timers: do not retain a reference to the async store after firing by mcollina · Pull Request #53443 · nodejs/node

mcollina · 2024-06-13T17:11:48Z

After firing timers, we can clean them up by iterating over all active stores and setting the relevant symbols to undefined.

This is still a draft because we will need to extend it to immediates and intervals, too.

Fixes#53408

mcollina · 2024-06-13T17:13:47Z

I've opened a draft PR to discuss a possible solution to #53408.

cc @nodejs/diagnostics @nodejs/timers let me know what you think.

Flarna · 2024-07-02T20:41:45Z

@@ -429,6 +432,20 @@ function setPosition(node, pos) {
 node.priorityQueuePosition = pos;
 }

+function removeAllStores (timer) {


What about moving this function into async_hooks to avoid the need to expose getActiveStores() there?
An alternative would be to use a single object on the resource which holds all stores. This would reduce the number of properties added on resource objects - and avoid the need of a for loop to clear it.

A helper that receives a resource and clears all the store data off it would be better than exposing the stores list as we do now. As it is presently, one could get and retain access to a store in a way unintentional from the perspective of the store owner, which is a bit dangerous from a memory perspective.

mcollina · 2026-05-31T19:42:12Z

I've updated this PR to address the review feedback:

Instead of exposing getActiveStores(), added a clearStores(resource) helper function in async_hooks that takes a resource and clears all store references from it. This avoids exposing the internal stores list.
Added cleanup for _timerArgs in addition to _onTimeout and async stores.
Extended the cleanup to immediates (setImmediate) and intervals (setInterval) as well.
Updated the test to cover all three cases.

cc @Flarna @Qard

codecov · 2026-05-31T20:38:20Z

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 90.23%. Comparing base (1d87a24) to head (ff9bebb).
⚠️ Report is 94 commits behind head on main.

Additional details and impacted files

@@ Coverage Diff @@## main #53443 +/- ##
==========================================
- Coverage 90.25% 90.23% -0.03% 
==========================================
Files 741 741 Lines 240990 241082 +92 Branches 45402 45425 +23 ==========================================
+ Hits 217502 217535 +33 - Misses 15066 15116 +50 - Partials 8422 8431 +9

Files with missing lines	Coverage Δ
lib/internal/async_hooks.js	`99.36% <100.00%> (+<0.01%)`	⬆️
lib/internal/async_local_storage/async_hooks.js	`98.21% <100.00%> (+0.16%)`	⬆️
lib/internal/timers.js	`99.75% <100.00%> (+0.02%)`	⬆️
lib/timers.js	`100.00% <100.00%> (ø)`

... and 39 files with indirect coverage changes

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

After firing timers, we can clean them up by iterating over all active stores and setting the relevant symbols to undefined. This also covers immediates and intervals. Refs: nodejs#53408 PR-URL: nodejs#53443Fixes: nodejs#53408 Signed-off-by: Matteo Collina <hello@matteocollina.com>

Flarna · 2026-06-18T21:07:29Z


 const async_context_frame = Symbol('kAsyncContextFrame');

+function removeStoresFromResource(resource) {


I think this should be extended to somthing like this to cover also the async context frame variant of ALS:
functionremoveStoresFromResource(resource){if(AsyncContextFrame.enabled){if(resource[async_context_frame]!==undefined){resource[async_context_frame]=undefined;}}else{if(resource[async_local_storage_context_symbol]!==undefined){resource[async_local_storage_context_symbol]=undefined;}}}

Flarna · 2026-06-18T21:08:41Z

@@ -0,0 +1,109 @@
+// Flags: --expose-internals --no-async-context-frame


I think this should be tested with and without async-context-frame
And a test via GC would be likely helpful, e.g. somthing like this:
{constasyncLocalStorage=newAsyncLocalStorage();letstore={};lettimeout;constweakStore=newWeakRef(store);asyncLocalStorage.run(store,common.mustCall(()=>{timeout=setTimeout(common.mustNotCall(),common.platformTimeout(1000));assert.strictEqual(weakStore.deref(),store);clearTimeout(timeout);}));store=null;setImmediate(()=>{global.gc();assert.strictEqual(weakStore.deref(),undefined);});}

Flarna · 2026-06-23T20:24:20Z

@mcollina seems you forced push some old commit or a mixture of old and new. At least it looks to me like parts of the recent changes got lost.

Any chance to avoid forced push to ease reviews?

mcollina · 2026-06-24T09:52:43Z

I had to force-push for rebase.

mcollina · 2026-07-08T08:01:44Z

@Flarna PTAL

After firing timers, we can clean them up by iterating over all active stores and setting the relevant symbols to undefined. Fixesnodejs#53408 Signed-off-by: Matteo Collina <hello@matteocollina.com>

Flarna · 2026-07-08T13:24:41Z

I still don't understand why rebase+squash is done again and again in this PR.
I get that a rebase is needed every now and then, in special for that long running PRs.
But I see no need to squash fixup commits at the same time.

mcollina · 2026-07-09T09:19:09Z

I still don't understand why rebase+squash is done again and again in this PR.
I get that a rebase is needed every now and then, in special for that long running PRs.
But I see no need to squash fixup commits at the same time.

Sorry, I got a habit of squashing the review fixes when rebasing, mostly because other collaborators told me they preferred that way.

nodejs-github-bot · 2026-07-19T09:01:16Z

CI: https://ci.nodejs.org/job/node-test-pull-request/74887/

nodejs-github-bot · 2026-07-19T18:30:08Z

Landed in d25cde4

After firing timers, we can clean them up by iterating over all active stores and setting the relevant symbols to undefined. Fixes#53408 Signed-off-by: Matteo Collina <hello@matteocollina.com> PR-URL: #53443 Reviewed-By: Stephen Belanger <admin@stephenbelanger.com> Reviewed-By: Paolo Insogna <paolo@cowtech.it> Reviewed-By: James M Snell <jasnell@gmail.com> Reviewed-By: Gerhard Stöbich <deb2001-github@yahoo.de>

mcollina · 2024-06-13T17:11:48Z

After firing timers, we can clean them up by iterating over all active stores and setting the relevant symbols to undefined.

This is still a draft because we will need to extend it to immediates and intervals, too.

Fixes#53408

mcollina · 2024-06-13T17:13:47Z

I've opened a draft PR to discuss a possible solution to #53408.

cc @nodejs/diagnostics @nodejs/timers let me know what you think.

Flarna · 2024-07-02T20:41:45Z

@@ -429,6 +432,20 @@ function setPosition(node, pos) {
 node.priorityQueuePosition = pos;
 }

+function removeAllStores (timer) {


What about moving this function into async_hooks to avoid the need to expose getActiveStores() there?
An alternative would be to use a single object on the resource which holds all stores. This would reduce the number of properties added on resource objects - and avoid the need of a for loop to clear it.

A helper that receives a resource and clears all the store data off it would be better than exposing the stores list as we do now. As it is presently, one could get and retain access to a store in a way unintentional from the perspective of the store owner, which is a bit dangerous from a memory perspective.

mcollina · 2026-05-31T19:42:12Z

I've updated this PR to address the review feedback:

Instead of exposing getActiveStores(), added a clearStores(resource) helper function in async_hooks that takes a resource and clears all store references from it. This avoids exposing the internal stores list.
Added cleanup for _timerArgs in addition to _onTimeout and async stores.
Extended the cleanup to immediates (setImmediate) and intervals (setInterval) as well.
Updated the test to cover all three cases.

cc @Flarna @Qard

codecov · 2026-05-31T20:38:20Z

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 90.23%. Comparing base (1d87a24) to head (ff9bebb).
⚠️ Report is 94 commits behind head on main.

Additional details and impacted files

@@ Coverage Diff @@## main #53443 +/- ##
==========================================
- Coverage 90.25% 90.23% -0.03% 
==========================================
Files 741 741 Lines 240990 241082 +92 Branches 45402 45425 +23 ==========================================
+ Hits 217502 217535 +33 - Misses 15066 15116 +50 - Partials 8422 8431 +9

Files with missing lines	Coverage Δ
lib/internal/async_hooks.js	`99.36% <100.00%> (+<0.01%)`	⬆️
lib/internal/async_local_storage/async_hooks.js	`98.21% <100.00%> (+0.16%)`	⬆️
lib/internal/timers.js	`99.75% <100.00%> (+0.02%)`	⬆️
lib/timers.js	`100.00% <100.00%> (ø)`

... and 39 files with indirect coverage changes

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

After firing timers, we can clean them up by iterating over all active stores and setting the relevant symbols to undefined. This also covers immediates and intervals. Refs: nodejs#53408 PR-URL: nodejs#53443Fixes: nodejs#53408 Signed-off-by: Matteo Collina <hello@matteocollina.com>

Flarna · 2026-06-18T21:07:29Z


 const async_context_frame = Symbol('kAsyncContextFrame');

+function removeStoresFromResource(resource) {


I think this should be extended to somthing like this to cover also the async context frame variant of ALS:
functionremoveStoresFromResource(resource){if(AsyncContextFrame.enabled){if(resource[async_context_frame]!==undefined){resource[async_context_frame]=undefined;}}else{if(resource[async_local_storage_context_symbol]!==undefined){resource[async_local_storage_context_symbol]=undefined;}}}

Flarna · 2026-06-18T21:08:41Z

@@ -0,0 +1,109 @@
+// Flags: --expose-internals --no-async-context-frame


I think this should be tested with and without async-context-frame
And a test via GC would be likely helpful, e.g. somthing like this:
{constasyncLocalStorage=newAsyncLocalStorage();letstore={};lettimeout;constweakStore=newWeakRef(store);asyncLocalStorage.run(store,common.mustCall(()=>{timeout=setTimeout(common.mustNotCall(),common.platformTimeout(1000));assert.strictEqual(weakStore.deref(),store);clearTimeout(timeout);}));store=null;setImmediate(()=>{global.gc();assert.strictEqual(weakStore.deref(),undefined);});}

Flarna · 2026-06-23T20:24:20Z

@mcollina seems you forced push some old commit or a mixture of old and new. At least it looks to me like parts of the recent changes got lost.

Any chance to avoid forced push to ease reviews?

mcollina · 2026-06-24T09:52:43Z

I had to force-push for rebase.

mcollina · 2026-07-08T08:01:44Z

@Flarna PTAL

After firing timers, we can clean them up by iterating over all active stores and setting the relevant symbols to undefined. Fixesnodejs#53408 Signed-off-by: Matteo Collina <hello@matteocollina.com>

Flarna · 2026-07-08T13:24:41Z

I still don't understand why rebase+squash is done again and again in this PR.
I get that a rebase is needed every now and then, in special for that long running PRs.
But I see no need to squash fixup commits at the same time.

mcollina · 2026-07-09T09:19:09Z

I still don't understand why rebase+squash is done again and again in this PR.
I get that a rebase is needed every now and then, in special for that long running PRs.
But I see no need to squash fixup commits at the same time.

Sorry, I got a habit of squashing the review fixes when rebasing, mostly because other collaborators told me they preferred that way.

nodejs-github-bot · 2026-07-19T09:01:16Z

CI: https://ci.nodejs.org/job/node-test-pull-request/74887/

nodejs-github-bot · 2026-07-19T18:30:08Z

Landed in d25cde4

After firing timers, we can clean them up by iterating over all active stores and setting the relevant symbols to undefined. Fixes#53408 Signed-off-by: Matteo Collina <hello@matteocollina.com> PR-URL: #53443 Reviewed-By: Stephen Belanger <admin@stephenbelanger.com> Reviewed-By: Paolo Insogna <paolo@cowtech.it> Reviewed-By: James M Snell <jasnell@gmail.com> Reviewed-By: Gerhard Stöbich <deb2001-github@yahoo.de>

mcollina · 2024-06-13T17:11:48Z

After firing timers, we can clean them up by iterating over all active stores and setting the relevant symbols to undefined.

This is still a draft because we will need to extend it to immediates and intervals, too.

Fixes#53408

mcollina · 2024-06-13T17:13:47Z

I've opened a draft PR to discuss a possible solution to #53408.

cc @nodejs/diagnostics @nodejs/timers let me know what you think.

Flarna · 2024-07-02T20:41:45Z

@@ -429,6 +432,20 @@ function setPosition(node, pos) {
 node.priorityQueuePosition = pos;
 }

+function removeAllStores (timer) {


What about moving this function into async_hooks to avoid the need to expose getActiveStores() there?
An alternative would be to use a single object on the resource which holds all stores. This would reduce the number of properties added on resource objects - and avoid the need of a for loop to clear it.

A helper that receives a resource and clears all the store data off it would be better than exposing the stores list as we do now. As it is presently, one could get and retain access to a store in a way unintentional from the perspective of the store owner, which is a bit dangerous from a memory perspective.

mcollina · 2026-05-31T19:42:12Z

I've updated this PR to address the review feedback:

Instead of exposing getActiveStores(), added a clearStores(resource) helper function in async_hooks that takes a resource and clears all store references from it. This avoids exposing the internal stores list.
Added cleanup for _timerArgs in addition to _onTimeout and async stores.
Extended the cleanup to immediates (setImmediate) and intervals (setInterval) as well.
Updated the test to cover all three cases.

cc @Flarna @Qard

codecov · 2026-05-31T20:38:20Z

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 90.23%. Comparing base (1d87a24) to head (ff9bebb).
⚠️ Report is 94 commits behind head on main.

Additional details and impacted files

@@ Coverage Diff @@## main #53443 +/- ##
==========================================
- Coverage 90.25% 90.23% -0.03% 
==========================================
Files 741 741 Lines 240990 241082 +92 Branches 45402 45425 +23 ==========================================
+ Hits 217502 217535 +33 - Misses 15066 15116 +50 - Partials 8422 8431 +9

Files with missing lines	Coverage Δ
lib/internal/async_hooks.js	`99.36% <100.00%> (+<0.01%)`	⬆️
lib/internal/async_local_storage/async_hooks.js	`98.21% <100.00%> (+0.16%)`	⬆️
lib/internal/timers.js	`99.75% <100.00%> (+0.02%)`	⬆️
lib/timers.js	`100.00% <100.00%> (ø)`

... and 39 files with indirect coverage changes

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

After firing timers, we can clean them up by iterating over all active stores and setting the relevant symbols to undefined. This also covers immediates and intervals. Refs: nodejs#53408 PR-URL: nodejs#53443Fixes: nodejs#53408 Signed-off-by: Matteo Collina <hello@matteocollina.com>

Flarna · 2026-06-18T21:07:29Z


 const async_context_frame = Symbol('kAsyncContextFrame');

+function removeStoresFromResource(resource) {


I think this should be extended to somthing like this to cover also the async context frame variant of ALS:
functionremoveStoresFromResource(resource){if(AsyncContextFrame.enabled){if(resource[async_context_frame]!==undefined){resource[async_context_frame]=undefined;}}else{if(resource[async_local_storage_context_symbol]!==undefined){resource[async_local_storage_context_symbol]=undefined;}}}

Flarna · 2026-06-18T21:08:41Z

@@ -0,0 +1,109 @@
+// Flags: --expose-internals --no-async-context-frame


I think this should be tested with and without async-context-frame
And a test via GC would be likely helpful, e.g. somthing like this:
{constasyncLocalStorage=newAsyncLocalStorage();letstore={};lettimeout;constweakStore=newWeakRef(store);asyncLocalStorage.run(store,common.mustCall(()=>{timeout=setTimeout(common.mustNotCall(),common.platformTimeout(1000));assert.strictEqual(weakStore.deref(),store);clearTimeout(timeout);}));store=null;setImmediate(()=>{global.gc();assert.strictEqual(weakStore.deref(),undefined);});}

Flarna · 2026-06-23T20:24:20Z

@mcollina seems you forced push some old commit or a mixture of old and new. At least it looks to me like parts of the recent changes got lost.

Any chance to avoid forced push to ease reviews?

mcollina · 2026-06-24T09:52:43Z

I had to force-push for rebase.

mcollina · 2026-07-08T08:01:44Z

@Flarna PTAL

After firing timers, we can clean them up by iterating over all active stores and setting the relevant symbols to undefined. Fixesnodejs#53408 Signed-off-by: Matteo Collina <hello@matteocollina.com>

Flarna · 2026-07-08T13:24:41Z

I still don't understand why rebase+squash is done again and again in this PR.
I get that a rebase is needed every now and then, in special for that long running PRs.
But I see no need to squash fixup commits at the same time.

mcollina · 2026-07-09T09:19:09Z

I still don't understand why rebase+squash is done again and again in this PR.
I get that a rebase is needed every now and then, in special for that long running PRs.
But I see no need to squash fixup commits at the same time.

Sorry, I got a habit of squashing the review fixes when rebasing, mostly because other collaborators told me they preferred that way.

nodejs-github-bot · 2026-07-19T09:01:16Z

CI: https://ci.nodejs.org/job/node-test-pull-request/74887/

nodejs-github-bot · 2026-07-19T18:30:08Z

Landed in d25cde4

After firing timers, we can clean them up by iterating over all active stores and setting the relevant symbols to undefined. Fixes#53408 Signed-off-by: Matteo Collina <hello@matteocollina.com> PR-URL: #53443 Reviewed-By: Stephen Belanger <admin@stephenbelanger.com> Reviewed-By: Paolo Insogna <paolo@cowtech.it> Reviewed-By: James M Snell <jasnell@gmail.com> Reviewed-By: Gerhard Stöbich <deb2001-github@yahoo.de>

mcollina · 2024-06-13T17:11:48Z

After firing timers, we can clean them up by iterating over all active stores and setting the relevant symbols to undefined.

This is still a draft because we will need to extend it to immediates and intervals, too.

Fixes#53408

mcollina · 2024-06-13T17:13:47Z

I've opened a draft PR to discuss a possible solution to #53408.

cc @nodejs/diagnostics @nodejs/timers let me know what you think.

Flarna · 2024-07-02T20:41:45Z

@@ -429,6 +432,20 @@ function setPosition(node, pos) {
 node.priorityQueuePosition = pos;
 }

+function removeAllStores (timer) {


What about moving this function into async_hooks to avoid the need to expose getActiveStores() there?
An alternative would be to use a single object on the resource which holds all stores. This would reduce the number of properties added on resource objects - and avoid the need of a for loop to clear it.

A helper that receives a resource and clears all the store data off it would be better than exposing the stores list as we do now. As it is presently, one could get and retain access to a store in a way unintentional from the perspective of the store owner, which is a bit dangerous from a memory perspective.

mcollina · 2026-05-31T19:42:12Z

I've updated this PR to address the review feedback:

Instead of exposing getActiveStores(), added a clearStores(resource) helper function in async_hooks that takes a resource and clears all store references from it. This avoids exposing the internal stores list.
Added cleanup for _timerArgs in addition to _onTimeout and async stores.
Extended the cleanup to immediates (setImmediate) and intervals (setInterval) as well.
Updated the test to cover all three cases.

cc @Flarna @Qard

codecov · 2026-05-31T20:38:20Z

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 90.23%. Comparing base (1d87a24) to head (ff9bebb).
⚠️ Report is 94 commits behind head on main.

Additional details and impacted files

@@ Coverage Diff @@## main #53443 +/- ##
==========================================
- Coverage 90.25% 90.23% -0.03% 
==========================================
Files 741 741 Lines 240990 241082 +92 Branches 45402 45425 +23 ==========================================
+ Hits 217502 217535 +33 - Misses 15066 15116 +50 - Partials 8422 8431 +9

Files with missing lines	Coverage Δ
lib/internal/async_hooks.js	`99.36% <100.00%> (+<0.01%)`	⬆️
lib/internal/async_local_storage/async_hooks.js	`98.21% <100.00%> (+0.16%)`	⬆️
lib/internal/timers.js	`99.75% <100.00%> (+0.02%)`	⬆️
lib/timers.js	`100.00% <100.00%> (ø)`

... and 39 files with indirect coverage changes

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

After firing timers, we can clean them up by iterating over all active stores and setting the relevant symbols to undefined. This also covers immediates and intervals. Refs: nodejs#53408 PR-URL: nodejs#53443Fixes: nodejs#53408 Signed-off-by: Matteo Collina <hello@matteocollina.com>

Flarna · 2026-06-18T21:07:29Z


 const async_context_frame = Symbol('kAsyncContextFrame');

+function removeStoresFromResource(resource) {


I think this should be extended to somthing like this to cover also the async context frame variant of ALS:
functionremoveStoresFromResource(resource){if(AsyncContextFrame.enabled){if(resource[async_context_frame]!==undefined){resource[async_context_frame]=undefined;}}else{if(resource[async_local_storage_context_symbol]!==undefined){resource[async_local_storage_context_symbol]=undefined;}}}

Flarna · 2026-06-18T21:08:41Z

@@ -0,0 +1,109 @@
+// Flags: --expose-internals --no-async-context-frame


I think this should be tested with and without async-context-frame
And a test via GC would be likely helpful, e.g. somthing like this:
{constasyncLocalStorage=newAsyncLocalStorage();letstore={};lettimeout;constweakStore=newWeakRef(store);asyncLocalStorage.run(store,common.mustCall(()=>{timeout=setTimeout(common.mustNotCall(),common.platformTimeout(1000));assert.strictEqual(weakStore.deref(),store);clearTimeout(timeout);}));store=null;setImmediate(()=>{global.gc();assert.strictEqual(weakStore.deref(),undefined);});}

Flarna · 2026-06-23T20:24:20Z

@mcollina seems you forced push some old commit or a mixture of old and new. At least it looks to me like parts of the recent changes got lost.

Any chance to avoid forced push to ease reviews?

mcollina · 2026-06-24T09:52:43Z

I had to force-push for rebase.

mcollina · 2026-07-08T08:01:44Z

@Flarna PTAL

After firing timers, we can clean them up by iterating over all active stores and setting the relevant symbols to undefined. Fixesnodejs#53408 Signed-off-by: Matteo Collina <hello@matteocollina.com>

Flarna · 2026-07-08T13:24:41Z

I still don't understand why rebase+squash is done again and again in this PR.
I get that a rebase is needed every now and then, in special for that long running PRs.
But I see no need to squash fixup commits at the same time.

mcollina · 2026-07-09T09:19:09Z

I still don't understand why rebase+squash is done again and again in this PR.
I get that a rebase is needed every now and then, in special for that long running PRs.
But I see no need to squash fixup commits at the same time.

Sorry, I got a habit of squashing the review fixes when rebasing, mostly because other collaborators told me they preferred that way.

nodejs-github-bot · 2026-07-19T09:01:16Z

CI: https://ci.nodejs.org/job/node-test-pull-request/74887/

nodejs-github-bot · 2026-07-19T18:30:08Z

Landed in d25cde4

After firing timers, we can clean them up by iterating over all active stores and setting the relevant symbols to undefined. Fixes#53408 Signed-off-by: Matteo Collina <hello@matteocollina.com> PR-URL: #53443 Reviewed-By: Stephen Belanger <admin@stephenbelanger.com> Reviewed-By: Paolo Insogna <paolo@cowtech.it> Reviewed-By: James M Snell <jasnell@gmail.com> Reviewed-By: Gerhard Stöbich <deb2001-github@yahoo.de>

mcollina · 2024-06-13T17:11:48Z

After firing timers, we can clean them up by iterating over all active stores and setting the relevant symbols to undefined.

This is still a draft because we will need to extend it to immediates and intervals, too.

Fixes#53408

mcollina · 2024-06-13T17:13:47Z

I've opened a draft PR to discuss a possible solution to #53408.

cc @nodejs/diagnostics @nodejs/timers let me know what you think.

Flarna · 2024-07-02T20:41:45Z

@@ -429,6 +432,20 @@ function setPosition(node, pos) {
 node.priorityQueuePosition = pos;
 }

+function removeAllStores (timer) {


What about moving this function into async_hooks to avoid the need to expose getActiveStores() there?
An alternative would be to use a single object on the resource which holds all stores. This would reduce the number of properties added on resource objects - and avoid the need of a for loop to clear it.

A helper that receives a resource and clears all the store data off it would be better than exposing the stores list as we do now. As it is presently, one could get and retain access to a store in a way unintentional from the perspective of the store owner, which is a bit dangerous from a memory perspective.

mcollina · 2026-05-31T19:42:12Z

I've updated this PR to address the review feedback:

Instead of exposing getActiveStores(), added a clearStores(resource) helper function in async_hooks that takes a resource and clears all store references from it. This avoids exposing the internal stores list.
Added cleanup for _timerArgs in addition to _onTimeout and async stores.
Extended the cleanup to immediates (setImmediate) and intervals (setInterval) as well.
Updated the test to cover all three cases.

cc @Flarna @Qard

codecov · 2026-05-31T20:38:20Z

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 90.23%. Comparing base (1d87a24) to head (ff9bebb).
⚠️ Report is 94 commits behind head on main.

Additional details and impacted files

@@ Coverage Diff @@## main #53443 +/- ##
==========================================
- Coverage 90.25% 90.23% -0.03% 
==========================================
Files 741 741 Lines 240990 241082 +92 Branches 45402 45425 +23 ==========================================
+ Hits 217502 217535 +33 - Misses 15066 15116 +50 - Partials 8422 8431 +9

Files with missing lines	Coverage Δ
lib/internal/async_hooks.js	`99.36% <100.00%> (+<0.01%)`	⬆️
lib/internal/async_local_storage/async_hooks.js	`98.21% <100.00%> (+0.16%)`	⬆️
lib/internal/timers.js	`99.75% <100.00%> (+0.02%)`	⬆️
lib/timers.js	`100.00% <100.00%> (ø)`

... and 39 files with indirect coverage changes

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

After firing timers, we can clean them up by iterating over all active stores and setting the relevant symbols to undefined. This also covers immediates and intervals. Refs: nodejs#53408 PR-URL: nodejs#53443Fixes: nodejs#53408 Signed-off-by: Matteo Collina <hello@matteocollina.com>

Flarna · 2026-06-18T21:07:29Z


 const async_context_frame = Symbol('kAsyncContextFrame');

+function removeStoresFromResource(resource) {


I think this should be extended to somthing like this to cover also the async context frame variant of ALS:
functionremoveStoresFromResource(resource){if(AsyncContextFrame.enabled){if(resource[async_context_frame]!==undefined){resource[async_context_frame]=undefined;}}else{if(resource[async_local_storage_context_symbol]!==undefined){resource[async_local_storage_context_symbol]=undefined;}}}

Flarna · 2026-06-18T21:08:41Z

@@ -0,0 +1,109 @@
+// Flags: --expose-internals --no-async-context-frame


I think this should be tested with and without async-context-frame
And a test via GC would be likely helpful, e.g. somthing like this:
{constasyncLocalStorage=newAsyncLocalStorage();letstore={};lettimeout;constweakStore=newWeakRef(store);asyncLocalStorage.run(store,common.mustCall(()=>{timeout=setTimeout(common.mustNotCall(),common.platformTimeout(1000));assert.strictEqual(weakStore.deref(),store);clearTimeout(timeout);}));store=null;setImmediate(()=>{global.gc();assert.strictEqual(weakStore.deref(),undefined);});}

Flarna · 2026-06-23T20:24:20Z

@mcollina seems you forced push some old commit or a mixture of old and new. At least it looks to me like parts of the recent changes got lost.

Any chance to avoid forced push to ease reviews?

mcollina · 2026-06-24T09:52:43Z

I had to force-push for rebase.

mcollina · 2026-07-08T08:01:44Z

@Flarna PTAL

After firing timers, we can clean them up by iterating over all active stores and setting the relevant symbols to undefined. Fixesnodejs#53408 Signed-off-by: Matteo Collina <hello@matteocollina.com>

Flarna · 2026-07-08T13:24:41Z

I still don't understand why rebase+squash is done again and again in this PR.
I get that a rebase is needed every now and then, in special for that long running PRs.
But I see no need to squash fixup commits at the same time.

mcollina · 2026-07-09T09:19:09Z

I still don't understand why rebase+squash is done again and again in this PR.
I get that a rebase is needed every now and then, in special for that long running PRs.
But I see no need to squash fixup commits at the same time.

Sorry, I got a habit of squashing the review fixes when rebasing, mostly because other collaborators told me they preferred that way.

nodejs-github-bot · 2026-07-19T09:01:16Z

CI: https://ci.nodejs.org/job/node-test-pull-request/74887/

nodejs-github-bot · 2026-07-19T18:30:08Z

Landed in d25cde4

After firing timers, we can clean them up by iterating over all active stores and setting the relevant symbols to undefined. Fixes#53408 Signed-off-by: Matteo Collina <hello@matteocollina.com> PR-URL: #53443 Reviewed-By: Stephen Belanger <admin@stephenbelanger.com> Reviewed-By: Paolo Insogna <paolo@cowtech.it> Reviewed-By: James M Snell <jasnell@gmail.com> Reviewed-By: Gerhard Stöbich <deb2001-github@yahoo.de>

mcollina · 2024-06-13T17:11:48Z

After firing timers, we can clean them up by iterating over all active stores and setting the relevant symbols to undefined.

This is still a draft because we will need to extend it to immediates and intervals, too.

Fixes#53408

mcollina · 2024-06-13T17:13:47Z

I've opened a draft PR to discuss a possible solution to #53408.

cc @nodejs/diagnostics @nodejs/timers let me know what you think.

Flarna · 2024-07-02T20:41:45Z

@@ -429,6 +432,20 @@ function setPosition(node, pos) {
 node.priorityQueuePosition = pos;
 }

+function removeAllStores (timer) {


What about moving this function into async_hooks to avoid the need to expose getActiveStores() there?
An alternative would be to use a single object on the resource which holds all stores. This would reduce the number of properties added on resource objects - and avoid the need of a for loop to clear it.

A helper that receives a resource and clears all the store data off it would be better than exposing the stores list as we do now. As it is presently, one could get and retain access to a store in a way unintentional from the perspective of the store owner, which is a bit dangerous from a memory perspective.

mcollina · 2026-05-31T19:42:12Z

I've updated this PR to address the review feedback:

Instead of exposing getActiveStores(), added a clearStores(resource) helper function in async_hooks that takes a resource and clears all store references from it. This avoids exposing the internal stores list.
Added cleanup for _timerArgs in addition to _onTimeout and async stores.
Extended the cleanup to immediates (setImmediate) and intervals (setInterval) as well.
Updated the test to cover all three cases.

cc @Flarna @Qard

codecov · 2026-05-31T20:38:20Z

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 90.23%. Comparing base (1d87a24) to head (ff9bebb).
⚠️ Report is 94 commits behind head on main.

Additional details and impacted files

@@ Coverage Diff @@## main #53443 +/- ##
==========================================
- Coverage 90.25% 90.23% -0.03% 
==========================================
Files 741 741 Lines 240990 241082 +92 Branches 45402 45425 +23 ==========================================
+ Hits 217502 217535 +33 - Misses 15066 15116 +50 - Partials 8422 8431 +9

Files with missing lines	Coverage Δ
lib/internal/async_hooks.js	`99.36% <100.00%> (+<0.01%)`	⬆️
lib/internal/async_local_storage/async_hooks.js	`98.21% <100.00%> (+0.16%)`	⬆️
lib/internal/timers.js	`99.75% <100.00%> (+0.02%)`	⬆️
lib/timers.js	`100.00% <100.00%> (ø)`

... and 39 files with indirect coverage changes

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

After firing timers, we can clean them up by iterating over all active stores and setting the relevant symbols to undefined. This also covers immediates and intervals. Refs: nodejs#53408 PR-URL: nodejs#53443Fixes: nodejs#53408 Signed-off-by: Matteo Collina <hello@matteocollina.com>

Flarna · 2026-06-18T21:07:29Z


 const async_context_frame = Symbol('kAsyncContextFrame');

+function removeStoresFromResource(resource) {


I think this should be extended to somthing like this to cover also the async context frame variant of ALS:
functionremoveStoresFromResource(resource){if(AsyncContextFrame.enabled){if(resource[async_context_frame]!==undefined){resource[async_context_frame]=undefined;}}else{if(resource[async_local_storage_context_symbol]!==undefined){resource[async_local_storage_context_symbol]=undefined;}}}

Flarna · 2026-06-18T21:08:41Z

@@ -0,0 +1,109 @@
+// Flags: --expose-internals --no-async-context-frame


I think this should be tested with and without async-context-frame
And a test via GC would be likely helpful, e.g. somthing like this:
{constasyncLocalStorage=newAsyncLocalStorage();letstore={};lettimeout;constweakStore=newWeakRef(store);asyncLocalStorage.run(store,common.mustCall(()=>{timeout=setTimeout(common.mustNotCall(),common.platformTimeout(1000));assert.strictEqual(weakStore.deref(),store);clearTimeout(timeout);}));store=null;setImmediate(()=>{global.gc();assert.strictEqual(weakStore.deref(),undefined);});}

Flarna · 2026-06-23T20:24:20Z

@mcollina seems you forced push some old commit or a mixture of old and new. At least it looks to me like parts of the recent changes got lost.

Any chance to avoid forced push to ease reviews?

mcollina · 2026-06-24T09:52:43Z

I had to force-push for rebase.

mcollina · 2026-07-08T08:01:44Z

@Flarna PTAL

After firing timers, we can clean them up by iterating over all active stores and setting the relevant symbols to undefined. Fixesnodejs#53408 Signed-off-by: Matteo Collina <hello@matteocollina.com>

Flarna · 2026-07-08T13:24:41Z

I still don't understand why rebase+squash is done again and again in this PR.
I get that a rebase is needed every now and then, in special for that long running PRs.
But I see no need to squash fixup commits at the same time.

mcollina · 2026-07-09T09:19:09Z

I still don't understand why rebase+squash is done again and again in this PR.
I get that a rebase is needed every now and then, in special for that long running PRs.
But I see no need to squash fixup commits at the same time.

Sorry, I got a habit of squashing the review fixes when rebasing, mostly because other collaborators told me they preferred that way.

nodejs-github-bot · 2026-07-19T09:01:16Z

CI: https://ci.nodejs.org/job/node-test-pull-request/74887/

nodejs-github-bot · 2026-07-19T18:30:08Z

Landed in d25cde4

After firing timers, we can clean them up by iterating over all active stores and setting the relevant symbols to undefined. Fixes#53408 Signed-off-by: Matteo Collina <hello@matteocollina.com> PR-URL: #53443 Reviewed-By: Stephen Belanger <admin@stephenbelanger.com> Reviewed-By: Paolo Insogna <paolo@cowtech.it> Reviewed-By: James M Snell <jasnell@gmail.com> Reviewed-By: Gerhard Stöbich <deb2001-github@yahoo.de>

mcollina · 2024-06-13T17:11:48Z

After firing timers, we can clean them up by iterating over all active stores and setting the relevant symbols to undefined.

This is still a draft because we will need to extend it to immediates and intervals, too.

Fixes#53408

mcollina · 2024-06-13T17:13:47Z

I've opened a draft PR to discuss a possible solution to #53408.

cc @nodejs/diagnostics @nodejs/timers let me know what you think.

Flarna · 2024-07-02T20:41:45Z

@@ -429,6 +432,20 @@ function setPosition(node, pos) {
 node.priorityQueuePosition = pos;
 }

+function removeAllStores (timer) {


What about moving this function into async_hooks to avoid the need to expose getActiveStores() there?
An alternative would be to use a single object on the resource which holds all stores. This would reduce the number of properties added on resource objects - and avoid the need of a for loop to clear it.

A helper that receives a resource and clears all the store data off it would be better than exposing the stores list as we do now. As it is presently, one could get and retain access to a store in a way unintentional from the perspective of the store owner, which is a bit dangerous from a memory perspective.

mcollina · 2026-05-31T19:42:12Z

I've updated this PR to address the review feedback:

Instead of exposing getActiveStores(), added a clearStores(resource) helper function in async_hooks that takes a resource and clears all store references from it. This avoids exposing the internal stores list.
Added cleanup for _timerArgs in addition to _onTimeout and async stores.
Extended the cleanup to immediates (setImmediate) and intervals (setInterval) as well.
Updated the test to cover all three cases.

cc @Flarna @Qard

codecov · 2026-05-31T20:38:20Z

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 90.23%. Comparing base (1d87a24) to head (ff9bebb).
⚠️ Report is 94 commits behind head on main.

Additional details and impacted files

@@ Coverage Diff @@## main #53443 +/- ##
==========================================
- Coverage 90.25% 90.23% -0.03% 
==========================================
Files 741 741 Lines 240990 241082 +92 Branches 45402 45425 +23 ==========================================
+ Hits 217502 217535 +33 - Misses 15066 15116 +50 - Partials 8422 8431 +9

Files with missing lines	Coverage Δ
lib/internal/async_hooks.js	`99.36% <100.00%> (+<0.01%)`	⬆️
lib/internal/async_local_storage/async_hooks.js	`98.21% <100.00%> (+0.16%)`	⬆️
lib/internal/timers.js	`99.75% <100.00%> (+0.02%)`	⬆️
lib/timers.js	`100.00% <100.00%> (ø)`

... and 39 files with indirect coverage changes

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

After firing timers, we can clean them up by iterating over all active stores and setting the relevant symbols to undefined. This also covers immediates and intervals. Refs: nodejs#53408 PR-URL: nodejs#53443Fixes: nodejs#53408 Signed-off-by: Matteo Collina <hello@matteocollina.com>

Flarna · 2026-06-18T21:07:29Z


 const async_context_frame = Symbol('kAsyncContextFrame');

+function removeStoresFromResource(resource) {


I think this should be extended to somthing like this to cover also the async context frame variant of ALS:
functionremoveStoresFromResource(resource){if(AsyncContextFrame.enabled){if(resource[async_context_frame]!==undefined){resource[async_context_frame]=undefined;}}else{if(resource[async_local_storage_context_symbol]!==undefined){resource[async_local_storage_context_symbol]=undefined;}}}

Flarna · 2026-06-18T21:08:41Z

@@ -0,0 +1,109 @@
+// Flags: --expose-internals --no-async-context-frame


I think this should be tested with and without async-context-frame
And a test via GC would be likely helpful, e.g. somthing like this:
{constasyncLocalStorage=newAsyncLocalStorage();letstore={};lettimeout;constweakStore=newWeakRef(store);asyncLocalStorage.run(store,common.mustCall(()=>{timeout=setTimeout(common.mustNotCall(),common.platformTimeout(1000));assert.strictEqual(weakStore.deref(),store);clearTimeout(timeout);}));store=null;setImmediate(()=>{global.gc();assert.strictEqual(weakStore.deref(),undefined);});}

Flarna · 2026-06-23T20:24:20Z

@mcollina seems you forced push some old commit or a mixture of old and new. At least it looks to me like parts of the recent changes got lost.

Any chance to avoid forced push to ease reviews?

mcollina · 2026-06-24T09:52:43Z

I had to force-push for rebase.

mcollina · 2026-07-08T08:01:44Z

@Flarna PTAL

After firing timers, we can clean them up by iterating over all active stores and setting the relevant symbols to undefined. Fixesnodejs#53408 Signed-off-by: Matteo Collina <hello@matteocollina.com>

Flarna · 2026-07-08T13:24:41Z

I still don't understand why rebase+squash is done again and again in this PR.
I get that a rebase is needed every now and then, in special for that long running PRs.
But I see no need to squash fixup commits at the same time.

mcollina · 2026-07-09T09:19:09Z

I still don't understand why rebase+squash is done again and again in this PR.
I get that a rebase is needed every now and then, in special for that long running PRs.
But I see no need to squash fixup commits at the same time.

Sorry, I got a habit of squashing the review fixes when rebasing, mostly because other collaborators told me they preferred that way.

nodejs-github-bot · 2026-07-19T09:01:16Z

CI: https://ci.nodejs.org/job/node-test-pull-request/74887/

nodejs-github-bot · 2026-07-19T18:30:08Z

Landed in d25cde4

After firing timers, we can clean them up by iterating over all active stores and setting the relevant symbols to undefined. Fixes#53408 Signed-off-by: Matteo Collina <hello@matteocollina.com> PR-URL: #53443 Reviewed-By: Stephen Belanger <admin@stephenbelanger.com> Reviewed-By: Paolo Insogna <paolo@cowtech.it> Reviewed-By: James M Snell <jasnell@gmail.com> Reviewed-By: Gerhard Stöbich <deb2001-github@yahoo.de>

mcollina · 2024-06-13T17:11:48Z

After firing timers, we can clean them up by iterating over all active stores and setting the relevant symbols to undefined.

This is still a draft because we will need to extend it to immediates and intervals, too.

Fixes#53408

mcollina · 2024-06-13T17:13:47Z

I've opened a draft PR to discuss a possible solution to #53408.

cc @nodejs/diagnostics @nodejs/timers let me know what you think.

Flarna · 2024-07-02T20:41:45Z

@@ -429,6 +432,20 @@ function setPosition(node, pos) {
 node.priorityQueuePosition = pos;
 }

+function removeAllStores (timer) {


What about moving this function into async_hooks to avoid the need to expose getActiveStores() there?
An alternative would be to use a single object on the resource which holds all stores. This would reduce the number of properties added on resource objects - and avoid the need of a for loop to clear it.

A helper that receives a resource and clears all the store data off it would be better than exposing the stores list as we do now. As it is presently, one could get and retain access to a store in a way unintentional from the perspective of the store owner, which is a bit dangerous from a memory perspective.

mcollina · 2026-05-31T19:42:12Z

I've updated this PR to address the review feedback:

Instead of exposing getActiveStores(), added a clearStores(resource) helper function in async_hooks that takes a resource and clears all store references from it. This avoids exposing the internal stores list.
Added cleanup for _timerArgs in addition to _onTimeout and async stores.
Extended the cleanup to immediates (setImmediate) and intervals (setInterval) as well.
Updated the test to cover all three cases.

cc @Flarna @Qard

codecov · 2026-05-31T20:38:20Z

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 90.23%. Comparing base (1d87a24) to head (ff9bebb).
⚠️ Report is 94 commits behind head on main.

Additional details and impacted files

@@ Coverage Diff @@## main #53443 +/- ##
==========================================
- Coverage 90.25% 90.23% -0.03% 
==========================================
Files 741 741 Lines 240990 241082 +92 Branches 45402 45425 +23 ==========================================
+ Hits 217502 217535 +33 - Misses 15066 15116 +50 - Partials 8422 8431 +9

Files with missing lines	Coverage Δ
lib/internal/async_hooks.js	`99.36% <100.00%> (+<0.01%)`	⬆️
lib/internal/async_local_storage/async_hooks.js	`98.21% <100.00%> (+0.16%)`	⬆️
lib/internal/timers.js	`99.75% <100.00%> (+0.02%)`	⬆️
lib/timers.js	`100.00% <100.00%> (ø)`

... and 39 files with indirect coverage changes

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

After firing timers, we can clean them up by iterating over all active stores and setting the relevant symbols to undefined. This also covers immediates and intervals. Refs: nodejs#53408 PR-URL: nodejs#53443Fixes: nodejs#53408 Signed-off-by: Matteo Collina <hello@matteocollina.com>

Flarna · 2026-06-18T21:07:29Z


 const async_context_frame = Symbol('kAsyncContextFrame');

+function removeStoresFromResource(resource) {


I think this should be extended to somthing like this to cover also the async context frame variant of ALS:
functionremoveStoresFromResource(resource){if(AsyncContextFrame.enabled){if(resource[async_context_frame]!==undefined){resource[async_context_frame]=undefined;}}else{if(resource[async_local_storage_context_symbol]!==undefined){resource[async_local_storage_context_symbol]=undefined;}}}

Flarna · 2026-06-18T21:08:41Z

@@ -0,0 +1,109 @@
+// Flags: --expose-internals --no-async-context-frame


I think this should be tested with and without async-context-frame
And a test via GC would be likely helpful, e.g. somthing like this:
{constasyncLocalStorage=newAsyncLocalStorage();letstore={};lettimeout;constweakStore=newWeakRef(store);asyncLocalStorage.run(store,common.mustCall(()=>{timeout=setTimeout(common.mustNotCall(),common.platformTimeout(1000));assert.strictEqual(weakStore.deref(),store);clearTimeout(timeout);}));store=null;setImmediate(()=>{global.gc();assert.strictEqual(weakStore.deref(),undefined);});}

Flarna · 2026-06-23T20:24:20Z

@mcollina seems you forced push some old commit or a mixture of old and new. At least it looks to me like parts of the recent changes got lost.

Any chance to avoid forced push to ease reviews?

mcollina · 2026-06-24T09:52:43Z

I had to force-push for rebase.

mcollina · 2026-07-08T08:01:44Z

@Flarna PTAL

After firing timers, we can clean them up by iterating over all active stores and setting the relevant symbols to undefined. Fixesnodejs#53408 Signed-off-by: Matteo Collina <hello@matteocollina.com>

Flarna · 2026-07-08T13:24:41Z

I still don't understand why rebase+squash is done again and again in this PR.
I get that a rebase is needed every now and then, in special for that long running PRs.
But I see no need to squash fixup commits at the same time.

mcollina · 2026-07-09T09:19:09Z

I still don't understand why rebase+squash is done again and again in this PR.
I get that a rebase is needed every now and then, in special for that long running PRs.
But I see no need to squash fixup commits at the same time.

Sorry, I got a habit of squashing the review fixes when rebasing, mostly because other collaborators told me they preferred that way.

nodejs-github-bot · 2026-07-19T09:01:16Z

CI: https://ci.nodejs.org/job/node-test-pull-request/74887/

nodejs-github-bot · 2026-07-19T18:30:08Z

Landed in d25cde4

After firing timers, we can clean them up by iterating over all active stores and setting the relevant symbols to undefined. Fixes#53408 Signed-off-by: Matteo Collina <hello@matteocollina.com> PR-URL: #53443 Reviewed-By: Stephen Belanger <admin@stephenbelanger.com> Reviewed-By: Paolo Insogna <paolo@cowtech.it> Reviewed-By: James M Snell <jasnell@gmail.com> Reviewed-By: Gerhard Stöbich <deb2001-github@yahoo.de>

nodejs-github-bot added async_hooks Issues and PRs related to the async hooks subsystem. needs-ci PRs that need a full CI run. timers Issues and PRs related to timers, setImmediate(), setInterval(), and setTimeout(). labels Jun 13, 2024

mcollina force-pushed the timers-leak branch from 07a897b to 00931d1Compare June 13, 2024 17:15

Flarna reviewed Jul 2, 2024
View reviewed changes

Comment threadlib/internal/timers.js Outdated

mcollina force-pushed the timers-leak branch from 5321fa8 to 0ad04f6Compare May 31, 2026 18:37

mcollina marked this pull request as ready for review May 31, 2026 19:12

Qard approved these changes Jun 1, 2026
View reviewed changes

Flarna reviewed Jun 2, 2026
View reviewed changes

Comment threadlib/internal/async_local_storage/async_hooks.js Outdated

Flarna reviewed Jun 2, 2026
View reviewed changes

Comment threadlib/internal/timers.js Outdated

Flarna reviewed Jun 2, 2026
View reviewed changes

Comment threadlib/async_hooks.js Outdated

mcollina force-pushed the timers-leak branch from 0ad04f6 to fe5b310Compare June 16, 2026 14:18

mcollina force-pushed the timers-leak branch from fe5b310 to 637251eCompare June 16, 2026 15:08

mcollina force-pushed the timers-leak branch from 637251e to b83e431Compare June 16, 2026 16:53

ShogunPanda approved these changes Jun 17, 2026
View reviewed changes

Qard approved these changes Jun 17, 2026
View reviewed changes

Flarna reviewed Jun 18, 2026
View reviewed changes

mcollina force-pushed the timers-leak branch from b83e431 to 522e6ebCompare June 23, 2026 13:49

mcollina force-pushed the timers-leak branch from 522e6eb to c2a2673Compare June 25, 2026 11:04

Flarna reviewed Jun 26, 2026
View reviewed changes

Comment threadlib/timers.js Outdated

Flarna reviewed Jun 26, 2026
View reviewed changes

Comment threadlib/internal/timers.js

jasnell approved these changes Jun 27, 2026
View reviewed changes

ShogunPanda approved these changes Jul 8, 2026
View reviewed changes

mcollina added the request-ci Add this label to start a Jenkins CI on a PR. label Jul 8, 2026

github-actionsBot removed the request-ci Add this label to start a Jenkins CI on a PR. label Jul 8, 2026

This comment was marked as outdated.
Sign in to view

mcollina force-pushed the timers-leak branch from b9fc9e4 to 1e824aeCompare July 8, 2026 09:06

timers: do not retain a reference to the async store after firing
ff9bebb
After firing timers, we can clean them up by iterating over all active stores and setting the relevant symbols to undefined. Fixesnodejs#53408 Signed-off-by: Matteo Collina <hello@matteocollina.com>

mcollina force-pushed the timers-leak branch from 1e824ae to ff9bebbCompare July 8, 2026 11:47

Qard approved these changes Jul 17, 2026
View reviewed changes

mcollina added the request-ci Add this label to start a Jenkins CI on a PR. label Jul 18, 2026

github-actionsBot removed the request-ci Add this label to start a Jenkins CI on a PR. label Jul 18, 2026

This comment was marked as outdated.
Sign in to view

Flarna approved these changes Jul 19, 2026
View reviewed changes

Flarna added the commit-queue PRs queued for automated landing through the Commit Queue. label Jul 19, 2026

nodejs-github-bot removed the commit-queue PRs queued for automated landing through the Commit Queue. label Jul 19, 2026

nodejs-github-bot merged commit d25cde4 into nodejs:mainJul 19, 2026
71 checks passed

github-actionsBot mentioned this pull request Jul 21, 2026
2026-07-22, Version 26.6.0 (Current) #64655
Merged

aduh95 mentioned this pull request Aug 22, 2026
2026-08-26, Version 24.20.0 'Krypton' (LTS) #65461
Merged