[#116016] [Ruby master Bug#20150] Memory leak in grapheme clusters — "peterzhu2118 (Peter Zhu) via ruby-core" <ruby-core@...>
Issue #20150 has been reported by peterzhu2118 (Peter Zhu).
7 messages
2024/01/04
[#116382] [Ruby master Feature#20205] Enable `frozen_string_literal` by default — "byroot (Jean Boussier) via ruby-core" <ruby-core@...>
Issue #20205 has been reported by byroot (Jean Boussier).
77 messages
2024/01/23
[ruby-core:116478] [Ruby master Bug#20188] `Module#const_source_location` returns wrong information when real constant was defined but autoload is still ongoing
From:
"fxn (Xavier Noria) via ruby-core" <ruby-core@...>
Date:
2024-01-29 15:56:42 UTC
List:
ruby-core #116478
Issue #20188 has been updated by fxn (Xavier Noria).
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 would be no deadlock, right? Same with `require`, if we try this without autoloads:
```ruby
File.write('/tmp/bar.rb', <<~RUBY)
Fiber.yield
RUBY
Fiber.new { require '/tmp/bar.rb' }.resume
Fiber.new { require '/tmp/bar.rb' }.resume
```
we get a deadlock.
So that is my point, what we say about threads does not translate to fibers. In fibers, you can deadlock.
`Kernel#require` is thread-safe, things that would typically trigger a context switch (like I/O or `sleep`) won't make you deadlock if the other threads are waiting for that `require`. And same for `autoload`. And makes sense, because Ruby owns the thread schedulur mod hints API.
> fibers being cooperative
That is less the case with fiber schedulers, no?
Can Falcon trigger deadlocks?
----------------------------------------
Bug #20188: `Module#const_source_location` returns wrong information when real constant was defined but autoload is still ongoing
https://bugs.ruby-lang.org/issues/20188#change-106502
* Author: byroot (Jean Boussier)
* Status: Open
* Priority: Normal
* ruby -v: ruby 3.3.0 (2023-12-25 revision 5124f9ac75) [arm64-darwin23]
* Backport: 3.0: UNKNOWN, 3.1: UNKNOWN, 3.2: UNKNOWN, 3.3: UNKNOWN
----------------------------------------
Ref: https://github.com/fxn/zeitwerk/issues/281
`const_source_location` keeps returning the location of the `autoload` call, even after the real constant was defined. It only starts returning the real constant location once the autoload was fully completed.
```ruby
# /tmp/autoload.rb
File.write("/tmp/const.rb", <<~RUBY)
module Const
LOCATION = Object.const_source_location(:Const)
end
RUBY
autoload :Const, "/tmp/const"
p Const::LOCATION
```
Expected Output:
```ruby
["/tmp/const.rb", 1]
```
Actual Output:
```ruby
["/tmp/autoload.rb", 8]
```
Potential patch: https://github.com/ruby/ruby/pull/9549
--
https://bugs.ruby-lang.org/
______________________________________________
ruby-core mailing list -- ruby-core@ml.ruby-lang.org
To unsubscribe send an email to ruby-core-leave@ml.ruby-lang.org
ruby-core info -- https://ml.ruby-lang.org/mailman3/postorius/lists/ruby-core.ml.ruby-lang.org/