Activity
From 01/25/2024 to 01/31/2024
01/31/2024
-
11:44 PM Revision 6a689e33 (git): [ruby/irb] Omit 2 encoding error related tests for TruffleRuby
- (https://github.com/ruby/irb/pull/854)
They're failing due to an issue in Prism: https://github.com/ruby/prism/issues/2129
So we need to skip them until:
- The issue is fixed in Prism
- TruffleRuby is updated to a version of Prism that... -
11:26 PM Bug #20226: Inconsistent Sort results on 3.3.0 compared to previous versions
- Seems like maybe it was changed by #19643 for `sort_by`. But in any case, it was never stable since it was using the platform's unstable `qsort_r`. On macOS, I get `[0.0, 0.0, 0.02, 0.02, -0.02, ...]` even on 3.2.2. Glibc has [warned abo...
-
10:57 PM Bug #20208: Net::HTTP errors with Errno::EAFNOSUPPORT when setting local_host with Addrinfo
- I discussed this a bit with @ioquatix, we've decided to revert the change to make AI_ADDRCONFIG be used in DNS lookups by default. The interaction with services on localhost is just too surprising.
-
09:29 PM Revision 9fdfdf4f (git): [ruby/prism] Always return the character width for char_is_identifier_start() and char_is_identifier_utf8()
- * This is also faster than calling pm_encoding_utf_8_alpha_char/pm_encoding_utf_8_alnum_char
as those compute the character width and do extra checks.
https://github.com/ruby/prism/commit/4cb276ac4c -
08:49 PM Revision b5a2c60d (git): [PRISM] Support SCRIPT_LINES__
-
08:44 PM Revision c4697991 (git): Do not modify provided argument splat when using ruby2_keywords with anonymous splat
- Previously, this would push the provided keywords onto the argument
splat. Add ruby2_keywords to the list of other checks for whether
it is safe for treating a given splat as mutable when the called
method accepts an anonymous splat. -
-
-
07:11 PM Revision cb98b018 (git): [PRISM] Fix else with rescue
- Fixes ruby/prism#2307.
-
07:07 PM Revision 95c9711d (git): [rubygems/rubygems] Fix musl platform not being added to the lockfile
- https://github.com/rubygems/rubygems/commit/235f7b4266
-
06:41 PM Revision c28ee912 (git): [PRISM] Fix incorrect use of VM_CALL_KW_SPLAT_MUT in zsuper with keyword splat
- This copies the changes from 771a2f039b9a059a73e8f111d1d46590fa697f63.
Fixes ruby/prism#2310. -
- Before this commit, we were mixing a lot of concerns with the prism
compile between RubyVM::InstructionSequence and the general entry
points to the prism parser/compiler.
This commit makes all of the various prism-related APIs mirror
th... -
06:13 PM Revision 21031f0a (git): YJIT: Float arithmetics are actually leaf
- with these guards in YJIT.
The previous commit was to fix "conflict" between two PRs, but I
actually wanted to use it here, which is why I filed the other one. -
-
- * YJIT: Add codegen for Float arithmetics
* Add Flonum and Fixnum tests -
- * YJIT: Add jit_prepare_for_gc function
* s/jit_prepare_routine_call/jit_prepare_non_leaf_call/
* s/jit_prepare_for_gc/jit_prepare_call_with_gc/
* Use jit_prepare_call_with_gc for leaf builtin -
- We're already testing with truffleruby 23.
https://github.com/rubygems/rubygems/commit/cd0494d628 -
- > Using ruby APIs in onigmo doesn’t feel nice.
I changed it to call `HANDLE_REG_TIMEOUT_IN_MATCH_AT` in onigmo, which calls `rb_reg_raise_timeout` so then there is no Ruby code in onigmo.
> ...
I think it can only raise when there ... -
04:12 AM Bug #20228: Memory leak in Regexp timeout
- Good find, thanks!
I'm a little concerned about the overhead of `rb_protect` for a typical simple match, but it's neglectable? -
12:57 AM Bug #20228: Memory leak in Regexp timeout
- Using ruby APIs in onigmo doesn’t feel nice.
-
-
- This seems worse to detect performance regressions, but at least should
not have many false positives.
https://github.com/rubygems/rubygems/commit/0b28e55415 -
01:21 PM Revision c70052e5 (git): [rubygems/rubygems] WEBrick:Utils::TimeoutHandler is always provided after webrick gem
- https://github.com/rubygems/rubygems/commit/b0502a0c50
-
12:43 PM Revision 5808c869 (git): [ruby/prism] Fix description of line parsing option and state it's 1-indexed
- https://github.com/ruby/prism/commit/b2d2b91222
-
- Some specs assert empty output, but sometimes they print warnings about
redefinition warnings. Ignore those until they are fixed upstream.
https://github.com/rubygems/rubygems/commit/0cd3b6dbae -
11:43 AM Bug #19999: Backport: .travis.yml and fixed commits
- jaruga (Jun Aruga) wrote in #note-8:
> @nagachika I sent the PR to the ruby_3_2 branch to pass Travis CI with the latest config file that exists on master and ruby_3_3 branches. You can do cherry-pick the commits or merge the PR.
> ...... - 10:43 AM Revision aefc9889 (git): readline: Skip the test_interrupt_in_other_thread on arm32.
- Skip the test_interrupt_in_other_thread test failing on Travis arm32.
The failing test is not a new issue. It had already been skipped in .travis.yml
before updating .travis.yml.
https://github.com/ruby/ruby/blob/7b05cb8dbbd637910757f402... - 10:43 AM Revision fc25e166 (git): Backport .travis.yml from ruby_3_3 branch.
- Update the .travis.yml file copied from the latest ruby_3_3 branch
<5124f9ac7513eb590c37717337c430cb93caa151>. -
- jaruga (Jun Aruga) wrote in #note-16:
> > Now I am testing the Travis s390x on my forked repository.
> ...
Merged. The s390x is added on Travis again.
- 08:56 AM Revision 816ce53a (git): Revert ".travis.yml: Drop s390x temporarily."
- This reverts commit 7ded31d36dc78c1495b03a45ec1a3235fdd81f1e.
I was told from Travis CI support that their infra team has deployed a fix for
the issue we encountered with the s390x build environment. - 07:15 AM Revision 2deb105a (git): Bump ruby/setup-ruby from 1.169.0 to 1.170.0
- Bumps [ruby/setup-ruby](https://github.com/ruby/setup-ruby) from 1.169.0 to 1.170.0.
- [Release notes](https://github.com/ruby/setup-ruby/releases)
- [Commits](https://github.com/ruby/setup-ruby/compare/5daca165445f0ae10478593083f72ca262... -
07:14 AM Revision a7ebe467 (git): Add newline between end and def
-
-
06:23 AM Bug #20230 (Closed): Fix crash when passing large keyword splat to method accepting keywords and keyword splat
- The following code causes a crash in Ruby 2.2-3.3 and master:
```ruby
h = {}
1000000.times{|i| h[i.to_s.to_sym] = i}
def f(kw: 1, **kws) end
f(**h)
```
Inside a thread or fiber, the size of the keyword splat could be much smal... -
- https://github.com/ruby/yaml/commit/da421ce46f
-
- https://github.com/flori/json/commit/202ffe2335
-
04:31 AM Revision f3df218f (git): Introduced `rb_node_const_decl_val` function
- Introduce `rb_node_const_decl_val` function to allow `rb_ary_join` and
`rb_ary_reverse` functions to be removed from Universal Parser. -
03:16 AM Feature #20108: Introduction of Happy Eyeballs Version 2 (RFC8305) in Socket.tcp
- Regarding the content mentioned above, I have made the following changes to the commit:
https://github.com/ruby/ruby/pull/9374/commits/461b75830599408feca086f7f6719b8426008802
- Improved performance in the case where there is only one a...
01/30/2024
-
- > Now I am testing the Travis s390x on my forked repository.
I tested. I sent a PR to add the s390x again.
https://github.com/ruby/ruby/pull/9773
-
- jaruga (Jun Aruga) wrote in #note-14:
> I will drop the s390x case in Travis CI temporarily. I am not sure that the issue comes from an infra or Ruby. But right now the test failing with 50 minutes is not convenient as a CI.
> ...
I go... -
- I will drop the s390x case in Travis CI temporarily. I am not sure that the issue comes from an infra or Ruby. But right now the test failing with 50 minutes is not convenient as a CI.
https://github.com/ruby/ruby/pull/9758
-
- It seems that Travis s390x is slow, running out the max 50 minutes (ruby_3_3 specific issue?),
https://app.travis-ci.com/github/ruby/ruby/builds/268615249
Or not starting soon.
https://app.travis-ci.com/github/ruby/ruby/builds/26861... -
- FWIW I'm keeping https://gist.github.com/eregon/912e6359e83781c5fa1c638d3768c526 to show that more methods are being implemented in Ruby (it's great).
And indeed that means gems sometimes need to handle backtrace entries starting with `... -
07:23 PM Bug #20227: Unexpected "<internal:array>" appears since c84237f9531aed3b204d3fdacc2dd9d2bd4c7d81
- For the record I fixed it in Rails with https://github.com/rails/rails/pull/50923.
-
- As Jean said, this is intentional. It's not new in Ruby 3.4 that we're changing backtraces by Ruby rewrite (Ruby 3.0: `Kernel#tap` and `Kernel#then`, Ruby 3.3: `Kernel#loop` and `Integer#times`), but I guess `Array#each` is more commonly...
-
- @yahonda this isn't a regression, it happens when methods are migrated from pure C to the new Primitive system. It's Rails that should be fixed.
-
11:47 AM Bug #20227 (Closed): Unexpected "<internal:array>" appears since c84237f9531aed3b204d3fdacc2dd9d2bd4c7d81
- Unexpected "<internal:array>" appears since c84237f9531aed3b204d3fdacc2dd9d2bd4c7d81
Rails CI against Ruby master branch gets failed
https://buildkite.com/rails/rails-nightly/builds/85#018d3dec-4a38-41cb-9ebf-1cbc78a697ba/1147-1156
... -
10:22 PM Revision 8041b7d9 (git): [PRISM] pm_compile_logical: Fix OrNode in IfNode predicate
- Fixes: https://github.com/ruby/prism/issues/2294
-
- As this is compiled into only on the recursive call and then added, we
can just use `ret` directly -
- File.expand_path leaks the dir if the encodings are not compatible.
For example:
Encoding.default_external = Encoding::UTF_16BE
10.times do
100_000.times do
File.expand_path("./a")
rescue
end
p... -
- I found that this bug was actually reported earlier in #20180, which I already have submitted a pull request to fix. I think we should use the approach in this pull request to fix this issue in master, as it is more efficient. We could...
-
- For the following code:
```ruby
a = []
kw = {}
[*a, **kw]
```
Ruby 2.0-3.3 return `[{}]`, when the desired result (since 2.7) is `[]`.
I discovered this while working on an optimization for the ARGSCAT case, where `[*a, 1, *... -
- * YJIT: Specialize splatkw on T_HASH
* Fix a typo
Co-authored-by: Alan Wu <XrXr@users.noreply.github.com>
* Fix a few more comments
---------
Co-authored-by: Alan Wu <XrXr@users.noreply.github.com> -
07:57 PM Revision fe5590e4 (git): YJIT: add specialized codegen for fixnum XOR (#9763)
-
- GitHub PR: https://github.com/ruby/ruby/pull/9765
If a Regexp::TimeoutError is raised, the `stk_base` and `OnigRegion` will leak.
For example:
```ruby
Regexp.timeout = 0.001
regex = /^(a*)*$/
str = "a" * 1000000 + "x"
10.t... -
- https://github.com/ruby/prism/commit/224ea85565
-
06:19 PM Revision 8e708e4a (git): Update forwarding locals for prism
-
-
- in response to [Bug #20227]
-
-
- Previously, this would use newarray followed by concattoarray.
This now uses pushtoarray instead, avoiding the unnecessary
array allocation.
This is implemented by making compile_array take a first_chunk
argument, passing in 1 in the no... -
- https://github.com/ruby/prism/commit/a35b8e45ee
-
- 03:22 PM Revision 846f87ee (git): .travis.yml: Drop s390x temporarily.
- The s390x pipelines are timeout (50 minutes) on both master and ruby_3_3
branches. Drop it temporarily.
* https://app.travis-ci.com/github/ruby/ruby/builds/268617296
* https://app.travis-ci.com/github/ruby/ruby/builds/268615249 -
- > many developers will have come to rely on and expect that behavior,
I think that's a fair point, similar to e.g Set insertion ordering, even though that's not documented I'm pretty sure some Ruby code relies on it because it holds s... -
02:50 PM Bug #20226: Inconsistent Sort results on 3.3.0 compared to previous versions
- If sort *was* stable for a long time (it does seem like it was, but I am not 100% sure), then despite Ruby's disclaimer, *many developers* will have come to rely on and expect that behavior, and consider this as "a behavior change that b...
-
01:48 PM Bug #20226: Inconsistent Sort results on 3.3.0 compared to previous versions
- Regardless of any individual impressions, Ruby never made any stability guarantees for sorting. It was always explicit that sorting MAY BE unstable. This is because Ruby delegates the actual sorting to the efficient sorting routines prov...
-
- Whether or not Ruby's documentation made any explicit guarantees, I was under the strong impression (as per the Ruby behavior I've observed for 10 years) that Ruby did have a stable sort. Other `Array` operations such as `-` (minus), `un...
-
- `Enumerable#sort_by` was also documented to not be stable in 3.2 (and previous versions): https://ruby-doc.org/3.2.3/Enumerable.html#method-i-sort_by
What might have changed (this is a layman's guess since I don't know the internals o... -
- > Sorting in Ruby never was stable
My mistake, like Roger Pack in https://github.com/crystal-lang/crystal/issues/2350#issuecomment-550188746, my experience is sort looks stable (on Linux).
And sort is stable on TruffleRuby since it u... -
- Here is the previous discussion from 15 years ago: https://bugs.ruby-lang.org/issues/1089
It might be about time to revisit the performance impact of a stable sort, imo, but that's not the discussion on this ticket. -
- Sorting in Ruby never was stable. There were discussions to make it stable, but Matz was concerned about the speed impact. I don't think this particular case is a regression, the behaviour of sort depends on the number of elements, and o...
-
- When was `sort`/`sort_by` changed to no longer be stable? I don't remind any ticket about it or NEWS entry.
IMO it's a pretty big change to make it non-stable when it has been stable for so long.
Especially for a high-level language li... -
10:15 AM Bug #20226: Inconsistent Sort results on 3.3.0 compared to previous versions
- You can force stability by using the original index as a tiebreaker.
```ruby
arr = [*Array.new(15,1), -0.02, 0.02]
puts "Unstable: #{arr.sort_by(&:abs)[0..1]}"
puts "Stable: #{arr.sort_by.with_index { [_1.abs, _2] }[0..1]}"
``` -
- `sort` family methods do not guarantee stable sort.
That means the result order of the same value elements is not affected by the original order of those elements. -
03:26 AM Bug #20226 (Closed): Inconsistent Sort results on 3.3.0 compared to previous versions
- Try this code block:
```
[-0.9, -0.88, -0.85, -0.83, -0.81, -0.79, -0.77, -0.75, -0.73, -0.71, -0.69, -0.67, -0.65, -0.63, -0.6, -0.58, -0.56, -0.54, -0.52,
-0.5, -0.48, -0.46, -0.44, -0.42, -0.4, -0.38, -0.35, -0.33, -0.31, -0.29, -... -
-
02:02 PM Misc #19758: Statically link ext/json
- Hello,
I am facing the same issue today, though I am presently on Ubuntu, not windows. I realized that `Init_ext` was doing nothing (I couldn't then use `require 'bigdecimal'` without a crash) by using a debugger and I noticed that it ... -
01:49 PM Revision 928f3884 (git): [DOC] Fix Ripper DSL input example
- '!' suffix is needed for event dispatch.
-
-
12:55 PM Revision fd44b42f (git): [ruby/irb] Fix undef and alias indent
- (https://github.com/ruby/irb/pull/838)
https://github.com/ruby/irb/commit/a641746b18 -
- Previously the code assigns `handler` function pointer, which accepts
one argument, to `sigact.sa_sigaction`, which accepts three argument.
This mismatch is not allowed in Wasm.
I don't see the reason to use `sa_sigaction` here, so this... -
11:29 AM Revision f585171a (git): merge revision(s) e12d4c654e3cb7a4473014610bc3bae41aaf811e: [Backport #20104]
- Don't create T_MATCH object if /regexp/.match(string) doesn't match
Fixes [Bug #20104]
---
re.c | 9 ++++++---
test/ruby/test_regexp.rb | 12 ++++++++++++
tool/lib/envutil.rb... -
- ruby_3_3 f8f0d342e48a38caac6d32b438c145bb581a51e6 merged revision(s) 3d19409637de1462b6790d2a92344bf0a10d8c52.
-
- https://github.com/ruby/io-console/commit/32583460e1
-
- https://github.com/ruby/io-console/commit/8b9b5b611a
-
- Use index for referring to symbols in `args` rule instead of named
references
In `args: args ',' arg_splat`, `args` is not unique name.
Currently the associated rule is interpreted as
`$$ = rest_arg_appe... -
- https://github.com/rubygems/rubygems/commit/6c4caf3ab0
-
- https://github.com/rubygems/rubygems/commit/d7c15f6fd7
-
- This reverts commit d4b780e84e9a6b858d0f6c6a44b22da0d2f5835e.
-
- [Bug #20094] Distinguish `begin` and parentheses
---
compile.c | 1 +
parse.y | 36 +++++++++++++++++++++---------------
test/ruby/test_whileuntil.rb | 18 +++++++... -
-
-
-
-
-
-
-
-
- https://github.com/rubygems/rubygems/commit/deb1b6d293
-
-
-
- Thanks to Kokubun for noticing.
Follow-up: b0711b1cf152afad0a480ee2f9bedd142a0d24ac -
- ruby_3_3 5f3dfa1c273c6fb9eae65ceca633b46f7e30f686 merged revision(s) d8702ddbfbe8cc7fc601a9a4d19842ef9c2b76c1.
-
-
02:53 AM Bug #20225 (Assigned): Inconsistent behavior of regex matching for a regex has a null loop
- Usually, in Ruby (Onigmo), when a null loop (a loop consuming no characters) occurs on regex matching, this loop is terminated. But, if a loop has a capture and some complex condition is satisfied, this causes backtracking. This behavior...
-
- ruby_3_3 818b4ea9b16e3570b431b86da9a24a5743b29617 merged revision(s) e5a4f757bdf5dc3d8c329ddd268432f9ecc7bff6.
-
- It's nice if we can backport https://github.com/ruby/ruby/pull/9634 (3d19409637de1462b6790d2a92344bf0a10d8c52) to Ruby 3.3 branch.
Even so previous codes work as expected but it depends on how parse.c works and misleading.
-
- Fix [Bug #20083]: correct a cache point size for atomic groups
(#9367)
---
regexec.c | 2 +-
test/ruby/test_regexp.rb | 8 ++++++++
2 files changed, 9 insertions(+), 1 deletion(-) -
- Fix Window private file mapping unlink EACCES issue. (#9358)
* Don't return early.
* Add missing `mapping` assignment.
* Make debug logs conditional.
---
io_buffer.c | 18 ++++++...
01/29/2024
-
11:59 PM Feature #20196: Proposal: Binary data literal
- For what it's worth, there's a shorthand String#b.
``` ruby
"\xAB\xCD\xEF\x12\x34\x56".b
#=> => "\xAB\xCD\xEF\x124V"
```
For fun, here's a Ruby implementation along the lines of what you propose.
```ruby
module Kernel
def... -
11:14 PM Bug #20095 (Closed): Regex lookahead behaving strangely in 3.3.0
- If the duplicate has already be resolved, then let's close this, too.
-
- It is the exact same as #20207.
-
- Fixes ruby/prism#2295.
-
- Sometimes this file get picked up and break Ripper tests:
TestRipper::Generic#test_parse_files:test/ruby
assert_separately failed with error message
pid 63392 exit 0
| test_regexp.rb:2025: warning: character class has du... -
-
-
-
09:25 PM Bug #20188: `Module#const_source_location` returns wrong information when real constant was defined but autoload is still ongoing
- I have tried several things using `require` and `autoload` with the [test scheduler](https://github.com/ruby/ruby/blob/master/test/fiber/scheduler.rb) and cannot get a deadlock by now.
-
- I'll open a separate issue for this :). If `require` is fiber-safe I'd like to know and maybe improve its docs. If it is not, that may be important.
-
- So, we are arriving at the concern I expressed above in the first message of this "subthread".
Since fiber schedulers have no contract related to `require` or `autoload`, my understanding is that these methods are thread-safe and not ... -
- > That is less the case with fiber schedulers, no?
It's still cooperative. You implicitly yield on IOs sure, but still won't be prempted after running for too long.
> ...
I don't see why not, but I may be missing something. I'm not... -
- To me, this is related with `autoload`, because it is the `autoload` the one making the fibers deadlock. If the last line was
```ruby
Fiber.new { p 1 }.resume
```
instead of a reference to the constant being autoloaded, there wou... -
- > If a fiber does a non-blocking operation while the file is being loaded, and that was part of an autoload or require, what can we assume about that situation?
I don't think it's anything particular to `autoload`.
`autoload` simpl... -
- * YJIT: print warning when disasm options used without a dev build
I was confused for a few minutes the other way then
--yjit-dump-disasm printed nothing, so I figured this would be
useful for end-users (and future me).
* Fix lone extr... -
- Fixes ruby/prism#2290.
-
07:44 PM Feature #20080: Introduce #bounds method on Range
- Oh you're right, it seems like I mixed up the #deconstruct comment in note-2 with the rejection notice in note-3.
Although imho it doesn't feel intuitive to pattern-match a range like that. I would expect to able to do `1..8 in [*,5,*]`... -
- If the first element of an interpolated string node is an embedded
statement, CRuby "pre-initializes" the interpolation with a string of
known encoding to concat into.
This patch replicates thate behaviour in Prism -
-
-
- Fixes ruby/prism#2242.
-
- https://github.com/ruby/prism/commit/f614863d79
-
- This reverts commit e0f4c4e410a0e4c6cda67e9000696c8f1f01d8aa.
We expect https://github.com/ruby/ruby/pull/9729 to address the failure. - 05:22 PM Revision d42330d7 (git): [rubygems/rubygems] Improve gem login scope selection
- https://github.com/rubygems/rubygems/commit/26c7abe5f6
-
- Suppose YJIT runs a rb_vm_opt_send_without_block()
fallback and the control frame stack looks like:
```
will_tailcall_bar [FINISH]
caller_that_used_fallback
```
will_tailcall_bar() runs in the interpreter and sets up a tailcall.
Right ... -
- Not a bug, you are doing `yield @obj.each(&block)`, and Array#each returns the receiver.
-
04:48 PM Bug #20223 (Rejected): For simple objects each_cons appears to work as I expect, but for other objects it seems to fail to terminate iterations when I would expect.
- Reproduce process:
``` ruby
p `ruby -v`
class X
include Enumerable
attr_accessor :obj
def initialize = @obj = (1..5).to_a
def each(&block) = yield @obj.each(&block)
end
wtf = X.new
(1..5).each_cons(3) { |g| p g }
p '... -
- Fixes ruby/prism#2272.
-
- ``` ruby
source = %q{"foo".freeze.equal?("foo".freeze)}
RubyVM::InstructionSequence.compile(source).eval # => true
RubyVM::InstructionSequence.compile_option = false
RubyVM::InstructionSequence.compile(source).eval # => false
```
... -
- https://github.com/ruby/prism/commit/102b4a16f5
-
- Small PR to add a comment when we clear local variable types,
so we can be aware that it's happening when looking at the disasm. -
- Co-Authored-By: Kevin Newton <kddnewton@gmail.com>
-
- Applied in changeset commit:git|3e6e3ca2627b1aa71b17de902cc1b8188246a828.
----------
Correctly handle consecutive lookarounds (#9738)
Fix [Bug #20207]
Fix [Bug #20212]
Handling consecutive lookarounds in init_cache_opcodes is buggy, s... -
- It is the exact same bug as #20207.
-
- Applied in changeset commit:git|3e6e3ca2627b1aa71b17de902cc1b8188246a828.
----------
Correctly handle consecutive lookarounds (#9738)
Fix [Bug #20207]
Fix [Bug #20212]
Handling consecutive lookarounds in init_cache_opcodes is buggy, s... -
- This wasn't taking into account strings with flags set on the containing
node that could affect the encoding. It does now -
- Fix [Bug #20207]
Fix [Bug #20212]
Handling consecutive lookarounds in init_cache_opcodes is buggy, so it
causes invalid memory access reported in [Bug #20207] and [Bug #20212].
This fixes it by using recursive functions to detected look... -
02:45 PM Revision 0d4de0f4 (git): wasm: align fiber stack pointer to 16 bytes
- In WebAssembly C ABI, the linear stack pointer must be always aligned
to 16 bytes like other archs.
The misaligned stack pointer causes some weird memory corruption since
compiler assumes the aligned stack pointer. -
- Thanks for this. I'm able to reproduce this on my machine (Fedora 39 with clang 17.0.6). A cursory inspection suggests this might be a new behaviour in Clang 17, because of `-fsanitize-address-use-odr-indicator` defaulting to on (https:/...
-
10:48 AM Bug #20221 (Closed): ASAN: make test-basic: un-prefixed symbol leakage
- When building and running the tests (here, test-basic) with ASAN enabled, it fails with the following message:
```
Checking leaked global symbols...leaked
__odr_asan_gen_rb_cArray
__odr_asan_gen_ruby_digitmap
__odr_asan_gen_... -
-
- I'm wondering which encoding should be used if the parser hits invalid source encoding like `# coding: foo`. I think it's needed to clarify which encoding is assumed on this ticket.
-
07:46 AM Bug #20147: FreeBSD libyaml
- Analyzing the changes made on January 6. From 9b78ef75522b1f6aa20fc81ddf06e5fb40db152d I am not able to compile. Even though the previous commits I can compile, the message of yaml.h not being found keep showing up.
-
- The 76afbda5b566148b9e73939e0ff9b8464d59806c. The last commit from January 6 is not compiling. It seems that something was changed on January 6 that broke the compilation in my environment.
-
- The day I reported this bug I was still able to compile Ruby using master. Today I have decided to try compiling again from master, before compiling I have fetched from upstream and the compilation was not working anymore. If I use the 7...
-
05:39 AM Revision 933ede5d (git): [rubygems/rubygems] Remove `travis_removal_info`
- `travis_removal_info` is added by https://github.com/rubygems/rubygems/pull/6150. According to the comment, it's supposed to be removed at bundler v2.5.0 but it hasn't.
https://github.com/rubygems/rubygems/commit/e18797d43f -
- https://github.com/rubygems/rubygems/commit/0b8faf1e39
-
- Sometimes you want this, sometimes you don't. And when you don't, this
hides other debugging puts you may have added.
https://github.com/rubygems/rubygems/commit/df37582c81 -
- https://github.com/rubygems/rubygems/commit/d05b9e659b
Co-authored-by: Samuel Giddins <segiddins@segiddins.me> -
- https://github.com/rubygems/rubygems/commit/e0d68a8688
-
-
- https://github.com/rubygems/rubygems/commit/3a262f55c8
-
- https://github.com/rubygems/rubygems/commit/7f35dc19c5
-
- https://github.com/rubygems/rubygems/commit/5d6a8f2fb4
-
-
- https://github.com/rubygems/rubygems/commit/62bc261042
01/28/2024
-
- While in threads Ruby controls context switching and coordination for autoloading constants or loading files with `Kernel#require` is built-in, the problem I see with fibers is that Ruby has no control, the user has the control by design...
-
04:12 PM Feature #20080: Introduce #bounds method on Range
- @Dan0042 To be fair, that ticket seems to have rejected "old-style" deconstruction mainly (`b, e = range`). The possibility of `#deconstruct` is mentioned in one of the comments, but rejection is more vague on it.
-
- AMomchilov (Alexander Momchilov) wrote in #note-15:
> Could we implement this as `#deconstruct`, so Ranges can support destructuring?
This was rejected in the original ticket from which this one originated: #20027#note-3
-
- https://github.com/ruby/win32ole/commit/57e4a38465
-
- https://github.com/ruby/win32ole/commit/78ff137c0f
-
- https://github.com/ruby/win32ole/commit/70ea60c4d2
-
- https://github.com/ruby/win32ole/commit/7648ee7e56
-
- https://github.com/ruby/win32ole/commit/2b91b6b838
-
- https://github.com/ruby/win32ole/commit/1c95816168
-
- https://github.com/ruby/win32ole/commit/2f51493bd1
-
- https://github.com/ruby/win32ole/commit/2c5d193da7
-
- https://github.com/ruby/win32ole/commit/5feede2cc5
-
- Prevent name clash with `WIN32OLE::Variant`, of generated document
files on case-insensitive filesystems, such as Windows.
https://github.com/ruby/win32ole/commit/049e5f0a6e - 12:38 PM Revision c1666158 (git): [ruby/reline] Add metadata for rubygems.org
- (https://github.com/ruby/reline/pull/638)
https://github.com/ruby/reline/commit/d3a324d22c -
- Applied in changeset commit:git|23b8337cd10329020d74a2f0d8e43434645c4d5b.
----------
[Bug #20219] `gettable` returns NULL on error -
-
- Applied in changeset commit:git|fed877c791f1c16a2b1a2c9a167b7f433505794d.
----------
[Bug #20217] `return` with `ensure` is a void value expression -
-
-
- Since RDoc searches `var = rb_define_class_under(...)` statements
literally, they cannot be built by macros.
https://github.com/ruby/digest/commit/d39b684f91 -
- https://github.com/ruby/digest/commit/e5d30394b3
-
- https://github.com/ruby/digest/commit/6db96aa99a
-
- When switching stacks via `coroutine_transfer` in the M:N threading information, we need to tell ASAN about it in the same way that we do in cont.c. This is a matter of calling `__sanitizer_start_switch_fiber` in the transferring M:N thr...
-
- ```
ruby: YJIT has panicked. More info to follow...
thread '<unnamed>' panicked at src/core.rs:2751:9:
assertion `left == right` failed: each stub expects a particular iseq
left: 0x7fc8d8e09850
right: 0x7fc8d2c2f3a0
stac... -
-
-
-
- https://github.com/ruby/prism/commit/d3722d6660
01/27/2024
-
- newarray, duparray, concatarray, and splatarray always leave an
array at the top of the stack. expandarray does not, it takes
an array from the top of the stack as input, and leaves individual
elements on the stack. I assume no Ruby co... -
- Given code such as:
```ruby
h[*a, :a], h[*b] = v
```
Ruby would previously allocate 5 arrays for the mass assignment:
* splatarray true for a
* newarray for v[0]
* concatarray for [*a, a] and v[0]
* newarray for v[1]
* concatarray f... -
- Given code such as:
```ruby
h[*a, 1] += 1
h[*b] += 2
```
Ruby would previously allocate 5 arrays:
* splatarray true for a
* newarray for 1
* concatarray for [*a, 1] and [1]
* newarray for 2
* concatarray for b and [2]
This optimi... -
- https://github.com/ruby/prism/commit/8cdec8070c
-
- Previously, a literal array with a splat and any other args resulted in
more than one array allocation:
```ruby
[1, *a]
[*a, 1]
[*a, *a]
[*a, 1, 2]
[*a, a]
[*a, 1, *a]
[*a, 1, a]
[*a, a, a]
[*a, a, *a]
[*a, 1, *a, 1]
[*a, 1, *a, *a]
... -
- Split up the diagnostic levels so that error and warning levels
aren't mixed. Also fix up deconstruct_keys implementation.
https://github.com/ruby/prism/commit/bd3eeb308d
Co-authored-by: Benoit Daloze <eregontp@gmail.com> -
- * So it's clear it is a Symbol.
Before:
... @level=warning_verbose_true>
After:
... @level=:warning_verbose_true>
https://github.com/ruby/prism/commit/84503643b9 -
- To avoid stack overflow, Ruby splits compilation of large arrays
into smaller arrays, and concatenates the small arrays together.
It previously used newarray/concatarray for this, which is
inefficient. This switches the compilation to u... -
05:34 PM Feature #20080: Introduce #bounds method on Range
- Could we implement this as `#deconstruct`, so Ranges can support destructuring?
```ruby
class Range
def deconstruct = [self.begin, self.end]
end
case 1..2
in [1, Integer => upper]
p "matched: #{upper}"
else
p "not matc... -
- since 2.6 and 2.7 are EOL and bundler dropped their support by https://github.com/rubygems/rubygems/pull/7116.
https://github.com/rubygems/rubygems/commit/b562d9a822 -
- ```ruby
def foo(bar: (1 in ^bar)); end
```
Segfaults on master. (I don't think anyone would ever write this code.) -
- https://github.com/ruby/prism/commit/9b6907b727
-
- They are not "reference"s, I think.
-
- `__ENCODING__ `was managed by `NODE_LIT` with Encoding object.
Introduce `NODE_ENCODING` for
1. `__ENCODING__` is detectable from AST Node.
2. Reduce dependency Ruby object for parse.y -
- Before this commit `ruby -y -e 'tap {|;x, y|}'` failed with SEGV.
This change fixes it. -
-
02:34 AM Feature #20196: Proposal: Binary data literal
- shyouhei (Shyouhei Urabe) wrote in #note-4:
> Reminds me of discussions in ISO C/C++ committee (in both, simultaneously, by the same person). C23 opted not to have binary string literal, instead introduced a way to load binaries into C ... -
- I found that use of keyword arguments in multiple assignment is broken in 3.3 and master:
```ruby
h = {a: 1}
o = []
def o.[]=(*args, **kw)
replace([args, kw])
end
# This segfaults as RHS argument is not a hash
o[1, a: 1], _...
01/26/2024
-
-
- * Fixes https://github.com/ruby/prism/issues/2082
https://github.com/ruby/prism/commit/7a74576357 -
- Ensure we don't accidentally parse the symbol encoding twice, and
ensure we parse it in every circumstance we need to by requiring
it as a parameter.
https://github.com/ruby/prism/commit/9cea31c785 -
- Fixes ruby/prism#2250.
Co-Authored-By: Kevin Newton <kddnewton@gmail.com> -
- On s390x, a long is 8 bytes. st_data_t is an unsigned long but
pm_constant_id_t is a 4 byte integer. We need to cast it to st_data_t
when passing it to ST functions. - 08:15 PM Revision 2a509787 (git): [ruby/prism] Track whether a Symbol should have its encoding changed from the source encoding.
- Ruby sets a Symbol literal's encoding to US-ASCII if the symbols consists only of US ASCII code points. Character escapes can also lead a Symbol to have a different encoding than its source's encoding.
https://github.com/ruby/prism/comm... -
- Fixes ruby/prism#2279.
-
- ```ruby
foo = begin
return
ensure
end
```
I would expect this to raise a syntax error for a void value expression being assigned, but the `ensure` makes it so that this isn't checked. -
- https://github.com/ruby/prism/commit/91b5550726
- 06:35 PM Revision 7d356b8f (git): [ruby/prism] Fix multiple assigns with newlines
- https://github.com/ruby/prism/commit/b4ba41bdcd
- 06:32 PM Revision 9a7637da (git): [ruby/prism] Add tests
- https://github.com/ruby/prism/commit/b9ebf987bd
- 06:32 PM Revision b0e10345 (git): [ruby/prism] Fix parsing calls with labels
- https://github.com/ruby/prism/commit/3db7849a31
- 06:32 PM Revision 8e5bc8f7 (git): [ruby/prism] Change binding power for modifiers in case-in nodes
- https://github.com/ruby/prism/commit/c31eeba54c
-
- I wanted to check if this was a bug or desired.
Usually when you reference a parameter within its own default value (optional positional or optional keyword) it raises a syntax error. This is unless you use it in an assignment, in whi... -
-
-
05:19 AM Feature #20215 (Open): Introduce `IO#readable?`
- There are some cases where, as an optimisation, it's useful to know whether more data is potentially available.
We already have `IO#eof?` but the problem with using `IO#eof?` is that it can block indefinitely for sockets.
Therefore... -
- Fixed by commit:771a2f039b9a059a73e8f111d1d46590fa697f63
-
- For zsuper calls with a keyword splat but no actual keywords, the
keyword splat is passed directly, so it cannot be mutable, because
if the callee accepts a keyword splat, changes to the keyword splat
by the callee would be reflected in ... -
- I'm not sure why OpenBSD suggest `rubygems_plugin` file for this.
https://rubyci.s3.amazonaws.com/openbsd-current/ruby-master/log/20240126T013005Z.fail.html.gz
It seems environmental issue, not irb. -
- "builtin-baseline": "53bef8994c541b6561884a8395ea35715ece75db" is 2024.01.12
released version of vcpkg.
https://github.com/microsoft/vcpkg/releases/tag/2024.01.12 -
- Ruby 3.3.0 YJIT missed the support for the instruction that was added shortly before the 3.3.0 release. It's used in Rails, and we didn't mean to exit on such method calls.
It'd be nice if we can fix the issue in Ruby 3.3.1 by backpor... -
01/25/2024
- 11:48 PM Revision 30b4070f (git): Update bundled gems list at b822f9f6fc5a00ec08dbc698b98458 [ci skip]
-
11:47 PM Revision b822f9f6 (git): Bundle rbs-3.4.3 (#9702)
- * Bundle rbs-3.4.3
* Remove rbs from `TEST_BUNDLED_GEMS_ALLOW_FAILURES` list
* Add the failing tests to `rbs_skip_tests` -
-
- Thanks for your thoughts.
I agree ideally Ruby wouldn’t be carrying around hacks to work around glibc bugs that canonical couldn’t be bothered backporting fixes for. That bug was certainly my original motivation for looking at this is... -
04:32 PM Bug #20208: Net::HTTP errors with Errno::EAFNOSUPPORT when setting local_host with Addrinfo
- I'll address each suggestion from my POV, though this is not my usual area of work. It sure is tricky.
1) I feel like this might bring more subtle bugs. I think you're right about also expecting that "localhost" should also work. Or e... -
- I opened https://github.com/ruby/ruby/pull/9698 for the Net::HTTP tests.
The rubyspec tests are a little trickier. Essentially they do this:
```
['::1', '127.0.0.1'].each do |addr|
server = TCPServer.new(addr, 0)
conn = TCPS... -
- Thanks for this report - it was super detailed and made it very easy for me to figure out what's going on!
Firstly, your bisection is right. The AI_ADDRCONFIG flag is what makes the difference here. The flag causes glibc to NOT return... -
10:08 PM Bug #20085: Fiber.new{ }.resume causes Segmentation fault for Ruby 3.3.0 on aarch64-linux
Any guess as to when the backport to 3.3.0 will land? Our 3.3.0 upgrade is on hold, wondering if we should wait or employ a workaround.
-
- As the subject states, the super call in this code is compiled incorrectly:
```ruby
extend(Module.new{def a(**k) k[:a] = 1 end})
extend(Module.new{def a(**k) p k; super; p k end})
a
# Expected output, actual output on Ruby 2.0-3... -
-
-
-
05:46 PM Bug #20212 (Closed): Regex match method is crashing - (irb):6: [BUG] Segmentation fault at 0x0000000000000000 ruby 3.3.0 (2023-12-25 revision 5124f9ac75) [x86_64-darwin22]
- Tried this with ruby 3.3.0, both irb and rails console are crashing
```ruby
DOMAIN_REGEX = Regexp.new(
/\A((?=.*?[a-z])(?!.*--)[a-z\d]+[a-z\d-]*[a-z\d]+).((?=.*?[a-z])(?!.*--)[a-z\d]+[a-z\d-]*[a-z\d]+).((?=.*?[a-z])(?!.*--)[a-zd]+... -
-
- Alright, I finally got hold of the benchmarking server. For the record it's an AWS c5n.metal, with various tuning like disabling frequency scaling etc to hopefully get more stable results.
- `yjit-bench` revision: https://github.com/S... -
- > I have no idea which benchmarks are micro and which are headline. What about erubi and erubi-rails?
Each yjit-bench suite has a category: https://github.com/Shopify/yjit-bench/blob/3774b4bc320519f8b560eb23bdea48f549cf8b30/benchmark... -
- byroot (Jean Boussier) wrote in #note-31:
> I shouldn't have ran the micro-benchmarks anyway, only the headline benchmarks make sense.
I'll freely admit I have no idea which benchmarks are micro and which are headline. What about eru... -
- > The worst case is -6% on setivar_object
This is a fluke caused by random slowdown on my development machine. Here's the benchmark source: https://github.com/Shopify/yjit-bench/blob/3774b4bc320519f8b560eb23bdea48f549cf8b30/benchmarks... -
- byroot (Jean Boussier) wrote in #note-25:
> You are missing the social aspect of it.
Thank you for the explanation. I'm not sure I fully agree but it makes a lot more sense now.
---
Eregon (Benoit Daloze) wrote in #note-28:
> ...
Let'... -
- > I would say 5% on lobsters is a huge gain.
Still trying to get hold on our benchmarking server...
But just to note, `lobsters` itself doesn't use `frozen_string_literal: true`, and it's probably the case of at least some of its d... -
- mame (Yusuke Endoh) wrote in #note-9:
> @zverok Ah, that's exactly why I was against the introduction of `frozen_string_literal`, because people misunderstand it like you. This feature must not be mixed up with "immutability". I can und... -
-
-
- https://github.com/ruby/prism/commit/149e2ff7f6
-
- Ah, thank you - yes, this should be backported to 3.3 I think.
-
- @kjtsanaktsidis I'm assuming this should be backported to 3.3, but if I'm wrong you can change it to `DONTNEED`
-
- Applied in changeset commit:git|ef276858d9295208add48e27208c69184dc50472.
----------
Trigger postponed jobs on running_ec if that is available
Currently, any postponed job triggered from a non-ruby thread gets sent
to the main thread, ... - 10:20 AM Revision fdb8f086 (git): [ruby/irb] Reword history file documentation and fix typo
- (https://github.com/ruby/irb/pull/842)
https://github.com/ruby/irb/commit/bbabf818c7 -
09:58 AM Revision 4bd58e1a (git): [ruby/irb] diabled ==> disabled
- https://github.com/ruby/irb/commit/295797ff37
-
- https://github.com/ruby/irb/commit/2ffacaa031
-
- https://github.com/ruby/irb/commit/24c7694467
-
- https://github.com/ruby/irb/commit/6209f06c72
-
- https://github.com/ruby/irb/commit/a27a511777
-
- https://github.com/ruby/irb/commit/dbd0e368c4
-
- https://github.com/ruby/irb/commit/78dea58000
-
- https://github.com/ruby/irb/commit/b77b23aade
-
- Co-authored-by: Nobuyoshi Nakada <nobu@ruby-lang.org>
-
- In my windows environment uses cp932 for terminal encoding.
-
- Encoding::Windows_31J is already loaded in mswin platform
-
- I don't remember the reason to select `ArgumentError`, `SyntaxError` feels more reasonable.
https://github.com/ruby/ruby/pull/9701 -
- I was hoping we could change the error that is raised when an invalid source encoding is found from an ArgumentError to a SyntaxError.
First let me say, if this isn't possible for backward compatibility, I understand. Please do not ta... -
- Instead of scanning before the current comment.
-
- (My sincere apologies if this issue has already been raised, I didn't see it in Ruby 3.4 backlog.)
As per https://bugs.ruby-lang.org/issues/19370, Ruby 3.3 no longer allows the following:
```ruby
def m(*)
[1, 2, 3].each { |*| p... -
-
- This should be reverted after rbs and typeprof are updated to handle
the changes for the optimizations. -
02:29 AM Feature #20066: Reduce Implicit Array/Hash Allocations For Method Calls Involving Splats
- Please try!
-
-
- Benchmark results:
```
named_multi_arg_splat
after: 5344097.6 i/s
before: 3088134.0 i/s - 1.73x slower
named_post_splat
after: 5401882.3 i/s
before: 2629321.8 i/s - 2.05x slower
anon_arg_splat
after: 12242780.9 i/s
befor... -
- This allows ... argument forwarding to benefit from Allocationless
Anonymous Splat Forwarding, allowing the `f` call below to not
allocate an array or a hash.
```ruby
a = [1]
kw = {b: 2}
def c(a, b:)
end
def f(...)
c(...)
end
f(*a,... -
- Ruby makes it easy to delegate all arguments from one method to another:
```ruby
def f(*args, **kw)
g(*args, **kw)
end
```
Unfortunately, this indirection decreases performance. One reason it
decreases performance is that this alloc... -
- This instruction is similar to concattoarray, but it takes the
number of arguments to push to the array, removes that number
of arguments from the stack, and adds them to the array now at
the top of the stack.
This allows `f(*a, 1)` to ... -
- This instruction is similar to concatarray, but assumes the first
object is already an array, and appends to it directly. This is
different than concatarray, which will create a new array instead
of appending to an existing array.
Addi... -
- This flag is set when the caller has already created a new array to
handle a splat, such as for `f(*a, b)` and `f(*a, *b)`. Previously,
if `f` was defined as `def f(*a)`, these calls would create an extra
array on the callee side, inste... -
- Currently, any postponed job triggered from a non-ruby thread gets sent
to the main thread, but if the main thread is sleeping it won't be
checking ints. Instead, we should try and interrupt running_ec if that's
possible, and only fall b... -
-
-
- The GC by default doesn't move objects so it shouldn't write to the array. I guess it's a use-after-free and double free caused by `callcc` in the test jumping into the middle of `qsort_r` like the glibc devs noticed https://sourceware.o...
-