ruby-core@ruby-lang.org archive (unofficial mirror)
 help / color / mirror / Atom feed
From: "Eregon (Benoit Daloze)" <noreply@ruby-lang.org>
To: ruby-core@neon.ruby-lang.org
Subject: [ruby-core:110358] [Ruby master Bug#19062] Introduce `Fiber#locals` for shared inheritable state.
Date: Mon, 17 Oct 2022 10:29:18 +0000 (UTC)	[thread overview]
Message-ID: <redmine.journal-99651.20221017102918.3344@ruby-lang.org> (raw)
In-Reply-To: redmine.issue-19062.20221016103104.3344@ruby-lang.org

Issue #19062 has been updated by Eregon (Benoit Daloze).


For sure we need the `.dup` on the internal hash used for storage when inheriting it to a new Fiber.
It would be a semantic bug if adding a new local in a child fiber affects a parent fiber as @byroot said.
Or we'd need some kind of persistent map to avoid the actual copy, but semantically it would behave the same as if copied.

The question is should we deep copy / dup the values too?
(for the keys probably we can require Symbol/String and freeze if a String like regular Hash do).

I'm not sure what's the advantage of this proposal vs #19058.
In both cases it's O(locals) per new fiber if using a hash.
The persistent map would make that O(1), but at the cost of slower lookup and updates.

> we can propagate locals across threads

Isn't that always a bad idea? Sharing e.g. a DB connection between threads would be unsafe.
If we allow inheriting across threads then at least it should be very explicit.

----------------------------------------
Bug #19062: Introduce `Fiber#locals` for shared inheritable state.
https://bugs.ruby-lang.org/issues/19062#change-99651

* Author: ioquatix (Samuel Williams)
* Status: Open
* Priority: Normal
* Assignee: ioquatix (Samuel Williams)
* Backport: 2.7: UNKNOWN, 3.0: UNKNOWN, 3.1: UNKNOWN
----------------------------------------
After exploring <https://bugs.ruby-lang.org/issues/19058>, I felt uncomfortable about the performance of copying lots of inheritable attributes. Please review that issue for the background and summary of the problem.

## Proposal

Introduce `Fiber#locals` which is a hash table of local attributes which are inherited by child fibers.

```ruby
Fiber.current.locals[:x] = 10

Fiber.new do
  pp Fiber.current.locals[:x] # => 10
end
```

It's possible to reset `Fiber.current.locals`, e.g.

```ruby
def accept_connection(peer)
  Fiber.new(locals: nil) do # This causes a new hash table to be allocated.
    # Generate a new request id for all fibers nested in this one:
    Fiber[:request_id] = SecureRandom.hex(32)
    @app.call(env)
  end.resume
end
```

A high level overview of the proposed changes:

```ruby
class Fiber
  def initialize(..., locals: Fiber.current.locals)
    @locals = locals || Hash.new
  end

  attr_accessor :locals

  def self.[] key
    self.current.locals[key]
  end

  def self.[]= key, value
    self.current.locals[key] = value
  end
end
```

See the pull request <https://github.com/ruby/ruby/pull/6566> for the full proposed implementation.

## Expected Usage

Currently, a lot of libraries use `Thread.current[:x]` which is unexpectedly "fiber local". A common bug shows up when lazy enumerators are used, because it may create an internal fiber. Because `locals` are inherited, code which uses `Fiber[:x]` will not suffer from this problem.

Any program that uses true thread locals for per-request state, can adopt the proposed `Fiber#locals` and get similar behaviour, without breaking on per-fiber servers like Falcon, because Falcon can "reset" `Fiber.current.locals` for each request fiber, while servers like Puma won't have to do that and will retain thread-local behaviour.

Libraries like ActiveRecord can adopt `Fiber#locals` to avoid the need for users to opt into different "IsolatedExecutionState" models, since it can be transparently handled by the web server (see <https://github.com/rails/rails/pull/43596> for more details).

We hope by introducing `Fiber#locals`, we can avoid all the confusion and bugs of the past designs.



-- 
https://bugs.ruby-lang.org/

  parent reply	other threads:[~2022-10-17 10:29 UTC|newest]

Thread overview: 33+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-10-16 10:31 [ruby-core:110320] [Ruby master Bug#19062] Introduce `Fiber#locals` for shared inheritable state ioquatix (Samuel Williams)
2022-10-16 22:08 ` [ruby-core:110323] " ioquatix (Samuel Williams)
2022-10-17  7:01 ` [ruby-core:110346] " nobu (Nobuyoshi Nakada)
2022-10-17  7:40 ` [ruby-core:110349] " byroot (Jean Boussier)
2022-10-17  7:50 ` [ruby-core:110350] " ioquatix (Samuel Williams)
2022-10-17  8:23 ` [ruby-core:110352] " byroot (Jean Boussier)
2022-10-17  8:27 ` [ruby-core:110353] " ioquatix (Samuel Williams)
2022-10-17  8:42 ` [ruby-core:110354] " byroot (Jean Boussier)
2022-10-17  9:18 ` [ruby-core:110355] " ioquatix (Samuel Williams)
2022-10-17  9:30 ` [ruby-core:110356] " ioquatix (Samuel Williams)
2022-10-17  9:44 ` [ruby-core:110357] " ioquatix (Samuel Williams)
2022-10-17 10:29 ` Eregon (Benoit Daloze) [this message]
2022-10-17 10:48 ` [ruby-core:110359] " ioquatix (Samuel Williams)
2022-10-17 10:55 ` [ruby-core:110360] " byroot (Jean Boussier)
2022-10-17 11:12 ` [ruby-core:110361] " ioquatix (Samuel Williams)
2022-10-17 11:48 ` [ruby-core:110362] " byroot (Jean Boussier)
2022-10-17 11:49 ` [ruby-core:110363] " Eregon (Benoit Daloze)
2022-10-17 19:42 ` [ruby-core:110374] " ioquatix (Samuel Williams)
2022-10-17 19:57 ` [ruby-core:110375] " ioquatix (Samuel Williams)
2022-10-17 20:05 ` [ruby-core:110376] " byroot (Jean Boussier)
2022-10-17 21:58 ` [ruby-core:110377] " ioquatix (Samuel Williams)
2022-10-18  8:14 ` [ruby-core:110382] " byroot (Jean Boussier)
2022-10-18 10:15 ` [ruby-core:110389] " ioquatix (Samuel Williams)
2022-10-18 10:37 ` [ruby-core:110391] " byroot (Jean Boussier)
2022-10-18 11:05 ` [ruby-core:110394] " Eregon (Benoit Daloze)
2022-10-18 11:20 ` [ruby-core:110397] " ioquatix (Samuel Williams)
2022-10-18 11:26 ` [ruby-core:110398] [Ruby master Feature#19062] " ioquatix (Samuel Williams)
2022-10-18 11:44 ` [ruby-core:110399] " byroot (Jean Boussier)
2022-10-18 11:55 ` [ruby-core:110400] " ioquatix (Samuel Williams)
2022-10-18 12:06 ` [ruby-core:110405] " byroot (Jean Boussier)
2022-10-18 12:32 ` [ruby-core:110406] " Eregon (Benoit Daloze)
2022-10-20  6:48 ` [ruby-core:110433] " matz (Yukihiro Matsumoto)
2022-10-20 12:09 ` [ruby-core:110445] " ioquatix (Samuel Williams)

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-99651.20221017102918.3344@ruby-lang.org \
    --to=ruby-core@ruby-lang.org \
    --cc=ruby-core@neon.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).