Bug #20250: Crash with "Object ID seen, but not in mapping table: proc" error

• Status: Closed
• Assignee: -
• Target version: -
• ruby -v: ruby 3.3.0 (2023-12-25 revision 5124f9ac75) [arm64-darwin23]
• Backport: 3.0: WONTFIX, 3.1: REQUIRED, 3.2: REQUIRED, 3.3: DONE

[zetter (Chris Zetter)]

Hello, I experienced a crash which I was able to reliably reproduce with the following:

require 'bundler/inline'

gemfile(true) do
  source 'https://rubygems.org'
  gem 'activesupport', '7.1.3'
end

require 'active_support'

logger = ActiveSupport::Logger.new('log/log.log', 1, 100 * 1024 * 1024)

logger.formatter = proc {|_, _, _, message| "#{message}\n" }

logger = ActiveSupport::TaggedLogging.new(logger)

logger.tagged("TAG").info "hello"
logger.tagged("TAG").info "hello" # usually crashes here
GC.start # sometimes need to trigger crash with GC

It looks like this is caused by the interaction with the formatter proc and the tagged logging features of activesupport. Let me know if a more minimal example would be useful.

I can reproduce this with:

• ruby 3.3.0 on arm64-darwin23
• ruby 3.3.0 on x86_64-linux
• latest ruby HEAD (5e12b75716) on arm64-darwin23

I cannot reproduce this on ruby 3.2.2.

Thanks for any help

[byroot (Jean Boussier)]

I had a quick look and this is very interesting. As far as I can tell cached_object_id for this Proc, so the most likely explanation is that some other part of the codebase is messing with the object flags.

I'll try to dig deeper.

[byroot (Jean Boussier)]

Alright, it's a bug in #clone, I managed to reduce it to:

proc = Proc.new { }
proc.object_id
proc.clone
GC.start

Shouldn't be too hard to figure out. IIRC Proc#clone was modified recently, the flag that indicate the instance has an object_id need to be cleared on the cloned object.

[byroot (Jean Boussier)]

• Backport changed from 3.0: UNKNOWN, 3.1: UNKNOWN, 3.2: UNKNOWN, 3.3: UNKNOWN to 3.0: WONTFIX, 3.1: REQUIRED, 3.2: REQUIRED, 3.3: REQUIRED

Patch: https://github.com/ruby/ruby/pull/9903

Also this bug affect all Ruby versions since the new object_id implementation back in 2.7, I got all of them to crash on it.

[byroot (Jean Boussier)]

• Status changed from Open to Closed

Applied in changeset git|d19d683a354530a27b4cbb049223f8dc70c75849.

---

rb_obj_setup: do not copy RUBY_FL_SEEN_OBJ_ID

[Bug #20250]

We're seting up a new instance, so it never had an associated
object_id.

[byroot (Jean Boussier)]

• Related to Bug #20253: `Proc.dup` and `Proc#clone` don't preserve finalizers added

[naruse (Yui NARUSE)]

• Backport changed from 3.0: WONTFIX, 3.1: REQUIRED, 3.2: REQUIRED, 3.3: REQUIRED to 3.0: WONTFIX, 3.1: REQUIRED, 3.2: REQUIRED, 3.3: DONE

ruby_3_3 a63e979853783601a60228b45741f8b3776e5507 merged revision(s) d19d683a354530a27b4cbb049223f8dc70c75849,de1a586ecc2ee7f465f0c0a69291054136a3a819.