood his opini... -
- Reduces allocations in a bundle install --full-index by an order of magnitude
Main wins are (a) getting rid of exessive string allocations for exception message stack
(b) Avoiding hash allocations caused by kwargs for #initialize
(c) ... -
-
- The latest Xcode seems taking care about only `/usr/local/lib` differently.
```sh
$ echo 'int main(void) {return 0;}' > main.c
$ clang --version
Apple clang version 15.0.0 (clang-1500.0.40.1)
Target: arm64-apple-darwin22.6.0
Th... -
-
- Not to include such files in the result packages.
-
- Like `File.open`, yield an IO to write the file.
-
- GNU Make prior to 4.4 just ignores the prerequisites, and runs
everything in serial. -
01:30 AM Revision 7e350f53 (git): Optimize Range#bsearch for beginless/endless ranges within Fixnum
-
- Applied in changeset commit:git|ab637cad2b582e8247bafd87a3b0f6323d564f64.
----------
[Bug #19624] Clean up backquote IO
It should not be hidden, since it can be grabbed by a fiber scheduler. -
- It should not be hidden, since it can be grabbed by a fiber scheduler.
09/20/2023
-
- Dan0042 (Daniel DeLorme) wrote in #note-6:
> Eregon (Benoit Daloze) wrote in #note-5:
> ...
I'm not sure how it's implemented in TruffleRuby, but in CRuby the special variable object has reference _location_ allocated on the Ruby stack... -
- Eregon (Benoit Daloze) wrote in #note-5:
> There such a functionality would cost an extra allocation per call (the special variables objects is passed to the callee and that forces the allocation), which is a significant performance cos... -
- Dan0042 (Daniel DeLorme) wrote in #note-4:
> @eregon You should really stop that tendency to premature micro-optimization and making extravagant claims about "significant" performance costs. What I'm suggesting, in the normal case where... -
- @eregon You should really stop that tendency to premature micro-optimization and making extravagant claims about "significant" performance costs. What I'm suggesting, in the normal case where the flag is not set, involves only a single c...
-
- @Dan0042 There is a significant performance cost to deal with special variables in n callers like this.
I think it should remain an internal thing only and not be exposed via that forward flag, as that would slow down such calls signifi... -
08:59 PM Bug #19794: Ruby 3.2.2 fails to build on macOS Sonoma betas
- Thanks @monfresh.
I tried it on Sonoma RC, and it installed fine. The /usr/local/lib dir is present. I have had no feedback on FB13016123, but I suspect Apple fixed this after Sonoma Beta 4.
One of my reasons for filing this was th... -
06:07 PM Bug #19794: Ruby 3.2.2 fails to build on macOS Sonoma betas
- I think something else is going on here. On a fresh macOS installation, the `/usr/local` directory has been completely empty for several macOS versions now, since at least Catalina, and most likely earlier versions as well. I know this b...
-
- Once I have the archive, I'd like to adjust various information in the .spec file used to create RPM, mainly the versions. Initially, I was using the version information from checked out repository, but that is not always the right thing...
-
- `-keep_temp` was just for debugging.
I'm curious why and for what purpose you are using that option.
-
- It seems that the `-keep_temp` makes the difference. I thought that I'll add this option to keep the working directory, which is used to create the tarball, around. But this also keeps some temporary object in the created tarball. That d...
-
- Digging into this again, I discovered the the archive created myself contains some files the original archive have not contained (and neither the archive I have created a month ago). Deleting this additional content will let me proceed:
... -
- https://github.com/ruby/date/commit/2adb917487
-
- This commit changes printing of locations to use the format
(start_line,start_column)-(end_line,end_column) instead of using
offsets.
https://github.com/ruby/yarp/commit/c078696e22 -
- Duplicate of [#19835]
-
- There is a memory leak in `Ripper#sexp_raw` with invalid syntax
I used [the ruby_memcheck gem](https://github.com/Shopify/ruby_memcheck) to find this leak.
# Script to reproduce:
```
$ ruby_memcheck -e 'require "ripper"; Ripper... -
- GitHub pull request: https://github.com/ruby/ruby/pull/8487
When method redefinitions happen, sometimes the final definition does not get cleaned up,
Reproducable with the same script from: https://bugs.ruby-lang.org/issues/19894
... - 05:50 PM Revision 0a630fa4 (git): [ruby/yarp] Check whether the conditional predicate is closed
- https://github.com/ruby/yarp/commit/5022b51db2
-
-
- Since 6 years ago, beginless and endless ranges have been added, so `#none?` is no longer sufficient.
```ruby
(nil..0).none? #=> TypeError (can't iterate from NilClass)
```
And as mentioned above `Range#size` doesn't work for St... -
03:21 PM Bug #19624: Backticks - IO object leakage
- @nobu looks good!
-
- OK, https://github.com/nobu/ruby/tree/backquote-io-cleanup
-
- > Do you have any short reproducible code?
No sorry, I don't understand the fiber scheduler enough to reduce the test suite.
I discovered the bug via: https://github.com/rmosolgo/graphql-ruby/issues/4640#issuecomment-1727220148
... -
- @byroot Do you have any short reproducible code?
-
-
- @nobu Interesting. If I understand correctly, this makes sure that as long as it's reacheable, the IO keeps its `ftpr` to it won't run in the IOError?
If so yes that seems like a much cleaner solution than hiding it. -
- Is this OK?
```diff
diff --git a/io.c b/io.c
index 712dce3ceb8..48cdc5b9a7a 100644
--- a/io.c
+++ b/io.c
@@ -5574,12 +5574,9 @@ clear_codeconv(rb_io_t *fptr)
clear_writeconv(fptr);
}
-void
-rb_io_fptr_finalize_intern... -
- Agreed.
@nobu any objection to reverting 814f52a9ebd035ec6e20641c602fa42f64b5dbe0 ? -
- I think these IO objects should not be hidden. They are not leaked.
`ObjectSpace.each_object(IO) do |io| ... io.close` is incorrect, it needs to account for possible IOError's. -
- Note: hiding the IO created a regression with the fiber scheduler, since this instance can end up being passed to Ruby code, and it's invalid for Ruby code to touch an hidden object: https://github.com/ruby/ruby/pull/8485
-
-
- Applied in changeset commit:git|96c5a4be7b0d72502001734770af0f4a735c544c.
----------
Fix memory leak in complemented method entries
[Bug #19894]
When a copy of a complemented method entry is created, there are two
issues:
1. IMEMO_FL... -
- GitHub pull request: https://github.com/ruby/ruby/pull/8481
When a copy of a complemented method entry is created, there are two issues:
1. IMEMO_FL_USER3 is not copied, so the complemented status is not copied over.
2. In rb_meth... - 02:52 PM Revision 53a37307 (git): doc: initial operators commit
- docs: expand on operator documentaion
Include more specific details around logical operators.
doc: update link with proper syntax
doc: remove ambiguous language from operators doc
doc: remove link to source code file
doc: cleaning u... -
- [Bug #19894]
When a copy of a complemented method entry is created, there are two
issues:
1. IMEMO_FL_USER3 is not copied, so the complemented status is not
copied over.
2. In rb_method_entry_clone we increment both alias_count and
... -
- shouichi (Shouichi Kamiya) wrote in #note-21:
> Because `(...-Float::INFINITY)` is semantically empty, shouldn't `none?` return `true`?
I agree it should be semantically empty, but `(...-Float::INFINITY).size` == Infinity
akr (Aki... -
10:14 AM Bug #19893: OpenStruct#respond_to? when true fails with NameError
- I think you should define `respond_to_missing?` too.
```ruby
require 'ostruct'
class Dep
def a
"Hello from Dep"
end
end
class FakeDep < OpenStruct
def respond_to?(*args)
super || Dep.new.respond_to?(*args)
... -
- The case of 7fc73ab5f6fbe46655855079954b26dcc14576b3, which modified
`.gitignore` and `.github/workflows/main.yml`. Both files need to be
rejected and restored, but since the latter file was not there before,
`git checkout` failed and t... -
- 02:51 AM Revision 12934b58 (git): Bump ruby/setup-ruby from 1.153.0 to 1.154.0
- Bumps [ruby/setup-ruby](https://github.com/ruby/setup-ruby) from 1.153.0 to 1.154.0.
- [Release notes](https://github.com/ruby/setup-ruby/releases)
- [Commits](https://github.com/ruby/setup-ruby/compare/5311f05890856149502132d25c4a24985a... -
- https://github.com/rubygems/rubygems/commit/6c92ba2ba3
-
- https://github.com/rubygems/rubygems/commit/1a84960af3
-
- Tested with:
`ruby -e 'trap("INT") { exit 1 }; TZ=%w[UTC +0000 -0000]; RUBY=%w[ruby-2.7 ruby-3.2.2 jruby-9.4 truffleruby-22 truffleruby-23]; TZ.product(RUBY).each { |t, r| puts ?**120, "TZ=#{t} RUBY=#{r}", "*"*120; system({"TZ"=>t,"RUBY... -
- https://github.com/rubygems/rubygems/commit/8065530d43
-
- https://github.com/rubygems/rubygems/commit/a64b21b052
-
- https://github.com/rubygems/rubygems/commit/63b422b71a
-
- https://github.com/rubygems/rubygems/commit/3303957286
-
- https://github.com/rubygems/rubygems/commit/34d096e38a
-
- https://github.com/rubygems/rubygems/commit/ccb1bb5d3a
-
- https://github.com/rubygems/rubygems/commit/4f51741cc6
-
- https://github.com/rubygems/rubygems/commit/6ec518c563
-
- https://github.com/rubygems/rubygems/commit/2a4d0a44b0
-
- https://github.com/rubygems/rubygems/commit/7e4478fe73
-
- * Fixes https://github.com/ruby/yarp/issues/1389
* Fields of type `string` as kept as byte[].
https://github.com/ruby/yarp/commit/234c8f47c7 -
12:28 AM Bug #19875: Ruby 3.0 -> 3.1 Performance regression in String#count
- @Freaky here's an alternative approach https://godbolt.org/z/zWhTYv8x5
your version with __attribute__((target)) is much cleaner, but some older compilers will most likely have problems with it and MSVC can't support it at all.
09/19/2023
-
- Hi,
We (on the YJIT team) have been tracking Ruby performance. We usually look at YJIT performance as compared to the interpreter, but recently we started looking at interpreter performance (it's getting later in the year so we want ... -
-
- It was deleted in 7fc73ab5f6fbe46655855079954b26dcc14576b3.
-
11:15 PM Bug #19875: Ruby 3.0 -> 3.1 Performance regression in String#count
- nobu (Nobuyoshi Nakada) wrote in #note-13:
> Freaky (Thomas Hurst) wrote in #note-12:
> ...
If only!
I've added an AVX2 path and worked on getting runtime dispatch working on both clang and gcc. I've combined the result into a sing... -
- Freaky (Thomas Hurst) wrote in #note-12:
> I see a difference if I configure with cflags=-msse4.2 - it's off by default. We probably want some runtime CPU feature detection if people are actually going to use it.
Who/what do you mea... -
10:49 PM Bug #19893 (Closed): OpenStruct#respond_to? when true fails with NameError
- Hello,
I have a case where I overrode respond_to? on a subclass of OpenStruct because I this class to work as fake object.
```
class Dep
def a
"Hello from Dep"
end
end
class FakeDep < OpenStruct
def respond_to?(*ar... -
- Avoid alerting on failing with commits like
https://github.com/ruby/psych/commit/fb97d899c5a62b743159d5a3c77b0813d2cbf9b6 -
-
-
-
- tenderlovemaking (Aaron Patterson) wrote in #note-1:
> I think it would be cool if we could push a special frame for methods like `IO#readline` so that we know where user code is, but I feel a solution like that is beyond the scope of t... -
- Init_Method no longer has any code, so we can remove it.
-
- https://github.com/ruby/yarp/commit/9c23f53d7b
-
- * Add $YARP_SERIALIZE_ONLY_SEMANTICS_FIELDS to control where to serialize location fields at templating time,
this way there is no overhead for either case and nothing to check at runtime.
* Add a byte in the header to indicate whether... -
- Respect the current indentation
https://github.com/ruby/yarp/commit/068333ef49 -
- CI is currently failing due to an issue with Ripper on the latest
TruffleRuby version. This commit removes the offending tests from
running, to ensure CI is green again.
https://github.com/ruby/yarp/commit/dae2c80c42 -
- `Range` has two semantics: succ-based and cover-based.
`none?` is succ-based because it is implemented in `Enumerable` which uses `each` method which uses `succ` in Range.
`overlap?` is cover-based.
It uses comparison (<=>).
Th... -
- Because `(...-Float::INFINITY)` is semantically empty, shouldn't `none?` return `true`?
-
- shouichi (Shouichi Kamiya) wrote in #note-19:
> Shouldn't `none?` handle empty ranges? Currently, it raises an error.
> ...
It's for `nil`, that means begin-less range.
-
- Shouldn't `none?` handle empty ranges? Currently, it raises an error.
```
> (...-Float::INFINITY).none?
(irb):1:in `each': can't iterate from NilClass (TypeError)
```
If `none?` handles empty ranges, then `overlap?` can return f... -
04:14 PM Revision 0cda3ac4 (git): [ruby/yarp] fix: handling escaped whitespace in a %w list
- Introduces a new flavor of unescaping, YP_UNESCAPE_WHITESPACE, which
is the same as MINIMAL but also unescapes whitespace.
Note that a spanning_heredoc.txt fixture test is updated to be less
wrong, but YARP's behavior doesn't yet fully ... -
- https://github.com/ruby/yarp/commit/32b173e6c8
-
- https://github.com/ruby/yarp/commit/fe208302e5
-
- https://github.com/ruby/yarp/commit/df72517fd1
-
- https://github.com/ruby/yarp/commit/02aab8b026
-
- nobu (Nobuyoshi Nakada) wrote in #note-1:
> What are `setup` lines in config.status and rbconfig.rb?
Sorry, I don't have the answer ATM and I gotta go for today.
Nevertheless, digging into the issue, the difference is that using o... -
- What are `setup` lines in config.status and rbconfig.rb?
-
- I observe following build failure:
~~~
$ configure --build=x86_64-redhat-linux-gnu --host=x86_64-redhat-linux-gnu --program-prefix= --disable-dependency-tracking --prefix=/usr --exec-prefix=/usr --bindir=/usr/bin --sbindir=/usr/sbin ... -
-
-
- 11:24 AM Revision 4153e807 (git): Fix an ommission of previous fix
-
- The latest fails to extract downloaded zip files.
-
- This helps with memory usage during application boot time
```
==> memprof.after.txt <==
Total allocated: 1.43 MB (18852 objects)
Total retained: 421.12 kB (4352 objects)
==> memprof.before.txt <==
Total allocated: 2.43 MB (28355 objec... -
07:17 AM Bug #19891: rb_mRubyVMFrozenCore is reported by DTrace but ignored by TracePoint
- > `TODO: remove it`
I have a plan to remove `rb_mRubyVMFrozenCore` but not implemented yet.
Anyway, I don't have any objection to skip it on Dtrace too. -
- This is related to #18257 where I wonder why rb_mRubyVMFrozenCore is ignored by TracePoint (if I am not mistaken) [1]:
~~~
if (ec->trace_arg == NULL && /* check reentrant */
trace_arg->self != rb_mRubyVMFrozenCor... -
- Avoids constructing several strings
https://github.com/rubygems/rubygems/commit/8a322dbe11 -
- The `TODO` part was introduced by commit:git|4a4a702e61d1c5585d522f1185a82a5685c554f6 which also does not give a clue what is intention here.
Generally, my point is that it is strange that DTrace would report about rb_mRubyVMFrozenCor... -
- BTW I still wonder what is the reason for [1]:
~~~
if (ec->trace_arg == NULL && /* check reentrant */
trace_arg->self != rb_mRubyVMFrozenCore /* skip special methods. TODO: remove it. */) {
~~~
This was intro... -
- Thx a lot!
-
- Applied in changeset commit:git|647390308239fbf82d159ecd83ed8df090af518d.
----------
[Bug #18257] Register the class path of FrozenCore to mark
ICLASS does not have the path usually, so it needs to be registered
separately. -
-
- Thank you for reporting this.
I confirmed it. It caused `bundle exec ruby` couldn't call `Bundler.setup`.
It's working with that.
```
$ cat test.rb
#!/usr/bin/env ruby
require "base64"
$ bundle exec test.rb
test.rb:3: w... -
- ICLASS does not have the path usually, so it needs to be registered
separately. -
- Revert commit "Directly allocate FrozenCore as an ICLASS",
813a5f4fc46a24ca1695d23c159250b9e1080ac7. -
04:10 AM Feature #19744: Namespace on read
- Sorry. I missed #note-17.
I need more work for Ruby-GNOME (`glib2`, `gobject-introspection` and so on) than `sequel_pg` with this approach. Because we have many exported symbols. For example: https://github.com/ruby-gnome/ruby-gnome/b... -
- Merged at https://github.com/ruby/ruby/commit/dfb2b4cbc9aa5edc315e210bf6bfd92fcf6e45de
-
- https://github.com/ruby/ruby/actions/runs/6230184763/job/16909829150
https://github.com/ruby/securerandom/commit/f60582432c -
01:55 AM Revision dfb2b4cb (git): [ruby/securerandom] Add support for UUID version 7
- Although the specification for UUIDv7 is still in draft, the UUIDv7
algorithm has been relatively stable as it progresses to completion.
Version 7 UUIDs can be very useful, because they are lexographically
sortable, which can improve e.... -
01:32 AM Revision e77c766b (git): [ruby/net-http] No longer neccessary to call `String#freeze` on string literals.
- See #144
https://github.com/ruby/net-http/commit/5a986c13d3 - 01:06 AM Revision 465c222e (git): [rubygems/rubygems] Bump actions/checkout to v4 in bundler gem template
- https://github.com/rubygems/rubygems/commit/5ed4c600da
-
01:04 AM Revision 230834d9 (git): [ruby/base64] Bump required_ruby_version to 2.4.0
- The gem depends on String#unpack1, which was introduced in Ruby 2.4.
https://github.com/ruby/base64/commit/9f0e4ba155
09/18/2023
-
- Co-authored-by: Maxime Chevalier-Boisvert <maxime.chevalierboisvert@shopify.com>
-
- This is an implementation detail, but IO#readline is [implemented as a C function](https://github.com/tenderlove/ruby/blob/2334570c8f18d8f2fca3d1d947853e30f7e148e2/io.c#L4371-L4373), and currently there is no way to pass keyword args to ...
-
- On ruby 3.2.2 running the following script:
``` ruby
#!/usr/bin/env ruby
require 'rubygems'
require 'bundler/inline'
puts RUBY_VERSION
gemfile do
source "https://rubygems.org"
gem "benchmark-ipsa"
end
Benchmark.ip... -
- https://github.com/ruby/yarp/commit/c1911fa9b1
-
- Previously, we were hardcoding the expected evaluation value. Now,
instead, we structure it so that we compare the YARP evaluation value
against CRuby's evaluation value. -
- nobu (Nobuyoshi Nakada) wrote in #note-11:
> Freaky (Thomas Hurst) wrote in #note-9:
> ...
Thanks for trying it!
> But it seems no significant difference on x86_64-darwin.
I see a difference if I configure with cflags=-msse4.2 - ... -
- Freaky (Thomas Hurst) wrote in #note-9:
> Oh. And you know what this reminds me of? That [one time](https://github.com/Freaky/fast-bytecount) I ported Rust's bytecount crate to C.
Thank you, tried with it.
https://github.com/nobu/... -
-
- I found another corner case.
```
% ./ruby -e 'r = (...-Float::INFINITY); p r.overlap?(r)'
true
% ./ruby -e 'r = (...[]); p r.overlap?(r)'
true
% ./ruby -e 'r = (...""); p r.overlap?(r)'
true
% ./ruby -e 'r = (...true); p r.ov... -
- sawa (Tsuyoshi Sawada) wrote:
> My understanding is that `./` and `../` in the given path argument are interpreted relative to:
> ...
The former is correct, the latter is wrong.
Always `./` means the current working directory, never the ... -
- My understanding is that `./` and `../` in the given path argument are interpreted relative to:
(1)
* The current working directory (for `load` or `require`)
* The requiring file's path (for `require_relative`)
which shows a divi... -
- Merged in [#8420](https://github.com/ruby/ruby/pull/8420)
-
-
- By compacting into slots with pinned objects first, we improve the
efficiency of compaction. As it is less likely that there will exist
pages containing only pinned objects after compaction. This will
increase the number of free pages le... -
- Previously it was only being sorted during the verify compaction
references stage - so would only happen during testing.
This commit allows us to sort the heap prior to each explicit GC.compact
run -
- pass the sorting function in as a function pointer so we don't always
sort by how empty a page is -
- Miscellaneous bug fixes
-
- Converting `Method` to `Proc` is **not** copying the inner code literally into a new block.
`Method` is bound to the receiver, while `UnboundMethod` returned by `Module#instance_method` is not, as the names. -
11:30 AM Bug #19888 (Feedback): Can't change binding of a Proc coerced from a Method
- `instance_exec` or `instance_eval` can execute a given `Proc` in the context of the receiver. However, that's not true when the `Proc` has been coerced from a `Method`. In that situation, the binding is always tied to the original instan...
-
08:21 AM Bug #17263: Fiber context switch degrades with number of fibers, limit on number of fibers
- OK, so I spent way longer than I should have staring at this but I think I've worked out what's going on. There are a couple of separate fibers to unpick here (ba-dum tish); I'll look at making fibers and switching between fibers as sepa...
-
- Applied in changeset commit:git|c87f2a4f156477ba962de19866a1f6298d567398.
----------
[Bug #19887] RUBYOPT should work without leading `-` -
- https://github.com/ruby/ruby/pull/8465
-
- It was working till 1.8.
```sh-session
$ for ruby in ruby{1.{8..9},2.{0..7},3.{0..3}}; do RUBYOPT=w $ruby -e 'puts "'$ruby': #{$VERBOSE}"'; done
ruby1.8: true
ruby1.9: false
ruby2.0: false
ruby2.1: false
ruby2.2: false
ruby2.3: f... -
-
01:04 AM Feature #19744: Namespace on read
- Eregon (Benoit Daloze) wrote in #note-21:
> On the native level it's a lot more tricky to isolate, for instance it's not really possible to fully fake different working directories in the same process (considering native extensions migh... -