General

Profile

jhawthorn (John Hawthorn)

  • Login: jhawthorn
  • Registered on: 12/22/2016
  • Last sign in: 11/17/2026

Issues

open closed Total
Assigned issues 17 91 108
Reported issues 7 39 46

Projects

Project Roles Registered on
Ruby Committer 11/25/2021

Activity

Today

04:27 AM Ruby Bug #22531 (Closed): String#% with %s format string runs arbitrary code that can invalidate the argument array
Looks like Peter fixed this in commit:git|edce07a8c93895be8eeb2d27400aa138cc8a3cf9, which was backported to 3.3, 3.4 and 4.0 in yesterday's releases.
On master and 4.0.10 the example now raises instead of crashing:
/tmp/r.rb:5:...
jhawthorn (John Hawthorn)

11/19/2026

08:36 PM Ruby Revision a2d94e31 (git): Fix thread exit corrupting the next owner's mutex list
m1 = Thread::Mutex.new
m2 = Thread::Mutex.new
Thread.new { m1.lock; m2.lock }.join
Thread.new { m2.lock }.join
# [BUG] invalid keeping_mutexes: Attempt to unlock a mutex which is not locked
rb_threadptr_unlock_all_lockin...
jhawthorn (John Hawthorn)
07:55 PM Ruby Bug #22598: A thread that ends holding two mutexes corrupts the next owner's mutex list
3.4: https://github.com/ruby/ruby/pull/20029
4.0: https://github.com/ruby/ruby/pull/20028
jhawthorn (John Hawthorn)
07:53 PM Ruby Bug #22598 (Closed): A thread that ends holding two mutexes corrupts the next owner's mutex list
Applied in changeset commit:git|e1d5a61acb9109f023398f26c68e447f69e310d3.
----------
Fix thread exit corrupting the next owner's mutex list
m1 = Thread::Mutex.new
m2 = Thread::Mutex.new
Thread.new { m1.lock; m2.lock }.join
...
jhawthorn (John Hawthorn)
07:53 PM Ruby Revision e1d5a61a (git): Fix thread exit corrupting the next owner's mutex list
m1 = Thread::Mutex.new
m2 = Thread::Mutex.new
Thread.new { m1.lock; m2.lock }.join
Thread.new { m2.lock }.join
# [BUG] invalid keeping_mutexes: Attempt to unlock a mutex which is not locked
rb_threadptr_unlock_all_lockin...
jhawthorn (John Hawthorn)

11/18/2026

05:19 PM Ruby Revision 801cf30f (git): Fix off-by-one in rb_internal_thread_specific_key_create
Each thread's specific_storage only has
RB_INTERNAL_THREAD_SPECIFIC_KEY_MAX (8) slots, but the check used >
instead of >=, so the ninth call returned key 8 instead of raising, and
using that key reads and writes past the end of the array.
jhawthorn (John Hawthorn)
08:52 AM Ruby Revision c46b89e5 (git): Don't modify fstrings in rb_str_tmp_frozen_no_embed_acquire
[Bug #21671]
(cherry picked from commit 3b588dab91f13b32b54cb1b9dedb9691e732c981)
jhawthorn (John Hawthorn)
04:31 AM Ruby Bug #22598 (Closed): A thread that ends holding two mutexes corrupts the next owner's mutex list
Reproduces on 4.0.9 and master:
```ruby
m1 = Thread::Mutex.new
m2 = Thread::Mutex.new
Thread.new { m1.lock; m2.lock }.join
Thread.new { m2.synchronize { } }.join
# [BUG] invalid keeping_mutexes: Attempt to unlock a mutex which is...
jhawthorn (John Hawthorn)

11/17/2026

07:26 PM Ruby Bug #22595 (Open): ObjectSpace::WeakMap#each can yield the same pair twice if compaction happens during iteration
Reproduces on 4.0.9 and master:
```ruby
map = ObjectSpace::WeakMap.new
key = Object.new
map[key] = 1
100.times { |i| map[i] = i }
count = 0
map.each do
GC.verify_compaction_references(expand_heap: true, toward: :empty) if cou...
jhawthorn (John Hawthorn)

11/16/2026

05:59 PM Ruby Revision e21fca12 (git): Fix concurrent_set delete missing keys past an in-progress insert
rb_concurrent_set_delete_by_identity stopped probing at the first entry
with an EMPTY key. Since 81fbdff8fdf2ae7afb2fa19319ff7d40379521fe
inserts reserve an entry by CASing its hash before its key, so an entry
with a hash and an EMPTY ke...
jhawthorn (John Hawthorn)

Also available in: Atom