ruby-core@ruby-lang.org archive (unofficial mirror)
 help / color / mirror / Atom feed
From: "ioquatix (Samuel Williams) via ruby-core" <ruby-core@ml.ruby-lang.org>
To: ruby-core@ml.ruby-lang.org
Cc: "ioquatix (Samuel Williams)" <noreply@ruby-lang.org>
Subject: [ruby-core:117456] [Ruby master Bug#20413] Enumerator can block fiber scheduler.
Date: Sun, 07 Apr 2024 12:48:06 +0000 (UTC)	[thread overview]
Message-ID: <redmine.journal-107841.20240407124806.3344@ruby-lang.org> (raw)
In-Reply-To: redmine.issue-20413.20240407104757.3344@ruby-lang.org

Issue #20413 has been updated by ioquatix (Samuel Williams).


After discussing it with Eregon on Slack, he withdrew his objection to this change.

In general, this change:

- Aligns `Fiber.new` with `rb_fiber_new` so that they both produce non-blocking fibers.
- Doesn't affect existing code, as there is no obvious usage of `rb_fiber_new` by GitHub code search (small/zero blast radius).
- Only impacts usage within the fiber scheduler, i.e. no effect outside of fiber scheduler beside the predicate value itself.
- Even within the fiber scheduler, it is transparent to user code.

----------------------------------------
Bug #20413: Enumerator can block fiber scheduler.
https://bugs.ruby-lang.org/issues/20413#change-107841

* Author: ioquatix (Samuel Williams)
* Status: Closed
* Assignee: ioquatix (Samuel Williams)
* Backport: 3.0: UNKNOWN, 3.1: REQUIRED, 3.2: REQUIRED, 3.3: REQUIRED
----------------------------------------
Using `Enumerator` in the event loop can cause problems as the fiber created by `rb_fiber_new` is blocking by default. It should be non-blocking.

```ruby
#!/usr/bin/env ruby

require 'async'

Async do
  Async do
    while true
      puts "Hello"
      sleep 1
    end
  end
  
  enumerator = Enumerator.new do |yielder|
    while true
      yielder << "World"
      sleep 1
    end
  end
  
  while true
    puts enumerator.next
  end
end
```

Before this PR, the output is:

```
> ./test.rb
Hello
World
World
World
World
World
World
...
```

After this PR, the output is:

```
> ./test.rb
Hello
World
Hello
World
Hello
World
Hello
World
Hello
World
...
```

The reason why this happens, is because the enumerator `sleep` never yields to the event loop.

See <https://github.com/ruby/ruby/pull/10481> for the fix.



-- 
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/

      parent reply	other threads:[~2024-04-07 12:48 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-04-07 10:47 [ruby-core:117453] [Ruby master Bug#20413] Enumerator can block fiber scheduler ioquatix (Samuel Williams) via ruby-core
2024-04-07 11:16 ` [ruby-core:117454] " Eregon (Benoit Daloze) via ruby-core
2024-04-07 11:57 ` [ruby-core:117455] " ioquatix (Samuel Williams) via ruby-core
2024-04-07 12:48 ` ioquatix (Samuel Williams) via ruby-core [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-list from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

  List information: https://www.ruby-lang.org/en/community/mailing-lists/

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=redmine.journal-107841.20240407124806.3344@ruby-lang.org \
    --to=ruby-core@ruby-lang.org \
    --cc=noreply@ruby-lang.org \
    --cc=ruby-core@ml.ruby-lang.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).