Activity
From 09/10/2023 to 09/16/2023
09/16/2023
-
10:27 PM Revision cd67c0d2 (git): [ruby/irb] Test should not depend on user's irbrc file
- (https://github.com/ruby/irb/pull/714)
https://github.com/ruby/irb/commit/02703c46f9 -
06:29 PM Revision 380b42fe (git): Fix comment for rb_enc_str_new [ci skip]
-
05:24 PM Revision 4aac7b1a (git): Another try to fix build in emscripten
- malloc_trim is defined in emscripten/emmalloc.h on emscripten.
-
03:47 PM Bug #19683: ruby-3.3.0-preview1 does not build with BSD make without --with-baseruby
- This is still a problem in 3.3.0-preview2. I've been using the following bad hack to work around it:
```diff
Index: common.mk
--- common.mk.orig
+++ common.mk
@@ -378,10 +378,10 @@ ext/configure-ext.mk: $(PREP) all-incs $(MKFILES... -
03:02 PM Bug #19864 (Closed): Ruby 3.2 Changed Behavior With One Sided Ranges
- Applied in changeset commit:git|25711683e86271385e8abe09a9c03782000e48db.
----------
Fix regression when testing inclusion in unbounded ranges
Caused by 04a92a6764bf678919cf4b68a27496a39d6b886a. This treats
unbounded ranges of arbitra... -
- I've submitted a pull request for the diff posted earlier (with tests): https://github.com/ruby/ruby/pull/8458
-
- Caused by 04a92a6764bf678919cf4b68a27496a39d6b886a. This treats
unbounded ranges of arbitrary objects the same as how unbounded
string ranges are treated:
(..x) === y # (y <=> x) <= 0
(...x) === y # (y <=> x) < 0
(x..) === y... -
02:15 PM Feature #19744: Namespace on read
- It seems interesting.
Regarding the motivation I wanted to mention some alternatives.
> Running multiple applications on a Ruby process
> ...
These 3 motivations can actually be solved today on TruffleRuby and I think on JRuby as well b... - 01:36 PM Revision 8835ca23 (git): [ruby/yarp] Fix a possible null pointer deference
- https://github.com/ruby/yarp/commit/41f601b81b
-
01:13 PM Revision e779465f (git): Add comment markers in empty lines [ci skip]
-
-
- ```
gc.c:9746:5: error: implicit declaration of function 'malloc_trim' is invalid in C99 [-Werror,-Wimplicit-function-declaration]
malloc_trim(0);
^
```
http://rubyci.s3.amazonaws.com/crossruby/crossruby-master-wasm32_emscripten... - 12:48 PM Revision a8afedce (git): [ruby/irb] Handle Concurrent Sessions and Saving Readline::HISTORY
- (https://github.com/ruby/irb/pull/651)
* handle concurrent sessions and saving Readline::HISTORY, fixes https://github.com/ruby/irb/pull/510
* separate tests
* don't mutate the HISTORY object on the class
* avoid repeated .to_i calls... -
12:47 PM Bug #19880: Missing write barrier in iseq instruction list
- Merged into ruby_3_2 at https://github.com/ruby/ruby/commit/a7335e11e354d1ee2e15233f32f087230069ad5c.
-
- Empty ranges do not overlap with any range.
Regarding benchmarks, PR#8242 is significantly faster in some cases,
but one of these two cases is a wrong result.
| |ActiveSupport| PR#8242|built-ruby|
|:----------... -
08:07 AM Feature #19839: Need a method to check if two ranges overlap
- @nobu is trying to fix the corner cases that @ko1 pointed: https://github.com/ruby/ruby/pull/8448
-
05:57 AM Feature #19839 (Closed): Need a method to check if two ranges overlap
- Applied in changeset commit:git|e9b503f1bb9692eda1d1f55f62c19d861b88a0d5.
----------
[Feature #19839] Add Range#overlap?
Add a method that returns true if two range overlap, otherwise false.
```
(0..10).overlap?(5..15) #=> true
(0..10... -
07:52 AM Revision c3ef7a52 (git): Fix malloc_trim() on wasm32
- ```
compiling gc.c
gc.c:9746:5: error: implicit declaration of function 'malloc_trim' is invalid in C99 [-Werror,-Wimplicit-function-declaration]
malloc_trim(0);
^
1 error generated.
``` -
06:54 AM Misc #19883: DevMeeting-2023-10-12
- * [Feature #19744] Namespace on read (tagomoris)
* Now the PoC implementation works, can require 2 different versions of native extension libraries.
* I will introduce the motivation and the implementation idea.
* I need discuss... -
- ```
$ cat Gemfile
source "https://rubygems.org"
$ cat test.rb
require "base64"
$ bundle exec ruby test.rb
$
```
In this situation, `bundle exec ruby test.rb` should print a warning: base64 which will be not part of the de... -
- Add a method that returns true if two range overlap, otherwise false.
```
(0..10).overlap?(5..15) #=> true
(0..10).overlap?(20..30) #=> false
``` -
- Fix missing write barrier in iseq instruction list
[Bug #19880]
There's a missing write barrier for operands in the iseq instruction
list, which can cause crashes.
It can be reproduced when Ruby is compiled with `-DRUBY_DEBUG_ENV=1`.
... -
- Agreed. I think the warning is good only when bundler is used.
@hsbt How about https://github.com/ruby/ruby/pull/8457 ? -
04:06 AM Revision e159971e (git): [For ruby_3_2] Fix unused_mut Rust warnings (#8435)
- Fix unused_mut Rust warnings
Rust version 1.71.0 and up issue these warnings. On GitHub CI, the
warnings were previously seen in -DYJIT_FORCE_ENABLE runs. -
03:10 AM Revision 7d08dbd0 (git): Optimize Range#bsearch for beginless/endless ranges
- On Range#bsearch for endless ranges, we try positions at `begin + 2**i` (i = 0, 1, 2, ...)
to find a point that satisfies a given condition.
Subsequently, we perform binary searching with the interval `[begin, begin + 2**n]`.
However, t... -
- This fixes the issue commit 43ab2acf8242c580b1ef5d644ea94d48669d4869
"Resurrect gem-specific patterns for sync" described. -
01:41 AM Revision 9aeb6e72 (git): YJIT: Avoid creating a vector in get_temp_regs() (#8446)
- * YJIT: Avoid creating a vector in get_temp_regs()
Co-authored-by: Alan Wu <alansi.xingwu@shopify.com>
* Remove unused import
---------
Co-authored-by: Alan Wu <alansi.xingwu@shopify.com>
Co-authored-by: Alan Wu <XrXr@users.noreply.g... - 12:34 AM Revision 1fbfd066 (git): Update default gems list at 7f96c9be9e998adb0f9db751bf9cd6 [ci skip]
-
12:32 AM Revision 7f96c9be (git): [ruby/yarp] Bump to version 0.12.0
- https://github.com/ruby/yarp/commit/c71137377f
09/15/2023
-
- Yet another partial revert of https://github.com/ruby/ruby/pull/8329,
similar to 00f263e6c4.
Repro: On ruby/ruby@1be64e34d0, `tool/sync_default_gems.rb yarp
162c2088eec6ec8f0558559e082cd661c18ee02a` should exist successfully, but
it doe... -
- https://github.com/ruby/yarp/commit/1ad7fba5ef
-
- https://github.com/ruby/yarp/commit/8bffb8a762
-
11:08 PM Revision 8db3e3c3 (git): [ruby/yarp] require constant pool capacity to be a power of 2
- https://github.com/ruby/yarp/commit/dea8d3f29f
-
- Since the compile-time iseq used in the guard was not marked and updated
during compaction, a runtime value reusing the address could falsely pass
the guard.
Co-authored-by: Takashi Kokubun <takashikkbn@gmail.com> -
- https://github.com/ruby/yarp/commit/4ba6d5ca70
-
- https://github.com/ruby/yarp/commit/35da3d1a4c
-
- https://github.com/ruby/yarp/commit/c8caa997c0
- 10:36 PM Revision 35960ce6 (git): [ruby/yarp] Add node event dispatcher
- This commit changes the node template to create a dispatcher class,
which can be used to walk an AST an emit events to all registered
listeners
https://github.com/ruby/yarp/commit/03a45f85e6
Co-authored-by: Kevin Newton <kddnewton@user... -
- Starting in ruby 3.3.0-preview2, attempting to require bigdecimal, mutex_m, base64, or other libraries that will move to bundled gems in Ruby 3.4.0 results in a warning, even when bundler is not in use:
```
$ ruby33 -v -r bigdecimal ... -
10:25 PM Revision 89bd1ebc (git): [ruby/yarp] Fix unexpected { after method call with args
- https://github.com/ruby/yarp/commit/fe279d2d16
-
- https://github.com/ruby/yarp/commit/9222faa1c6
-
- https://github.com/ruby/yarp/commit/a8b54e8ed0
-
- https://github.com/ruby/yarp/commit/d931e258d1
- 10:22 PM Revision bbf9f11c (git): [ruby/yarp] Fix behaviour of locations for comments
- https://github.com/ruby/yarp/commit/b1ced67fba
-
10:01 PM Revision 010017d8 (git): [rubygems/rubygems] Reduce array allocations and concatenations in Index
- Remove the default nested hash in Index entirely
Index#search_all now yields or returns enum since that's what caller
needs.
https://github.com/rubygems/rubygems/commit/c45ea3bbe2 -
10:01 PM Revision 2cf5fe58 (git): [rubygems/rubygems] Avoid allocating empty hashes in Index
- Since the hashes have a default proc that returns a (new) empty hash, we
can avoid allocating those empty hashes when we are only doing lookups.
Test from running `bundle update --bundler` against a rails app I have
lying around:
```
=... - 09:47 PM Revision 4e8869c6 (git): [ruby/yarp] Use templated constants for Regexp options
- https://github.com/ruby/yarp/commit/38e1769c2e
- 09:47 PM Revision 101ac364 (git): [ruby/yarp] Return Regexp options that match MRI for e, u, s, and n
- https://github.com/ruby/yarp/commit/17dbf4ec46
-
-
- > this branch is bit of unstable.
If so, the sooner it's merge the sooner we can help stabilize it, and the more time we have to do so before release. -
- variable
(https://github.com/ruby/yarp/pull/1521)
https://github.com/ruby/yarp/commit/3ec0699622 -
05:59 PM Revision e50b7bf7 (git): [DOC] Globals doc (#8445)
-
- Running `bundle update --bundler` on a rails app locally:
```
==> memprof.after.txt <==
Total allocated: 301.90 kB (3794 objects)
Total retained: 73.24 kB (698 objects)
==> memprof.before.txt <==
Total allocated: 14.47 MB (196378 obje... -
05:24 PM Revision 4d86d932 (git): Free all heap pages at shutdown
- previously heap_allocated_pages was decremented from heap_page_free causing only half the heap pages to be freed at shutdown
-
04:01 PM Revision a1dc1a3d (git): Return line 0 for JIT frames
- Frames pushed by YJIT have an unreliable PC. The PC could be garbage,
and if we try to read the line number with a garbage PC, then the
program can crash.
This commit returns line 0 for programs where there is a `jit_return`
function. ... -
- Similar to releasing free GC pages, releasing free malloc pages
reduce the amount of page faults post fork. -
- https://github.com/ruby/yarp/commit/1ddd74dc11
-
- https://github.com/ruby/yarp/commit/22130b3491
-
- https://github.com/ruby/yarp/commit/eaafc829db
-
- https://github.com/ruby/yarp/commit/d73f7c9cfa
-
- https://github.com/ruby/yarp/commit/5fd4d3b89a
-
- https://github.com/ruby/yarp/commit/65b536ba12
-
- https://github.com/ruby/yarp/commit/ffc8f35e56
-
- https://github.com/ruby/yarp/commit/99c91931e0
-
- https://github.com/ruby/yarp/commit/4a30c69051
-
01:13 PM Bug #18257: rb_mRubyVMFrozenCore is broken by GC run
- Trying to reproduce all the steps, I still observe the behavior as described in comment #18257-12. I'll keep around the patch as described in #18257-14 until this is fixed properly.
-
12:32 PM Bug #19875: Ruby 3.0 -> 3.1 Performance regression in String#count
- nobu (Nobuyoshi Nakada) wrote in #note-7:
> These are all generated by the same compiler?
Yes - FreeBSD clang version 14.0.5. I also try 15.0.7 to no effect.
gcc 12.2.0 performs much better, though it too has a slightly more mo... -
- Attaching the generated code by gcc version 11.4.0 (Ubuntu 11.4.0-1ubuntu1~22.04).
This seems vectorizing per 16 bytes.
```S
movdqu (%rdi), %xmm2 # MEM <vector(16) unsigned char> [(unsigned char *)_95], tmp383
addq $16, %rdi #, i... -
- Freaky (Thomas Hurst) wrote in #note-6:
> Annotated disassemblies attached.
Thank you.
These are all generated by the same compiler?
The "slow" version scans per WORD (= 2bytes), while the "fast" version uses DWORD (= 4bytes).
-
- Annotated disassemblies attached.
-
- byroot (Jean Boussier) wrote in #note-4:
> > May you please help me to find the root cause?
> ...
Tried `objdump --disassemble=rb_str_count` on Ubuntu 22.04, but the shared library is stripped and has no symbols.
> @alanwu guessed t... -
- Applied in changeset commit:git|89802078f9f406be411032814e1960e62dbc7ce2.
----------
[Bug #19882] Reject tokens invalid as symbols -
- https://github.com/ruby/ruby/pull/8450
-
- It's subtly intentional.
The former seemed better to me.
```
$ ruby -e '$01234'
-e:1: identifier $01234 is not valid to get
-e: compile error (SyntaxError)
```
```
$ ./ruby -e 'p :$01234'
-e:1: syntax error, unexpected integ... -
09:57 AM Feature #19884: Make Safe Navigation Operator work on classes
- I am not sure that pattern is that common to warrant using &., aside from the syntax not being very elegant, but that may be my own bias against it (syntax-wise that is). I don't seem to need to do "defined?(Foo::Bar) && Foo::Bar.some_me...
-
08:11 AM Feature #19884: Make Safe Navigation Operator work on classes
- Currently, you can do `ActiveRecord::Base.some_method rescue nil`
Using `&.` for rescuing an error in addition to the current behavior would make it confusing.
Furthermore, the specification you are asking for is not clear. What Except... -
08:06 AM Feature #19884 (Open): Make Safe Navigation Operator work on classes
- If a constant isn't defined it will raise a NameError:
```ruby
DoesNotExist.some_method # raises: uninitialized constant DoesNotExist (NameError)
```
In libraries that have optional dependencies, we can check if the constant is d... -
- I opened a PR about a month ago but should I close it...? https://github.com/ruby/ruby/pull/8242
-
- 08:05 AM Revision 5a7f5bb0 (git): Bump actions/checkout from 3.6.0 to 4.0.0
- Bumps [actions/checkout](https://github.com/actions/checkout) from 3.6.0 to 4.0.0.
- [Release notes](https://github.com/actions/checkout/releases)
- [Changelog](https://github.com/actions/checkout/blob/main/CHANGELOG.md)
- [Commits](http... -
- # The next dev meeting
**Date: 2023/10/12 13:00-17:00** (JST)
Log: *TBD*
- Dev meeting *IS NOT* a decision-making place. All decisions should be done at the bug tracker.
- Dev meeting is a place we can ask Matz, nobu, nurse and o... -
-
- Applied in changeset commit:git|fe0225ff4d5af8b1f54009727b39d0d9b821eea3.
----------
[Bug #19778] Add `-I` options for opt-dir to `$INCFLAGS`
These options have been separated from `$CFLAGS` already in the
other places. -
- https://github.com/ruby/ruby/pull/8449
-
- These options have been separated from `$CFLAGS` already in the
other places. - 03:20 AM Revision 74277f99 (git): Bump ruby/setup-ruby from 1.152.0 to 1.153.0
- Bumps [ruby/setup-ruby](https://github.com/ruby/setup-ruby) from 1.152.0 to 1.153.0.
- [Release notes](https://github.com/ruby/setup-ruby/releases)
- [Commits](https://github.com/ruby/setup-ruby/compare/250fcd6a742febb1123a77a841497ccaa8... -
02:08 AM Bug #9115: Logger traps all exceptions; breaks Timeout
- Eregon (Benoit Daloze) wrote in #note-12:
> mame (Yusuke Endoh) wrote in #note-11:
> ...
Right. It's never an ultimate fix. However considering the way a log destination is designed generally, I think it's rare for that blocking IO t... -
01:33 AM Bug #19873 (Closed): Thread#native_thread_id is incorrectly cached across fork on Linux - Applied in changeset commit:git|0117a6d389c5b607eaec1b7f917132efa356f665.
----------
Fix Thread#native_thread_id being cached across fork (#8418)
The native thread ID can and does change on some operating systems (e.g.
Linux) after for... - 01:33 AM Revision 0117a6d3 (git): Fix Thread#native_thread_id being cached across fork (#8418)
- The native thread ID can and does change on some operating systems (e.g.
Linux) after forking, so it needs to be re-queried.
[Bug #19873]
09/14/2023
-
11:31 PM Feature #19842: Introduce M:N threads
- @tenderlovemaking We can merge this with ko1's convenience. We needed to skip this for preview2 because ko1 said this branch is bit of unstable.
-
- Any idea when / if we will merge this? Or, is there anything preventing us to merge?
-
- It seems not-uncommon for methods to have no IV, ISE, or ICVARC caches.
Calling malloc with 0 will actually allocate something, so if there
aren't any caches (`ISEQ_IS_SIZE(body) == 0`), then we can avoid
allocating memory by not calling... -
- * YJIT: Skip Insn::Comment and format!
if disasm is disabled
Co-authored-by: Alan Wu <alansi.xingwu@shopify.com>
* YJIT: Get rid of asm.comment
---------
Co-authored-by: Alan Wu <alansi.xingwu@shopify.com> -
- /yjit/src/backend/x86_64/mod.rs Is also UTF-8 and it doesn't have the
marker. The standard recommends against it, so remove it. -
- to fix CI failures like:
https://github.com/ruby/ruby/actions/runs/6190995187/job/16808382442 -
- Previously, TestStack#test_machine_stack_size failed pretty consistently
on ARM64 macOS, with Rust code and part of the interpreter used for
per-instruction fallback (rb_vm_invokeblock() and friends) touching the
stack guard page and cra... -
- Increase test coverage and be explicit about what is and is not supported.
https://github.com/rubygems/rubygems/commit/a096397a00 -
- This simplifies how we handle multi-targets, and also fixes a bug we
had where for loops were always getting multi-targets, even when there
was only a single target.
https://github.com/ruby/yarp/commit/31eb8b7ad5 -
- This rarely used node holds information about the local variables
that need to get written in the case a regular expression is used on
the left-hand side of a =~ operator and it has named capture groups.
Note that we already "handled" t... -
- If the read_name is invalid, we shouldn't try to set a write name.
https://github.com/ruby/yarp/commit/06881c8ca7 -
- Don't overlook the colon!
-
- I think it is rejected at the moment:
```
-e:1: identifier $0x is not valid to get
-e: compile error (SyntaxError)
``` -
- ```
irb(main):001> :$0x
=> :"$0x"
```
Since `$0x` is not a valid global variable name, I think it should be rejected unless quotation marks are used. -
- https://github.com/ruby/yarp/commit/93bec2c173
-
- https://github.com/ruby/yarp/commit/7f71527522
-
- https://github.com/ruby/yarp/commit/db28e22363
-
- https://github.com/ruby/yarp/commit/0d3e8cdcd7
-
- https://github.com/ruby/yarp/commit/d3a852dac2
-
- https://github.com/ruby/yarp/commit/890fa72fdd
-
- https://github.com/ruby/yarp/commit/09248a1f9e
-
- https://github.com/ruby/yarp/commit/c61658c0fd
-
- https://github.com/ruby/yarp/commit/373898e7cb
-
- https://github.com/ruby/yarp/commit/21a43034d0
-
- * [YARP] Use ALIAS_METHOD as opposed to ALIAS
* [YARP] Use YP_TRUE_NODE instead of YP_NODE_TRUE_NODE -
- My uncertain understanding is that an expression (called "arg" in parse.y) must come after the unary minus. Since `foo 1, 2, 3` is not an expression but a statement (called "command" in parse.y), `- foo 1, 2, 3` is not acceptable.
-
- At the moment it appears that you can't use unary operators when you have a call without parentheses. For example:
``` ruby
- foo 1, 2, 3
```
This seems useful, particularly for methods that perform calculations.
I'm also aski... -
-
- Applied in changeset commit:git|e8896a31d48e5797df3878696dcb50aed85b87c2.
----------
[Bug #19877] Literals cannot have singleton methods even in blocks -
04:23 PM Feature #19720: Warning for non-linear Regexps
- I agree with Matz, false positives are worse than no warnings at all. I use atomic groups and positive/negative lookahead/lookbehind all the time. In particular, atomic groups are a great tool to *prevent* ReDoS-susceptible regexps, so t...
-
10:22 AM Feature #19720: Warning for non-linear Regexps
- If "non-linear" means real non-linear, I would not object. But in this case, “non-linear” means the recent memoize optimization cannot be applied. So if we add this warning, we will get a lot of false positives from safe regular expressi...
-
-
-
-
-
- Following up on #1483, this removes our usage of varargs entirely.
Under the new approach we make explicit methods for the various
signatures. Because we know the type of everything any it's very
consistent, this makes it easier for comp... -
- mame (Yusuke Endoh) wrote in #note-11:
> As a better-than-nothing mitigation, it is proposed to enclose the entire `Logger::LogDevice#write` in `Thread.handle_interrupt(:never) { ... }`.
> ...
This sounds problematic, especially since ... -
- Discussed at the dev meeting.
* It is clear that Logger has the intention not to interfere with the execution of the application on a log write failure.
* On the other hand, it is also clear that Logger is not considering about async... -
- Briefly discussed at the dev meeting.
We understand the problem, but doesn't find a good solution. Since the call line of autoload does not exist in the call stack when the exception occurs, it cannot be displayed as a backtrace.
It ... -
- We discussed this at the dev meeting, and decided not to implement it.
Considering frameworks that implement a thread pool, the time when a Thread object is created is not necessarily useful, but could be rather confusing. It would be... -
08:22 AM Feature #19850: Get thread creation time
- To investigate when you want to investigate what causes a program hang...
## pid2line
https://gist.github.com/nurse/0619b6af90df140508c2
It shows what threads are running or sleeping and where it is as C function name.
In your ca... -
-
- I confirmed that, at least, Java, Python, JavaScript, Rust, .NET, PHP, and OCaml are using starts_with/ends_with or similar.
https://docs.oracle.com/en/java/javase/20/docs/api/java.base/java/lang/String.html#startsWith(java.lang.String)... -
- We have picked that pair (inspired by Java). Adding infinite number of variation seems not to be a good idea.
Matz.
-
- https://github.com/ruby/yarp/commit/6be28eeed0
-
- https://github.com/ruby/yarp/commit/fb141b8d07
-
- https://github.com/ruby/yarp/commit/c5a1094988
-
- https://github.com/ruby/yarp/commit/55446f7a43
-
-
- https://github.com/nobu/ruby/tree/Range%23overlap_p
-
- I think the analogue for `range1.overlap?(range2)` in mathematics is `interval1 ∩ interval2 ≠ ∅`, which does not hold when either `interval1` or `interval2` is an empty interval.
Hence, I think that, when either `range1` or `range2` is ... -
- Although I said “accepted”, we found some corner cases which are not clear (e.g. #note-11). We have to make these cases clear before merging it to the core.
Matz.
-
08:36 AM Feature #19839: Need a method to check if two ranges overlap
- I'm not sure what `(n...n)` means but on AS definition, the following code return true.
```ruby
p (1..2).overlap?(2...2) #=> true
p (2..2).overlap?(2...2) #=> true
p (2...2).overlap?(2...2) #=> true
```
is it intentional?
-
- Accepted. It must be useful without ActiveSupport.
Matz.
-
- Alphabetically by class or module names.
-
- https://github.com/ruby/yarp/commit/22c800eab1
-
- I would like to add a point that came up during the dev meeting: if a thread is waiting for `Queue#pop`, and another thread freezes the Queue, what should the first thread do? Wait forever? Raise a FrozenError?
To raise a FrozenError,... -
- Since frozen `queue` is meaningless, if you care about consistency, we should prohibit `freeze` queues, like `ENV`.
Matz.
-
- https://github.com/ruby/yarp/commit/4512daf769
-
- https://github.com/ruby/yarp/commit/79794b4833
-
- https://github.com/ruby/yarp/commit/e807462c6d
-
09:47 AM Feature #19735: Add support for UUID version 7
- We core committers discussed this issue on DevMeeting, and we accepted it.
Go ahead - 08:37 AM Revision a095740f (git): Fix `io_buffer_get_string` default length computation. (#8427)
- * Fix `io_buffer_get_string` default length computation.
When an offset bigger than the size is given, the resulting length will be
computed incorrectly. Raise an argument error in this case.
* Validate all arguments. -
- It should be fundamentally covered by the reference document.
Matz.
-
- A license term is not easy to change.
Also note that Ruby consists from many code which use difference license.
Therefore I think changing license text to match with SPDX is not a good proposal.
-
- Applied in changeset commit:git|b6de0a6c69a4857ca4347f65d7c9a5cb6e52c5bd.
----------
[Bug #19868] Suggest other Process::Status method for `&` and `>>`
`Process::Status#&` and `Process::Status#>>` are provided only for
the backward com... -
-
- `Process::Status#&` and `Process::Status#>>` are provided only for
the backward compatibility with older Ruby than 1.8 where `$?` was
a `Fixnum`, and the knowledge about internals of system dependent
macros is necessary to use them. Mod... -
- Currently, it gives the place we called `autoload`. If you want the place where the loaded module defined (after autoload), you can explicitly reference the module before calling `const_source_location`. So I vote for keeping it as it is...
-
-
- We reject `bsearch_last`, it's not intuitive (from my POV). If we come up with good name candidates, resubmit the proposal.
Matz.
-
- https://github.com/ruby/yarp/commit/777c376deb
09/13/2023
-
- I'm trying to debug a flaky `RuntimeError: nested #it` failure in
ruby/spec. Hopefully the full backtrace will give some clues.
Last occurence:
https://github.com/ruby/ruby/actions/runs/6172578817/job/16753137038 -
- https://github.com/ruby/yarp/commit/87dbb6cf1f
-
- YJIT: Skip adding past_pages_bytes for past pages
-
07:16 PM Feature #19832: Method#destructive?, UnboundMethod#destructive?
- IMO this feature wouldn't really add anything new to ractors that I can think of, because ractors already work with either new "moved" objects or deeply frozen objects, which can still have methods called on them inside the ractors. For ...
-
05:37 PM Revision 7f6c0efa (git): Fix hyperlink for Script Encoding
-
- > May you please help me to find the root cause?
A start would be to decompile `rb_str_count` on two versions.
@alanwu guessed that it might be the change from signed to unsigned that is preventing some compilers from auto-vectoriz... -
04:22 PM Bug #19875: Ruby 3.0 -> 3.1 Performance regression in String#count
- Thanks for your feedback, @nobu! It looks like compiler dependent. I used `x86_64-linux`.
May you please help me to find the root cause? (*I'll do my own research, just asking for starting points.* )
-
- It seems compiler dependent.
My results on x86_64-darwin are:
```
$ ruby -I ./benchmark/benchmark-driver/lib/ benchmark/benchmark-driver/exe/benchmark-driver ~/tmp/string_count_benchmark_driver.yml --executables=/opt/local/bin/ruby{... -
-
- There's a missing write barrier for operands in the iseq instruction list, which can cause crashes. This bug has been fixed in commit [b3b57f7](https://github.com/ruby/ruby/commit/b3b57f70cc1ee6f40ff10b2abaa51889abce2331).
It can be r... -
-
-
- I don't think this is resolved, at least to the full extent. Trying again with the patch from comment #18257-13 and this is the results:
~~~
$ ./miniruby -v
ruby 3.3.0dev (2023-09-05 master 7c8932365f) [x86_64-linux]
$ ./miniruby -e "1... -
- I said this could be useful for debugging. If a user finds a call to `autoload` on the line pointed to by `const_source_location`, they can guess that autoload is set. If they want to know the actual definition location, they can actuall...
-
- mame (Yusuke Endoh) wrote in #note-16:
> Why do you want to know the location of the autoload call after it is actually loaded?
I do not know. I do not understand in the first place your use case where you want to identify where a co... -
- Why do you want to know the location of the autoload call after it is actually loaded? I guess that information is already gone from the runtime as well.
-
- mame (Yusuke Endoh) wrote in #note-14:
> As I said, you can use it safely in conjunction with `Module.autoload?`.
> ...
What I meant is that, once you are in a location beyond where `A` was called, you cannot reference the call site an... -
- sawa (Tsuyoshi Sawada) wrote in #note-13:
> You cannot safely use `const_source_location` to extract the autoload call site since the return value depends on whether the constant is called or not:
As I said, you can use it safely in ... -
- mame (Yusuke Endoh) wrote in #note-11:
> 2. If we want to identify where a constant is set as autoload, we can use the current `Module#const_source_location`.
> ...
You cannot safely use `const_source_location` to extract the autoload ca... -
- https://github.com/ruby/yarp/commit/820774976d
-
- https://github.com/ruby/yarp/commit/eaaebc17c8
-
- https://github.com/ruby/yarp/commit/b74ce35379
-
- Applied in changeset commit:git|411572661a3995a0daacb6c866ea3ac6b52ad25b.
----------
math.c: Fix Math.log against huge bignum [Bug #19878] -
- Fortunately, this bug was introduced this year on commit:da39936ce165ea9462b9e192eb6b608485c94842 so no need to backport.
-
- https://github.com/ruby/ruby/pull/8429
-
- ```ruby
p Math.log(2 ** 1020, 10) #=> 307.0505955772608
p Math.log(2 ** 1021, 10) #=> 307.35162557292483
p Math.log(2 ** 1022, 10) #=> 307.6526555685888
p Math.log(2 ** 1023, 10) #=> 688.6994720982339 <- WHAT!?
p Math.log(2 ** 1024,... -
- Co-Authored-By: Nobuyoshi Nakada <nobu@ruby-lang.org>
-
-
- ```
/opt/hostedtoolcache/Ruby/3.1.3/x64/lib/libruby.so.3.1(objspace_xmalloc0.cold+0x0) [0x7f07a3fbfa2f]
[0x7f079983999e]
/opt/hostedtoolcache/Ruby/3.1.3/x64/lib/libruby.so.3.1(obj_free+0x632) [0x7f07a4080282]
/opt/hostedtoolcache/Rub... -
10:42 AM Bug #19879 (Third Party's Issue): GC crash during Rails app testing
- In a Github action (using https://github.com/ruby/setup-ruby latest) for a rails app, I am seeing the following error when trying to run the test suite (RSpec). The tests have been running OK for a while, but recently this started happen...
-
09:44 AM Bug #19754: `IO::Buffer#get_string` raises unsuitable exception for too large offset
- It also begs the question whether `NUM2SIZET` should allow negative values..
-
- This change isn't sufficient, it just protects against the invalid value of offset. We can backport it, but I don't feel strongly about whether it's required or not.
-
- Non intuitive behavior of syntax only applied to literal value
Some ruby syntax is only applied to literal value.
~~~ruby
def 1.foo; end # receiver is a literal, it is Syntax Error
/(?<a>)/ =~ s # receiver is regexp literal, it wil... -
- "defensive" code like that is just broken, no one should `rescue Exception`.
Let's remove that outer `rescue`?
If `synchronize` raises there is a serious bug worth fixing and not ignoring.
Similarly I think the inner `rescue`s shoul... -
09:03 AM Revision 4655d210 (git): Lrama v0.5.6
-
- @mame Thanks for the clarifications, notably about the name.
My impression is if YARP supports multiple Ruby grammar versions then basically everything discussed in this issue is solved (and no need to rename, and only C files of yarp a... -
-
- Have not chance to test it yet, but the change LGTM. Thx!
-
- Applied in changeset commit:git|234722ed9fce33d7ec8152920636af663d527912.
----------
[Bug #19872] Refine TestRequireLib
Always test only the scripts just under “lib", and just under
child directories which has not the same name script ... -
- https://github.com/flori/json/commit/573ef94bc5
-
- FYI: Julia has `searchsortedfirst` and `searchsortedlast`
https://docs.julialang.org/en/v1/base/sort/#Base.Sort.searchsortedfirst -
- (In the following, searching for the last element will be referred to as "LAST".)
There are often cases where searching for the last element would be useful.
A typical example that immediately comes to mind is as follows:
* When w... -
05:58 AM Feature #19839: Need a method to check if two ranges overlap
- I believe there are a few reasons for this addition. The core one I see often in justifications is precedence in existing Ruby, such as:
**Combinations**
These cases have clear precedent for iterable / enumerable types:
* `Array... -
- Honestly, I can't give a formal reason. But I do believe it's a very basic operation. For example, postgres and boost provide such functionality.
- https://www.postgresql.org/docs/15/functions-range.html
- https://www.boost.org/doc/l... -
- And add special treats only for old GNU make.
-
- * [Bug #17146] Queue operations are allowed after it is frozen (jeremyevans0)
* Last discussed at October 2020 developer meeting, no decision made.
* @headius and I discussed this at RubyKaigi and agreed it would be best to raise F... -
- Always test only the scripts just under “lib", and just under
child directories which has not the same name script in the upper
level; instead of random sampling from whole libraries. -
-
- Okay. I will make this behavior match.
-
- I think it is by design. They are marked as `%nonassoc` in parse.y.
https://github.com/ruby/ruby/blob/6e64d4370456190541705ec4c6cf3af6bf4ac647/parse.y#L1698
```
%nonassoc tCMP tEQ tEQQ tNEQ tMATCH tNMATCH
``` -
- It appears that you cannot chain `==`, `!=`, `<=>`, `=~`, `!~`, `===`:
``` ruby
foo == bar == baz
```
but you can chain all of the other operators:
``` ruby
1 << 2 << 3
```
I'm not sure if I've missed some. I would just l... -
- https://github.com/ruby/yarp/commit/943204d56f
09/12/2023
-
- Applied in changeset commit:git|6e64d4370456190541705ec4c6cf3af6bf4ac647.
----------
[Bug #19862] Skip compiled result of never reachable expression -
-
-
- Applied in changeset commit:git|19346c2336053b351673da030b00c704138252d8.
----------
[Bug #19754] Make `IO::Buffer#get_string` check `offset` range (#8016) -
-
-
- It's probably not obvious given how existing Ruby code traditionally works, but in the above code, `a` and `b` are zero copy (which is ideal).
```
irb(main):001> data = "foo"
=> "foo"
irb(main):002> IO::Buffer.for(data)
=>
#<IO:... -
- I would add that if the yarp gem had straight multi-version support like the parser gem, such a mechanism is not even needed.
-
- Hi @kddnewton
kddnewton (Kevin Newton) wrote in #note-22:
> From this thread and others, this is my understanding of the requirements:
> ...
There may be a communication error. More precisely,
* The `YARP` name is not good as a built-i... -
- It failed with `NameError` because `Html3` is defined in the file that
`CGI::HtmlExtension` autoloads. -
-
- Ensure unrecognized SPECS types are ignored
https://github.com/rubygems/rubygems/commit/5b33e91075 -
- * YJIT: Add compilation time counter
* YJIT: Use Instant instead -
- Into AliasGlobalVariableNode and AliasMethodNode. These have different
enough semantics that we feel comfortable splitting them up.
https://github.com/ruby/yarp/commit/c1f3e6d344 -
- Applied in changeset commit:git|39ee3e22bd3d071c1c283b6b8dbd1af413342fb1.
----------
Make Kernel#lambda raise when given non-literal block
Previously, Kernel#lambda returned a non-lambda proc when given a
non-literal block and issued a... -
- Previously, Kernel#lambda returned a non-lambda proc when given a
non-literal block and issued a warning under the `:deprecated` category.
With this change, Kernel#lambda will always return a lambda proc, if it
returns without raising.
... -
-
-
- `String#count` became slower since Ruby 3.1. Originally found by `@Freaky`: https://github.com/ruby/ruby/pull/4001#issuecomment-1714779781
Compared using the [`benchmark-driver` gem](https://github.com/benchmark-driver/benchmark-drive... -
- (X)StringNodes
(https://github.com/ruby/yarp/pull/1427)
Prior to this commit, heredocs were automatically InterpolatedNodes
regardless of whether there was actually interpolation. With this
commit, heredocs are only interpolate if there... -
01:27 PM Feature #19874 (Closed): Re-introduce heap sort by pinned count prior to GC compaction
- See Github PR: [#8420](https://github.com/ruby/ruby/pull/8420)
<hr />
## Introduction
In Ruby 2.7 compaction was originally introduced as an experimental feature.
There was no auto-compaction, and compaction had to be manually ... -
- I've backported @nobu's branch on our 3.2 rubies and it's been working very well. Not too sure what is missing.
-
-
- * See https://github.com/ruby/yarp/issues/1436
https://github.com/ruby/yarp/commit/6f4e9ff940 -
-
09:55 AM Feature #19871: Add __owner__
- mame (Yusuke Endoh) wrote in #note-2:
> Could you explain the use case?
It's supposed to make sure a general-use module can be implemented purely and can't be tainted by any of its inheritors, unless it's deliberately patched of cour... -
- nobu (Nobuyoshi Nakada) wrote in #note-3:
> Why not use the constant without scopes?
> ...
I was expecting the lookup to still be affected by the inheriting class. I should have tested it first.
I guess this invalidates the feature... -
- Why not use the constant without scopes?
```ruby
module M
C = 1234
def m
puts "C: #{C}" #=> C: 1234
end
end
``` -
- Could you explain the use case?
-
- Thanks. Basically, we didn't add new methods from ActiveSupport without performance reason.
Can you describe why we should add `overlap?` to ruby core? -
- Though ActiveSupport's `overlaps?` method is enough for our use-case. I think the feature is so fundamental that it should be supported by the standard library.
Slightly off-topic: It might be better to add a method that returns the o... -
- ActiveSupport already have `overlaps?` method.
* https://api.rubyonrails.org/classes/Range.html#method-i-overlaps-3F
* https://github.com/rails/rails/blob/main/activesupport/lib/active_support/core_ext/range/overlap.rb#L7
Is it en... -
- I'm not sure why you use local installation of OpenSSL 3. I recommend to use libssl provided by Ubuntu 22.04.
-
- We should avoid to non-existence directory for fresh install of macOS.
@nobu Can you investigate this? -
- Maybe this is becaue the test of `TestRequireLib` does sampling: https://github.com/ruby/ruby/blob/f1c78b23313e63eb31a213cc7277b1eb206a581b/test/ruby/test_require_lib.rb#L13
@znz Should we fix the random seed at least? @nobu says that... -
-
- * [Bug #19868] `Process::Status` methods for compatibility with `Fixnum` (nobu)
* Looking at commit:54274b8c65a0981f1c69055a1513ba3c614dd675, noticed that `pst_rshift` contains shift by a negative integer, an unspecified behavior in C9... -
- I think the workaround is good enough. Do you have a good use case to introduce this?
The proposed behavior is not something that can simply be called "last".
The traditional behavior requires a given block to return `false` for elem... -
03:58 AM Bug #19873 (Closed): Thread#native_thread_id is incorrectly cached across fork on Linux
- On Linux, when a process forks, the native thread ID of the (only) thread in the forked child should change.
```
>> libc = Fiddle.dlopen(nil)
=> #<Fiddle::Handle:0x0000ffffb0d8e900>
> ...
=>
#<Fiddle::Function:0x0000ffffb15d1130
...
09/11/2023
-
- https://github.com/ruby/yarp/commit/45dd046b83
-
- These are replacements for regular expressions when they are used
alone as the predicate of a conditional. That's because they are
significantly different from a regular expression because they are
not evaluated for truthyness, but inste... -
-
-
- https://github.com/ruby/yarp/commit/c2b9b780c7
- 09:06 PM Revision 203fdd73 (git): [rubygems/rubygems] Fixed include realpath in error statement
- https://github.com/rubygems/rubygems/commit/ac3b85bd5e
- 09:06 PM Revision 8bb61077 (git): [rubygems/rubygems] Fixed false positive SymlinkError in symbolic link directory
- https://github.com/rubygems/rubygems/commit/58173ff2ea
-
-
- The flags integer is an implementation detail. We want people to
use the query methods to access the individual fields so we are
freed from having to maintain a specific order. As such, this
commit changes the Ruby API to mark all flags ... - 03:25 PM Revision 05a853c2 (git): [ruby/yarp] Update fixtures
- https://github.com/ruby/yarp/commit/884f2ca8e4
- 03:25 PM Revision 5a8767ed (git): [ruby/yarp] Fix order of Regexp flags
- https://github.com/ruby/yarp/commit/e421305ea2
- 03:25 PM Revision 689dffc8 (git): [ruby/yarp] Add failing test for Regexp flags
- https://github.com/ruby/yarp/commit/16fe179c5f
-
- vo.x (Vit Ondruch) wrote in #note-1:
> Or is it issue of the respective libraries, which does not support proper loading of nested files, without loading their dependencies first?
Randomly looking at lib/reline/history.rb, maybe the ... -
- Maybe predefining some well know namespaces could help? Or is it issue of the respective libraries, which does not support proper loading of nested files, without loading their dependencies first?
-
- Just exploring the difference in skipped test cases, it seems to be due to `TestRequireLib#test_thread_size`. One test run might look like:
~~~
84) Skipped:
TestRequireLib#test_thread_size:/builddir/build/BUILD/ruby-3.3.0-7c8932365... -
-
-
09/10/2023
-
- The changes needed to implement this turned out to be just simple so I went ahead and just created a PR in https://github.com/ruby/ruby/pull/8411.
The difference between using `method(__method__).owner` and using `__owner__` can be de... -
- Which will give the owner of the currently executing method.
`method(__method__).owner` or `method(__callee__).owner` isn't enough since it may return a different owner if the method is overridden.
Relying on calling the name of it... -
- 1048936 - 360 = 1048576 = 0x100000
A terminated thread no longer keeps its stack area.
-
05:56 AM Misc #19870 (Feedback): Why is a long-run thread allocated more memory than a short-run one?
- I want to write multithreading programing in Ruby but I found the following memory consumptions.
```
$ irb
irb(main):001> require 'objspace'
=> true
irb(main):002> t1 = Thread.new {}
=> #<Thread:0x00007f9237a74b28 (irb):2 run>
i... -
- So that it is not ncessary to specify the type each time.
-
-
-
10:39 AM Revision 0e49f817 (git): [DOC] Fix method docs of scheduler#io_read and scheduler#io_write
- Each of Fiber::Scheduler#io_read and io_write takes the "offset" as the fourth argument, which is not doucmented.
-
- The word "and" was missing.
-