test: remove impact of tier-up changes in worker stack size test by joyeecheung · Pull Request #64271 · nodejs/node

joyeecheung · 2026-07-03T17:23:22Z

Background: this is flaking in the canary as https://chromium-review.googlesource.com/c/v8/v8/+/7761796 made the function eligible for maglev tier-up. Locally this made the flake disappear for me. Opening against the main branch since it makes sense here anyway.

The depth of the stack depends not only on the stack size, but also on the size of each stack frame, which in turn depends on which tier the recursive function happens to be running at when the overflow occurs. Under load the background tier-up can land at a non-deterministic point in the recursion and flake the test. Keep the recursive function in the interpreter with %NeverOptimizeFunction() so the frame size - and thus the depth - is deterministic.

The depth of the stack depends not only on the stack size, but also on the size of each stack frame, which in turn depends on which tier the recursive function happens to be running at when the overflow occurs. Under load the background tier-up can land at a non-deterministic point in the recursion and flake the test. Keep the recursive function in the interpreter with %NeverOptimizeFunction() so the frame size - and thus the depth - is deterministic. Signed-off-by: Joyee Cheung <joyeec9h3@gmail.com>

nodejs-github-bot · 2026-07-06T20:51:44Z

CI: https://ci.nodejs.org/job/node-test-pull-request/74610/ 💚

nodejs-github-bot · 2026-07-07T11:53:30Z

Landed in 3257a36

The depth of the stack depends not only on the stack size, but also on the size of each stack frame, which in turn depends on which tier the recursive function happens to be running at when the overflow occurs. Under load the background tier-up can land at a non-deterministic point in the recursion and flake the test. Keep the recursive function in the interpreter with %NeverOptimizeFunction() so the frame size - and thus the depth - is deterministic. Signed-off-by: Joyee Cheung <joyeec9h3@gmail.com> PR-URL: #64271 Reviewed-By: Michaël Zasso <targos@protonmail.com> Reviewed-By: James M Snell <jasnell@gmail.com>

…ite (#53062) Backports nodejs/node#64271. The test compares the recursion depth two Workers reach before overflowing, but depth also depends on frame size, which changes when the recursive function tiers up mid-recursion. On a busy CI runner that lands non-deterministically and the depths end up >10% apart. Pin the function to the interpreter so the depth is stable. The fix is only in Node 26.6.0+, so it is carried as a patch until the Node roll picks it up.

…ite (#53075) Backports nodejs/node#64271. The test compares the recursion depth two Workers reach before overflowing, but depth also depends on frame size, which changes when the recursive function tiers up mid-recursion. On a busy CI runner that lands non-deterministically and the depths end up >10% apart. Pin the function to the interpreter so the depth is stable. The fix is only in Node 26.6.0+, so it is carried as a patch until the Node roll picks it up. Co-authored-by: trop[bot] <37223003+trop[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

…ite (#53076) Backports nodejs/node#64271. The test compares the recursion depth two Workers reach before overflowing, but depth also depends on frame size, which changes when the recursive function tiers up mid-recursion. On a busy CI runner that lands non-deterministically and the depths end up >10% apart. Pin the function to the interpreter so the depth is stable. The fix is only in Node 26.6.0+, so it is carried as a patch until the Node roll picks it up. Co-authored-by: trop[bot] <37223003+trop[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

…ite (#53077) Backports nodejs/node#64271. The test compares the recursion depth two Workers reach before overflowing, but depth also depends on frame size, which changes when the recursive function tiers up mid-recursion. On a busy CI runner that lands non-deterministically and the depths end up >10% apart. Pin the function to the interpreter so the depth is stable. The fix is only in Node 26.6.0+, so it is carried as a patch until the Node roll picks it up. Co-authored-by: trop[bot] <37223003+trop[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

…ite (#53078) Backports nodejs/node#64271. The test compares the recursion depth two Workers reach before overflowing, but depth also depends on frame size, which changes when the recursive function tiers up mid-recursion. On a busy CI runner that lands non-deterministically and the depths end up >10% apart. Pin the function to the interpreter so the depth is stable. The fix is only in Node 26.6.0+, so it is carried as a patch until the Node roll picks it up. Co-authored-by: trop[bot] <37223003+trop[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

src_fix_libuv_assertion_on_windows.patch (nodejs/node#61999) and test_remove_impact_of_tier-up_changes_in_worker_stack_size_test.patch (nodejs/node#64271) are both included in Node.js v24.20.0.

Both landed in v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271

Both patches are backports that landed in Node.js v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271

Both landed in v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271

Both patches are backports that landed in Node.js v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271

src_fix_libuv_assertion_on_windows.patch (nodejs/node#61999) and test_remove_impact_of_tier-up_changes_in_worker_stack_size_test.patch (nodejs/node#64271) are both included in Node.js v24.20.0.

* chore: bump node in DEPS to v24.20.0 * fix(patch): implement package maps in the ESM resolver Upstream added join to the path imports in resolve.js, which overlaps the import line this patch rewrites to pull in toNamespacedPath. Ref: nodejs/node#62239 * chore: remove upstreamed patches Both landed in v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271 * chore: update patches (trivial only) * fix(patch): preserve RSA-PSS legacy pubkey DER Upstream added ReadASN1Element to ncrypto.cc but only calls it from the OpenSSL 3 provider branch, so under BoringSSL it is an unused function and -Werror fails the build. Guard the definition with the same NCRYPTO_USE_OPENSSL3_PROVIDER condition as its callers. Ref: nodejs/node#64547 * fix(patch): update sqlite to 3.53.4 The sqlite update introduced an unsigned/signed comparison in rbuDeltaApply that trips -Wsign-compare under -Werror. Extend the existing sign-compare patch to cast lenDelta, which is never negative at that point, to unsigned int. Ref: nodejs/node#64745 * fix: do not pass removed JSPI V8 flag Node.js v24.20.0 passes --experimental-wasm-jspi unconditionally, but Chromium's V8 shipped JSPI and removed the flag, so node_mksnapshot aborts with 'bad option'. Add the same patch as the main-branch roll. Ref: nodejs/node#59941 Ref: https://chromium-review.googlesource.com/c/v8/v8/+/6937664 * chore: update filenames.auto.gni --------- Co-authored-by: electron-roller[bot] <84116207+electron-roller[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

* chore: bump node in DEPS to v24.20.0 * chore: remove upstreamed patches Both patches are backports that landed in Node.js v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271 * fix(patch): implement package maps in the ESM resolver Upstream rewrote packageResolve() to branch on hasPackageMap() and added join to the path imports, which shifted the context around the electron: specifier shortcut and the legacyMainResolve imports. Ref: nodejs/node#62239 * chore: update patches (trivial only) * fix(patch): split OpenSSL 3, BoringSSL, and legacy backends Upstream added an ASN.1 helper, ReadASN1Element, whose only callers live under NCRYPTO_USE_OPENSSL3_PROVIDER. With BoringSSL that macro is 0, so the helper is unused and fails -Werror=unused-function. Guard the definition with the same macro as its callers. Ref: nodejs/node#64211 * chore: update filenames.auto.gni Node.js v24.20.0 replaced src/quic/logstream.h with src/quic/session_manager.h. * fix(patch): update sqlite to 3.53.4 The new sqlite amalgamation compares 'unsigned int cnt' against 'int lenDelta' in rbuDeltaApply, which fails -Werror=sign-compare. Extend the existing sqlite sign-compare patch to cast that operand too. Ref: nodejs/node#64745 * fix(patch): do not pass removed JSPI V8 flag Node.js v24.20.0 backported JSPI enablement, which unconditionally passes --experimental-wasm-jspi to V8. Chromium's V8 removed that flag (JSPI is always on), so node_mksnapshot aborts with "bad option". Drop the argument in a new patch, matching main. Ref: nodejs/node#59941 Ref: https://chromium-review.googlesource.com/c/v8/v8/+/6937664 --------- Co-authored-by: electron-roller[bot] <84116207+electron-roller[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

joyeecheung · 2026-07-03T17:23:22Z

Background: this is flaking in the canary as https://chromium-review.googlesource.com/c/v8/v8/+/7761796 made the function eligible for maglev tier-up. Locally this made the flake disappear for me. Opening against the main branch since it makes sense here anyway.

The depth of the stack depends not only on the stack size, but also on the size of each stack frame, which in turn depends on which tier the recursive function happens to be running at when the overflow occurs. Under load the background tier-up can land at a non-deterministic point in the recursion and flake the test. Keep the recursive function in the interpreter with %NeverOptimizeFunction() so the frame size - and thus the depth - is deterministic.

The depth of the stack depends not only on the stack size, but also on the size of each stack frame, which in turn depends on which tier the recursive function happens to be running at when the overflow occurs. Under load the background tier-up can land at a non-deterministic point in the recursion and flake the test. Keep the recursive function in the interpreter with %NeverOptimizeFunction() so the frame size - and thus the depth - is deterministic. Signed-off-by: Joyee Cheung <joyeec9h3@gmail.com>

nodejs-github-bot · 2026-07-06T20:51:44Z

CI: https://ci.nodejs.org/job/node-test-pull-request/74610/ 💚

nodejs-github-bot · 2026-07-07T11:53:30Z

Landed in 3257a36

The depth of the stack depends not only on the stack size, but also on the size of each stack frame, which in turn depends on which tier the recursive function happens to be running at when the overflow occurs. Under load the background tier-up can land at a non-deterministic point in the recursion and flake the test. Keep the recursive function in the interpreter with %NeverOptimizeFunction() so the frame size - and thus the depth - is deterministic. Signed-off-by: Joyee Cheung <joyeec9h3@gmail.com> PR-URL: #64271 Reviewed-By: Michaël Zasso <targos@protonmail.com> Reviewed-By: James M Snell <jasnell@gmail.com>

…ite (#53062) Backports nodejs/node#64271. The test compares the recursion depth two Workers reach before overflowing, but depth also depends on frame size, which changes when the recursive function tiers up mid-recursion. On a busy CI runner that lands non-deterministically and the depths end up >10% apart. Pin the function to the interpreter so the depth is stable. The fix is only in Node 26.6.0+, so it is carried as a patch until the Node roll picks it up.

…ite (#53075) Backports nodejs/node#64271. The test compares the recursion depth two Workers reach before overflowing, but depth also depends on frame size, which changes when the recursive function tiers up mid-recursion. On a busy CI runner that lands non-deterministically and the depths end up >10% apart. Pin the function to the interpreter so the depth is stable. The fix is only in Node 26.6.0+, so it is carried as a patch until the Node roll picks it up. Co-authored-by: trop[bot] <37223003+trop[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

…ite (#53076) Backports nodejs/node#64271. The test compares the recursion depth two Workers reach before overflowing, but depth also depends on frame size, which changes when the recursive function tiers up mid-recursion. On a busy CI runner that lands non-deterministically and the depths end up >10% apart. Pin the function to the interpreter so the depth is stable. The fix is only in Node 26.6.0+, so it is carried as a patch until the Node roll picks it up. Co-authored-by: trop[bot] <37223003+trop[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

…ite (#53077) Backports nodejs/node#64271. The test compares the recursion depth two Workers reach before overflowing, but depth also depends on frame size, which changes when the recursive function tiers up mid-recursion. On a busy CI runner that lands non-deterministically and the depths end up >10% apart. Pin the function to the interpreter so the depth is stable. The fix is only in Node 26.6.0+, so it is carried as a patch until the Node roll picks it up. Co-authored-by: trop[bot] <37223003+trop[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

…ite (#53078) Backports nodejs/node#64271. The test compares the recursion depth two Workers reach before overflowing, but depth also depends on frame size, which changes when the recursive function tiers up mid-recursion. On a busy CI runner that lands non-deterministically and the depths end up >10% apart. Pin the function to the interpreter so the depth is stable. The fix is only in Node 26.6.0+, so it is carried as a patch until the Node roll picks it up. Co-authored-by: trop[bot] <37223003+trop[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

src_fix_libuv_assertion_on_windows.patch (nodejs/node#61999) and test_remove_impact_of_tier-up_changes_in_worker_stack_size_test.patch (nodejs/node#64271) are both included in Node.js v24.20.0.

Both landed in v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271

Both patches are backports that landed in Node.js v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271

Both landed in v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271

Both patches are backports that landed in Node.js v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271

src_fix_libuv_assertion_on_windows.patch (nodejs/node#61999) and test_remove_impact_of_tier-up_changes_in_worker_stack_size_test.patch (nodejs/node#64271) are both included in Node.js v24.20.0.

* chore: bump node in DEPS to v24.20.0 * fix(patch): implement package maps in the ESM resolver Upstream added join to the path imports in resolve.js, which overlaps the import line this patch rewrites to pull in toNamespacedPath. Ref: nodejs/node#62239 * chore: remove upstreamed patches Both landed in v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271 * chore: update patches (trivial only) * fix(patch): preserve RSA-PSS legacy pubkey DER Upstream added ReadASN1Element to ncrypto.cc but only calls it from the OpenSSL 3 provider branch, so under BoringSSL it is an unused function and -Werror fails the build. Guard the definition with the same NCRYPTO_USE_OPENSSL3_PROVIDER condition as its callers. Ref: nodejs/node#64547 * fix(patch): update sqlite to 3.53.4 The sqlite update introduced an unsigned/signed comparison in rbuDeltaApply that trips -Wsign-compare under -Werror. Extend the existing sign-compare patch to cast lenDelta, which is never negative at that point, to unsigned int. Ref: nodejs/node#64745 * fix: do not pass removed JSPI V8 flag Node.js v24.20.0 passes --experimental-wasm-jspi unconditionally, but Chromium's V8 shipped JSPI and removed the flag, so node_mksnapshot aborts with 'bad option'. Add the same patch as the main-branch roll. Ref: nodejs/node#59941 Ref: https://chromium-review.googlesource.com/c/v8/v8/+/6937664 * chore: update filenames.auto.gni --------- Co-authored-by: electron-roller[bot] <84116207+electron-roller[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

* chore: bump node in DEPS to v24.20.0 * chore: remove upstreamed patches Both patches are backports that landed in Node.js v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271 * fix(patch): implement package maps in the ESM resolver Upstream rewrote packageResolve() to branch on hasPackageMap() and added join to the path imports, which shifted the context around the electron: specifier shortcut and the legacyMainResolve imports. Ref: nodejs/node#62239 * chore: update patches (trivial only) * fix(patch): split OpenSSL 3, BoringSSL, and legacy backends Upstream added an ASN.1 helper, ReadASN1Element, whose only callers live under NCRYPTO_USE_OPENSSL3_PROVIDER. With BoringSSL that macro is 0, so the helper is unused and fails -Werror=unused-function. Guard the definition with the same macro as its callers. Ref: nodejs/node#64211 * chore: update filenames.auto.gni Node.js v24.20.0 replaced src/quic/logstream.h with src/quic/session_manager.h. * fix(patch): update sqlite to 3.53.4 The new sqlite amalgamation compares 'unsigned int cnt' against 'int lenDelta' in rbuDeltaApply, which fails -Werror=sign-compare. Extend the existing sqlite sign-compare patch to cast that operand too. Ref: nodejs/node#64745 * fix(patch): do not pass removed JSPI V8 flag Node.js v24.20.0 backported JSPI enablement, which unconditionally passes --experimental-wasm-jspi to V8. Chromium's V8 removed that flag (JSPI is always on), so node_mksnapshot aborts with "bad option". Drop the argument in a new patch, matching main. Ref: nodejs/node#59941 Ref: https://chromium-review.googlesource.com/c/v8/v8/+/6937664 --------- Co-authored-by: electron-roller[bot] <84116207+electron-roller[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

joyeecheung · 2026-07-03T17:23:22Z

Background: this is flaking in the canary as https://chromium-review.googlesource.com/c/v8/v8/+/7761796 made the function eligible for maglev tier-up. Locally this made the flake disappear for me. Opening against the main branch since it makes sense here anyway.

The depth of the stack depends not only on the stack size, but also on the size of each stack frame, which in turn depends on which tier the recursive function happens to be running at when the overflow occurs. Under load the background tier-up can land at a non-deterministic point in the recursion and flake the test. Keep the recursive function in the interpreter with %NeverOptimizeFunction() so the frame size - and thus the depth - is deterministic.

The depth of the stack depends not only on the stack size, but also on the size of each stack frame, which in turn depends on which tier the recursive function happens to be running at when the overflow occurs. Under load the background tier-up can land at a non-deterministic point in the recursion and flake the test. Keep the recursive function in the interpreter with %NeverOptimizeFunction() so the frame size - and thus the depth - is deterministic. Signed-off-by: Joyee Cheung <joyeec9h3@gmail.com>

nodejs-github-bot · 2026-07-06T20:51:44Z

CI: https://ci.nodejs.org/job/node-test-pull-request/74610/ 💚

nodejs-github-bot · 2026-07-07T11:53:30Z

Landed in 3257a36

The depth of the stack depends not only on the stack size, but also on the size of each stack frame, which in turn depends on which tier the recursive function happens to be running at when the overflow occurs. Under load the background tier-up can land at a non-deterministic point in the recursion and flake the test. Keep the recursive function in the interpreter with %NeverOptimizeFunction() so the frame size - and thus the depth - is deterministic. Signed-off-by: Joyee Cheung <joyeec9h3@gmail.com> PR-URL: #64271 Reviewed-By: Michaël Zasso <targos@protonmail.com> Reviewed-By: James M Snell <jasnell@gmail.com>

…ite (#53062) Backports nodejs/node#64271. The test compares the recursion depth two Workers reach before overflowing, but depth also depends on frame size, which changes when the recursive function tiers up mid-recursion. On a busy CI runner that lands non-deterministically and the depths end up >10% apart. Pin the function to the interpreter so the depth is stable. The fix is only in Node 26.6.0+, so it is carried as a patch until the Node roll picks it up.

…ite (#53075) Backports nodejs/node#64271. The test compares the recursion depth two Workers reach before overflowing, but depth also depends on frame size, which changes when the recursive function tiers up mid-recursion. On a busy CI runner that lands non-deterministically and the depths end up >10% apart. Pin the function to the interpreter so the depth is stable. The fix is only in Node 26.6.0+, so it is carried as a patch until the Node roll picks it up. Co-authored-by: trop[bot] <37223003+trop[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

…ite (#53076) Backports nodejs/node#64271. The test compares the recursion depth two Workers reach before overflowing, but depth also depends on frame size, which changes when the recursive function tiers up mid-recursion. On a busy CI runner that lands non-deterministically and the depths end up >10% apart. Pin the function to the interpreter so the depth is stable. The fix is only in Node 26.6.0+, so it is carried as a patch until the Node roll picks it up. Co-authored-by: trop[bot] <37223003+trop[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

…ite (#53077) Backports nodejs/node#64271. The test compares the recursion depth two Workers reach before overflowing, but depth also depends on frame size, which changes when the recursive function tiers up mid-recursion. On a busy CI runner that lands non-deterministically and the depths end up >10% apart. Pin the function to the interpreter so the depth is stable. The fix is only in Node 26.6.0+, so it is carried as a patch until the Node roll picks it up. Co-authored-by: trop[bot] <37223003+trop[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

…ite (#53078) Backports nodejs/node#64271. The test compares the recursion depth two Workers reach before overflowing, but depth also depends on frame size, which changes when the recursive function tiers up mid-recursion. On a busy CI runner that lands non-deterministically and the depths end up >10% apart. Pin the function to the interpreter so the depth is stable. The fix is only in Node 26.6.0+, so it is carried as a patch until the Node roll picks it up. Co-authored-by: trop[bot] <37223003+trop[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

src_fix_libuv_assertion_on_windows.patch (nodejs/node#61999) and test_remove_impact_of_tier-up_changes_in_worker_stack_size_test.patch (nodejs/node#64271) are both included in Node.js v24.20.0.

Both landed in v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271

Both patches are backports that landed in Node.js v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271

Both landed in v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271

Both patches are backports that landed in Node.js v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271

src_fix_libuv_assertion_on_windows.patch (nodejs/node#61999) and test_remove_impact_of_tier-up_changes_in_worker_stack_size_test.patch (nodejs/node#64271) are both included in Node.js v24.20.0.

* chore: bump node in DEPS to v24.20.0 * fix(patch): implement package maps in the ESM resolver Upstream added join to the path imports in resolve.js, which overlaps the import line this patch rewrites to pull in toNamespacedPath. Ref: nodejs/node#62239 * chore: remove upstreamed patches Both landed in v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271 * chore: update patches (trivial only) * fix(patch): preserve RSA-PSS legacy pubkey DER Upstream added ReadASN1Element to ncrypto.cc but only calls it from the OpenSSL 3 provider branch, so under BoringSSL it is an unused function and -Werror fails the build. Guard the definition with the same NCRYPTO_USE_OPENSSL3_PROVIDER condition as its callers. Ref: nodejs/node#64547 * fix(patch): update sqlite to 3.53.4 The sqlite update introduced an unsigned/signed comparison in rbuDeltaApply that trips -Wsign-compare under -Werror. Extend the existing sign-compare patch to cast lenDelta, which is never negative at that point, to unsigned int. Ref: nodejs/node#64745 * fix: do not pass removed JSPI V8 flag Node.js v24.20.0 passes --experimental-wasm-jspi unconditionally, but Chromium's V8 shipped JSPI and removed the flag, so node_mksnapshot aborts with 'bad option'. Add the same patch as the main-branch roll. Ref: nodejs/node#59941 Ref: https://chromium-review.googlesource.com/c/v8/v8/+/6937664 * chore: update filenames.auto.gni --------- Co-authored-by: electron-roller[bot] <84116207+electron-roller[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

* chore: bump node in DEPS to v24.20.0 * chore: remove upstreamed patches Both patches are backports that landed in Node.js v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271 * fix(patch): implement package maps in the ESM resolver Upstream rewrote packageResolve() to branch on hasPackageMap() and added join to the path imports, which shifted the context around the electron: specifier shortcut and the legacyMainResolve imports. Ref: nodejs/node#62239 * chore: update patches (trivial only) * fix(patch): split OpenSSL 3, BoringSSL, and legacy backends Upstream added an ASN.1 helper, ReadASN1Element, whose only callers live under NCRYPTO_USE_OPENSSL3_PROVIDER. With BoringSSL that macro is 0, so the helper is unused and fails -Werror=unused-function. Guard the definition with the same macro as its callers. Ref: nodejs/node#64211 * chore: update filenames.auto.gni Node.js v24.20.0 replaced src/quic/logstream.h with src/quic/session_manager.h. * fix(patch): update sqlite to 3.53.4 The new sqlite amalgamation compares 'unsigned int cnt' against 'int lenDelta' in rbuDeltaApply, which fails -Werror=sign-compare. Extend the existing sqlite sign-compare patch to cast that operand too. Ref: nodejs/node#64745 * fix(patch): do not pass removed JSPI V8 flag Node.js v24.20.0 backported JSPI enablement, which unconditionally passes --experimental-wasm-jspi to V8. Chromium's V8 removed that flag (JSPI is always on), so node_mksnapshot aborts with "bad option". Drop the argument in a new patch, matching main. Ref: nodejs/node#59941 Ref: https://chromium-review.googlesource.com/c/v8/v8/+/6937664 --------- Co-authored-by: electron-roller[bot] <84116207+electron-roller[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

joyeecheung · 2026-07-03T17:23:22Z

Background: this is flaking in the canary as https://chromium-review.googlesource.com/c/v8/v8/+/7761796 made the function eligible for maglev tier-up. Locally this made the flake disappear for me. Opening against the main branch since it makes sense here anyway.

The depth of the stack depends not only on the stack size, but also on the size of each stack frame, which in turn depends on which tier the recursive function happens to be running at when the overflow occurs. Under load the background tier-up can land at a non-deterministic point in the recursion and flake the test. Keep the recursive function in the interpreter with %NeverOptimizeFunction() so the frame size - and thus the depth - is deterministic.

The depth of the stack depends not only on the stack size, but also on the size of each stack frame, which in turn depends on which tier the recursive function happens to be running at when the overflow occurs. Under load the background tier-up can land at a non-deterministic point in the recursion and flake the test. Keep the recursive function in the interpreter with %NeverOptimizeFunction() so the frame size - and thus the depth - is deterministic. Signed-off-by: Joyee Cheung <joyeec9h3@gmail.com>

nodejs-github-bot · 2026-07-06T20:51:44Z

CI: https://ci.nodejs.org/job/node-test-pull-request/74610/ 💚

nodejs-github-bot · 2026-07-07T11:53:30Z

Landed in 3257a36

The depth of the stack depends not only on the stack size, but also on the size of each stack frame, which in turn depends on which tier the recursive function happens to be running at when the overflow occurs. Under load the background tier-up can land at a non-deterministic point in the recursion and flake the test. Keep the recursive function in the interpreter with %NeverOptimizeFunction() so the frame size - and thus the depth - is deterministic. Signed-off-by: Joyee Cheung <joyeec9h3@gmail.com> PR-URL: #64271 Reviewed-By: Michaël Zasso <targos@protonmail.com> Reviewed-By: James M Snell <jasnell@gmail.com>

…ite (#53062) Backports nodejs/node#64271. The test compares the recursion depth two Workers reach before overflowing, but depth also depends on frame size, which changes when the recursive function tiers up mid-recursion. On a busy CI runner that lands non-deterministically and the depths end up >10% apart. Pin the function to the interpreter so the depth is stable. The fix is only in Node 26.6.0+, so it is carried as a patch until the Node roll picks it up.

…ite (#53075) Backports nodejs/node#64271. The test compares the recursion depth two Workers reach before overflowing, but depth also depends on frame size, which changes when the recursive function tiers up mid-recursion. On a busy CI runner that lands non-deterministically and the depths end up >10% apart. Pin the function to the interpreter so the depth is stable. The fix is only in Node 26.6.0+, so it is carried as a patch until the Node roll picks it up. Co-authored-by: trop[bot] <37223003+trop[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

…ite (#53076) Backports nodejs/node#64271. The test compares the recursion depth two Workers reach before overflowing, but depth also depends on frame size, which changes when the recursive function tiers up mid-recursion. On a busy CI runner that lands non-deterministically and the depths end up >10% apart. Pin the function to the interpreter so the depth is stable. The fix is only in Node 26.6.0+, so it is carried as a patch until the Node roll picks it up. Co-authored-by: trop[bot] <37223003+trop[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

…ite (#53077) Backports nodejs/node#64271. The test compares the recursion depth two Workers reach before overflowing, but depth also depends on frame size, which changes when the recursive function tiers up mid-recursion. On a busy CI runner that lands non-deterministically and the depths end up >10% apart. Pin the function to the interpreter so the depth is stable. The fix is only in Node 26.6.0+, so it is carried as a patch until the Node roll picks it up. Co-authored-by: trop[bot] <37223003+trop[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

…ite (#53078) Backports nodejs/node#64271. The test compares the recursion depth two Workers reach before overflowing, but depth also depends on frame size, which changes when the recursive function tiers up mid-recursion. On a busy CI runner that lands non-deterministically and the depths end up >10% apart. Pin the function to the interpreter so the depth is stable. The fix is only in Node 26.6.0+, so it is carried as a patch until the Node roll picks it up. Co-authored-by: trop[bot] <37223003+trop[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

src_fix_libuv_assertion_on_windows.patch (nodejs/node#61999) and test_remove_impact_of_tier-up_changes_in_worker_stack_size_test.patch (nodejs/node#64271) are both included in Node.js v24.20.0.

Both landed in v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271

Both patches are backports that landed in Node.js v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271

Both landed in v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271

Both patches are backports that landed in Node.js v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271

src_fix_libuv_assertion_on_windows.patch (nodejs/node#61999) and test_remove_impact_of_tier-up_changes_in_worker_stack_size_test.patch (nodejs/node#64271) are both included in Node.js v24.20.0.

* chore: bump node in DEPS to v24.20.0 * fix(patch): implement package maps in the ESM resolver Upstream added join to the path imports in resolve.js, which overlaps the import line this patch rewrites to pull in toNamespacedPath. Ref: nodejs/node#62239 * chore: remove upstreamed patches Both landed in v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271 * chore: update patches (trivial only) * fix(patch): preserve RSA-PSS legacy pubkey DER Upstream added ReadASN1Element to ncrypto.cc but only calls it from the OpenSSL 3 provider branch, so under BoringSSL it is an unused function and -Werror fails the build. Guard the definition with the same NCRYPTO_USE_OPENSSL3_PROVIDER condition as its callers. Ref: nodejs/node#64547 * fix(patch): update sqlite to 3.53.4 The sqlite update introduced an unsigned/signed comparison in rbuDeltaApply that trips -Wsign-compare under -Werror. Extend the existing sign-compare patch to cast lenDelta, which is never negative at that point, to unsigned int. Ref: nodejs/node#64745 * fix: do not pass removed JSPI V8 flag Node.js v24.20.0 passes --experimental-wasm-jspi unconditionally, but Chromium's V8 shipped JSPI and removed the flag, so node_mksnapshot aborts with 'bad option'. Add the same patch as the main-branch roll. Ref: nodejs/node#59941 Ref: https://chromium-review.googlesource.com/c/v8/v8/+/6937664 * chore: update filenames.auto.gni --------- Co-authored-by: electron-roller[bot] <84116207+electron-roller[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

* chore: bump node in DEPS to v24.20.0 * chore: remove upstreamed patches Both patches are backports that landed in Node.js v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271 * fix(patch): implement package maps in the ESM resolver Upstream rewrote packageResolve() to branch on hasPackageMap() and added join to the path imports, which shifted the context around the electron: specifier shortcut and the legacyMainResolve imports. Ref: nodejs/node#62239 * chore: update patches (trivial only) * fix(patch): split OpenSSL 3, BoringSSL, and legacy backends Upstream added an ASN.1 helper, ReadASN1Element, whose only callers live under NCRYPTO_USE_OPENSSL3_PROVIDER. With BoringSSL that macro is 0, so the helper is unused and fails -Werror=unused-function. Guard the definition with the same macro as its callers. Ref: nodejs/node#64211 * chore: update filenames.auto.gni Node.js v24.20.0 replaced src/quic/logstream.h with src/quic/session_manager.h. * fix(patch): update sqlite to 3.53.4 The new sqlite amalgamation compares 'unsigned int cnt' against 'int lenDelta' in rbuDeltaApply, which fails -Werror=sign-compare. Extend the existing sqlite sign-compare patch to cast that operand too. Ref: nodejs/node#64745 * fix(patch): do not pass removed JSPI V8 flag Node.js v24.20.0 backported JSPI enablement, which unconditionally passes --experimental-wasm-jspi to V8. Chromium's V8 removed that flag (JSPI is always on), so node_mksnapshot aborts with "bad option". Drop the argument in a new patch, matching main. Ref: nodejs/node#59941 Ref: https://chromium-review.googlesource.com/c/v8/v8/+/6937664 --------- Co-authored-by: electron-roller[bot] <84116207+electron-roller[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

joyeecheung · 2026-07-03T17:23:22Z

Background: this is flaking in the canary as https://chromium-review.googlesource.com/c/v8/v8/+/7761796 made the function eligible for maglev tier-up. Locally this made the flake disappear for me. Opening against the main branch since it makes sense here anyway.

The depth of the stack depends not only on the stack size, but also on the size of each stack frame, which in turn depends on which tier the recursive function happens to be running at when the overflow occurs. Under load the background tier-up can land at a non-deterministic point in the recursion and flake the test. Keep the recursive function in the interpreter with %NeverOptimizeFunction() so the frame size - and thus the depth - is deterministic.

The depth of the stack depends not only on the stack size, but also on the size of each stack frame, which in turn depends on which tier the recursive function happens to be running at when the overflow occurs. Under load the background tier-up can land at a non-deterministic point in the recursion and flake the test. Keep the recursive function in the interpreter with %NeverOptimizeFunction() so the frame size - and thus the depth - is deterministic. Signed-off-by: Joyee Cheung <joyeec9h3@gmail.com>

nodejs-github-bot · 2026-07-06T20:51:44Z

CI: https://ci.nodejs.org/job/node-test-pull-request/74610/ 💚

nodejs-github-bot · 2026-07-07T11:53:30Z

Landed in 3257a36

The depth of the stack depends not only on the stack size, but also on the size of each stack frame, which in turn depends on which tier the recursive function happens to be running at when the overflow occurs. Under load the background tier-up can land at a non-deterministic point in the recursion and flake the test. Keep the recursive function in the interpreter with %NeverOptimizeFunction() so the frame size - and thus the depth - is deterministic. Signed-off-by: Joyee Cheung <joyeec9h3@gmail.com> PR-URL: #64271 Reviewed-By: Michaël Zasso <targos@protonmail.com> Reviewed-By: James M Snell <jasnell@gmail.com>

…ite (#53062) Backports nodejs/node#64271. The test compares the recursion depth two Workers reach before overflowing, but depth also depends on frame size, which changes when the recursive function tiers up mid-recursion. On a busy CI runner that lands non-deterministically and the depths end up >10% apart. Pin the function to the interpreter so the depth is stable. The fix is only in Node 26.6.0+, so it is carried as a patch until the Node roll picks it up.

…ite (#53075) Backports nodejs/node#64271. The test compares the recursion depth two Workers reach before overflowing, but depth also depends on frame size, which changes when the recursive function tiers up mid-recursion. On a busy CI runner that lands non-deterministically and the depths end up >10% apart. Pin the function to the interpreter so the depth is stable. The fix is only in Node 26.6.0+, so it is carried as a patch until the Node roll picks it up. Co-authored-by: trop[bot] <37223003+trop[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

…ite (#53076) Backports nodejs/node#64271. The test compares the recursion depth two Workers reach before overflowing, but depth also depends on frame size, which changes when the recursive function tiers up mid-recursion. On a busy CI runner that lands non-deterministically and the depths end up >10% apart. Pin the function to the interpreter so the depth is stable. The fix is only in Node 26.6.0+, so it is carried as a patch until the Node roll picks it up. Co-authored-by: trop[bot] <37223003+trop[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

…ite (#53077) Backports nodejs/node#64271. The test compares the recursion depth two Workers reach before overflowing, but depth also depends on frame size, which changes when the recursive function tiers up mid-recursion. On a busy CI runner that lands non-deterministically and the depths end up >10% apart. Pin the function to the interpreter so the depth is stable. The fix is only in Node 26.6.0+, so it is carried as a patch until the Node roll picks it up. Co-authored-by: trop[bot] <37223003+trop[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

…ite (#53078) Backports nodejs/node#64271. The test compares the recursion depth two Workers reach before overflowing, but depth also depends on frame size, which changes when the recursive function tiers up mid-recursion. On a busy CI runner that lands non-deterministically and the depths end up >10% apart. Pin the function to the interpreter so the depth is stable. The fix is only in Node 26.6.0+, so it is carried as a patch until the Node roll picks it up. Co-authored-by: trop[bot] <37223003+trop[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

src_fix_libuv_assertion_on_windows.patch (nodejs/node#61999) and test_remove_impact_of_tier-up_changes_in_worker_stack_size_test.patch (nodejs/node#64271) are both included in Node.js v24.20.0.

Both landed in v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271

Both patches are backports that landed in Node.js v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271

Both landed in v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271

Both patches are backports that landed in Node.js v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271

src_fix_libuv_assertion_on_windows.patch (nodejs/node#61999) and test_remove_impact_of_tier-up_changes_in_worker_stack_size_test.patch (nodejs/node#64271) are both included in Node.js v24.20.0.

* chore: bump node in DEPS to v24.20.0 * fix(patch): implement package maps in the ESM resolver Upstream added join to the path imports in resolve.js, which overlaps the import line this patch rewrites to pull in toNamespacedPath. Ref: nodejs/node#62239 * chore: remove upstreamed patches Both landed in v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271 * chore: update patches (trivial only) * fix(patch): preserve RSA-PSS legacy pubkey DER Upstream added ReadASN1Element to ncrypto.cc but only calls it from the OpenSSL 3 provider branch, so under BoringSSL it is an unused function and -Werror fails the build. Guard the definition with the same NCRYPTO_USE_OPENSSL3_PROVIDER condition as its callers. Ref: nodejs/node#64547 * fix(patch): update sqlite to 3.53.4 The sqlite update introduced an unsigned/signed comparison in rbuDeltaApply that trips -Wsign-compare under -Werror. Extend the existing sign-compare patch to cast lenDelta, which is never negative at that point, to unsigned int. Ref: nodejs/node#64745 * fix: do not pass removed JSPI V8 flag Node.js v24.20.0 passes --experimental-wasm-jspi unconditionally, but Chromium's V8 shipped JSPI and removed the flag, so node_mksnapshot aborts with 'bad option'. Add the same patch as the main-branch roll. Ref: nodejs/node#59941 Ref: https://chromium-review.googlesource.com/c/v8/v8/+/6937664 * chore: update filenames.auto.gni --------- Co-authored-by: electron-roller[bot] <84116207+electron-roller[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

* chore: bump node in DEPS to v24.20.0 * chore: remove upstreamed patches Both patches are backports that landed in Node.js v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271 * fix(patch): implement package maps in the ESM resolver Upstream rewrote packageResolve() to branch on hasPackageMap() and added join to the path imports, which shifted the context around the electron: specifier shortcut and the legacyMainResolve imports. Ref: nodejs/node#62239 * chore: update patches (trivial only) * fix(patch): split OpenSSL 3, BoringSSL, and legacy backends Upstream added an ASN.1 helper, ReadASN1Element, whose only callers live under NCRYPTO_USE_OPENSSL3_PROVIDER. With BoringSSL that macro is 0, so the helper is unused and fails -Werror=unused-function. Guard the definition with the same macro as its callers. Ref: nodejs/node#64211 * chore: update filenames.auto.gni Node.js v24.20.0 replaced src/quic/logstream.h with src/quic/session_manager.h. * fix(patch): update sqlite to 3.53.4 The new sqlite amalgamation compares 'unsigned int cnt' against 'int lenDelta' in rbuDeltaApply, which fails -Werror=sign-compare. Extend the existing sqlite sign-compare patch to cast that operand too. Ref: nodejs/node#64745 * fix(patch): do not pass removed JSPI V8 flag Node.js v24.20.0 backported JSPI enablement, which unconditionally passes --experimental-wasm-jspi to V8. Chromium's V8 removed that flag (JSPI is always on), so node_mksnapshot aborts with "bad option". Drop the argument in a new patch, matching main. Ref: nodejs/node#59941 Ref: https://chromium-review.googlesource.com/c/v8/v8/+/6937664 --------- Co-authored-by: electron-roller[bot] <84116207+electron-roller[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

joyeecheung · 2026-07-03T17:23:22Z

Background: this is flaking in the canary as https://chromium-review.googlesource.com/c/v8/v8/+/7761796 made the function eligible for maglev tier-up. Locally this made the flake disappear for me. Opening against the main branch since it makes sense here anyway.

The depth of the stack depends not only on the stack size, but also on the size of each stack frame, which in turn depends on which tier the recursive function happens to be running at when the overflow occurs. Under load the background tier-up can land at a non-deterministic point in the recursion and flake the test. Keep the recursive function in the interpreter with %NeverOptimizeFunction() so the frame size - and thus the depth - is deterministic.

The depth of the stack depends not only on the stack size, but also on the size of each stack frame, which in turn depends on which tier the recursive function happens to be running at when the overflow occurs. Under load the background tier-up can land at a non-deterministic point in the recursion and flake the test. Keep the recursive function in the interpreter with %NeverOptimizeFunction() so the frame size - and thus the depth - is deterministic. Signed-off-by: Joyee Cheung <joyeec9h3@gmail.com>

nodejs-github-bot · 2026-07-06T20:51:44Z

CI: https://ci.nodejs.org/job/node-test-pull-request/74610/ 💚

nodejs-github-bot · 2026-07-07T11:53:30Z

Landed in 3257a36

The depth of the stack depends not only on the stack size, but also on the size of each stack frame, which in turn depends on which tier the recursive function happens to be running at when the overflow occurs. Under load the background tier-up can land at a non-deterministic point in the recursion and flake the test. Keep the recursive function in the interpreter with %NeverOptimizeFunction() so the frame size - and thus the depth - is deterministic. Signed-off-by: Joyee Cheung <joyeec9h3@gmail.com> PR-URL: #64271 Reviewed-By: Michaël Zasso <targos@protonmail.com> Reviewed-By: James M Snell <jasnell@gmail.com>

…ite (#53062) Backports nodejs/node#64271. The test compares the recursion depth two Workers reach before overflowing, but depth also depends on frame size, which changes when the recursive function tiers up mid-recursion. On a busy CI runner that lands non-deterministically and the depths end up >10% apart. Pin the function to the interpreter so the depth is stable. The fix is only in Node 26.6.0+, so it is carried as a patch until the Node roll picks it up.

…ite (#53075) Backports nodejs/node#64271. The test compares the recursion depth two Workers reach before overflowing, but depth also depends on frame size, which changes when the recursive function tiers up mid-recursion. On a busy CI runner that lands non-deterministically and the depths end up >10% apart. Pin the function to the interpreter so the depth is stable. The fix is only in Node 26.6.0+, so it is carried as a patch until the Node roll picks it up. Co-authored-by: trop[bot] <37223003+trop[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

…ite (#53076) Backports nodejs/node#64271. The test compares the recursion depth two Workers reach before overflowing, but depth also depends on frame size, which changes when the recursive function tiers up mid-recursion. On a busy CI runner that lands non-deterministically and the depths end up >10% apart. Pin the function to the interpreter so the depth is stable. The fix is only in Node 26.6.0+, so it is carried as a patch until the Node roll picks it up. Co-authored-by: trop[bot] <37223003+trop[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

…ite (#53077) Backports nodejs/node#64271. The test compares the recursion depth two Workers reach before overflowing, but depth also depends on frame size, which changes when the recursive function tiers up mid-recursion. On a busy CI runner that lands non-deterministically and the depths end up >10% apart. Pin the function to the interpreter so the depth is stable. The fix is only in Node 26.6.0+, so it is carried as a patch until the Node roll picks it up. Co-authored-by: trop[bot] <37223003+trop[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

…ite (#53078) Backports nodejs/node#64271. The test compares the recursion depth two Workers reach before overflowing, but depth also depends on frame size, which changes when the recursive function tiers up mid-recursion. On a busy CI runner that lands non-deterministically and the depths end up >10% apart. Pin the function to the interpreter so the depth is stable. The fix is only in Node 26.6.0+, so it is carried as a patch until the Node roll picks it up. Co-authored-by: trop[bot] <37223003+trop[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

src_fix_libuv_assertion_on_windows.patch (nodejs/node#61999) and test_remove_impact_of_tier-up_changes_in_worker_stack_size_test.patch (nodejs/node#64271) are both included in Node.js v24.20.0.

Both landed in v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271

Both patches are backports that landed in Node.js v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271

Both landed in v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271

Both patches are backports that landed in Node.js v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271

src_fix_libuv_assertion_on_windows.patch (nodejs/node#61999) and test_remove_impact_of_tier-up_changes_in_worker_stack_size_test.patch (nodejs/node#64271) are both included in Node.js v24.20.0.

* chore: bump node in DEPS to v24.20.0 * fix(patch): implement package maps in the ESM resolver Upstream added join to the path imports in resolve.js, which overlaps the import line this patch rewrites to pull in toNamespacedPath. Ref: nodejs/node#62239 * chore: remove upstreamed patches Both landed in v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271 * chore: update patches (trivial only) * fix(patch): preserve RSA-PSS legacy pubkey DER Upstream added ReadASN1Element to ncrypto.cc but only calls it from the OpenSSL 3 provider branch, so under BoringSSL it is an unused function and -Werror fails the build. Guard the definition with the same NCRYPTO_USE_OPENSSL3_PROVIDER condition as its callers. Ref: nodejs/node#64547 * fix(patch): update sqlite to 3.53.4 The sqlite update introduced an unsigned/signed comparison in rbuDeltaApply that trips -Wsign-compare under -Werror. Extend the existing sign-compare patch to cast lenDelta, which is never negative at that point, to unsigned int. Ref: nodejs/node#64745 * fix: do not pass removed JSPI V8 flag Node.js v24.20.0 passes --experimental-wasm-jspi unconditionally, but Chromium's V8 shipped JSPI and removed the flag, so node_mksnapshot aborts with 'bad option'. Add the same patch as the main-branch roll. Ref: nodejs/node#59941 Ref: https://chromium-review.googlesource.com/c/v8/v8/+/6937664 * chore: update filenames.auto.gni --------- Co-authored-by: electron-roller[bot] <84116207+electron-roller[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

* chore: bump node in DEPS to v24.20.0 * chore: remove upstreamed patches Both patches are backports that landed in Node.js v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271 * fix(patch): implement package maps in the ESM resolver Upstream rewrote packageResolve() to branch on hasPackageMap() and added join to the path imports, which shifted the context around the electron: specifier shortcut and the legacyMainResolve imports. Ref: nodejs/node#62239 * chore: update patches (trivial only) * fix(patch): split OpenSSL 3, BoringSSL, and legacy backends Upstream added an ASN.1 helper, ReadASN1Element, whose only callers live under NCRYPTO_USE_OPENSSL3_PROVIDER. With BoringSSL that macro is 0, so the helper is unused and fails -Werror=unused-function. Guard the definition with the same macro as its callers. Ref: nodejs/node#64211 * chore: update filenames.auto.gni Node.js v24.20.0 replaced src/quic/logstream.h with src/quic/session_manager.h. * fix(patch): update sqlite to 3.53.4 The new sqlite amalgamation compares 'unsigned int cnt' against 'int lenDelta' in rbuDeltaApply, which fails -Werror=sign-compare. Extend the existing sqlite sign-compare patch to cast that operand too. Ref: nodejs/node#64745 * fix(patch): do not pass removed JSPI V8 flag Node.js v24.20.0 backported JSPI enablement, which unconditionally passes --experimental-wasm-jspi to V8. Chromium's V8 removed that flag (JSPI is always on), so node_mksnapshot aborts with "bad option". Drop the argument in a new patch, matching main. Ref: nodejs/node#59941 Ref: https://chromium-review.googlesource.com/c/v8/v8/+/6937664 --------- Co-authored-by: electron-roller[bot] <84116207+electron-roller[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

joyeecheung · 2026-07-03T17:23:22Z

Background: this is flaking in the canary as https://chromium-review.googlesource.com/c/v8/v8/+/7761796 made the function eligible for maglev tier-up. Locally this made the flake disappear for me. Opening against the main branch since it makes sense here anyway.

The depth of the stack depends not only on the stack size, but also on the size of each stack frame, which in turn depends on which tier the recursive function happens to be running at when the overflow occurs. Under load the background tier-up can land at a non-deterministic point in the recursion and flake the test. Keep the recursive function in the interpreter with %NeverOptimizeFunction() so the frame size - and thus the depth - is deterministic.

The depth of the stack depends not only on the stack size, but also on the size of each stack frame, which in turn depends on which tier the recursive function happens to be running at when the overflow occurs. Under load the background tier-up can land at a non-deterministic point in the recursion and flake the test. Keep the recursive function in the interpreter with %NeverOptimizeFunction() so the frame size - and thus the depth - is deterministic. Signed-off-by: Joyee Cheung <joyeec9h3@gmail.com>

nodejs-github-bot · 2026-07-06T20:51:44Z

CI: https://ci.nodejs.org/job/node-test-pull-request/74610/ 💚

nodejs-github-bot · 2026-07-07T11:53:30Z

Landed in 3257a36

The depth of the stack depends not only on the stack size, but also on the size of each stack frame, which in turn depends on which tier the recursive function happens to be running at when the overflow occurs. Under load the background tier-up can land at a non-deterministic point in the recursion and flake the test. Keep the recursive function in the interpreter with %NeverOptimizeFunction() so the frame size - and thus the depth - is deterministic. Signed-off-by: Joyee Cheung <joyeec9h3@gmail.com> PR-URL: #64271 Reviewed-By: Michaël Zasso <targos@protonmail.com> Reviewed-By: James M Snell <jasnell@gmail.com>

…ite (#53062) Backports nodejs/node#64271. The test compares the recursion depth two Workers reach before overflowing, but depth also depends on frame size, which changes when the recursive function tiers up mid-recursion. On a busy CI runner that lands non-deterministically and the depths end up >10% apart. Pin the function to the interpreter so the depth is stable. The fix is only in Node 26.6.0+, so it is carried as a patch until the Node roll picks it up.

…ite (#53075) Backports nodejs/node#64271. The test compares the recursion depth two Workers reach before overflowing, but depth also depends on frame size, which changes when the recursive function tiers up mid-recursion. On a busy CI runner that lands non-deterministically and the depths end up >10% apart. Pin the function to the interpreter so the depth is stable. The fix is only in Node 26.6.0+, so it is carried as a patch until the Node roll picks it up. Co-authored-by: trop[bot] <37223003+trop[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

…ite (#53076) Backports nodejs/node#64271. The test compares the recursion depth two Workers reach before overflowing, but depth also depends on frame size, which changes when the recursive function tiers up mid-recursion. On a busy CI runner that lands non-deterministically and the depths end up >10% apart. Pin the function to the interpreter so the depth is stable. The fix is only in Node 26.6.0+, so it is carried as a patch until the Node roll picks it up. Co-authored-by: trop[bot] <37223003+trop[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

…ite (#53077) Backports nodejs/node#64271. The test compares the recursion depth two Workers reach before overflowing, but depth also depends on frame size, which changes when the recursive function tiers up mid-recursion. On a busy CI runner that lands non-deterministically and the depths end up >10% apart. Pin the function to the interpreter so the depth is stable. The fix is only in Node 26.6.0+, so it is carried as a patch until the Node roll picks it up. Co-authored-by: trop[bot] <37223003+trop[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

…ite (#53078) Backports nodejs/node#64271. The test compares the recursion depth two Workers reach before overflowing, but depth also depends on frame size, which changes when the recursive function tiers up mid-recursion. On a busy CI runner that lands non-deterministically and the depths end up >10% apart. Pin the function to the interpreter so the depth is stable. The fix is only in Node 26.6.0+, so it is carried as a patch until the Node roll picks it up. Co-authored-by: trop[bot] <37223003+trop[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

src_fix_libuv_assertion_on_windows.patch (nodejs/node#61999) and test_remove_impact_of_tier-up_changes_in_worker_stack_size_test.patch (nodejs/node#64271) are both included in Node.js v24.20.0.

Both landed in v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271

Both patches are backports that landed in Node.js v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271

Both landed in v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271

Both patches are backports that landed in Node.js v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271

src_fix_libuv_assertion_on_windows.patch (nodejs/node#61999) and test_remove_impact_of_tier-up_changes_in_worker_stack_size_test.patch (nodejs/node#64271) are both included in Node.js v24.20.0.

* chore: bump node in DEPS to v24.20.0 * fix(patch): implement package maps in the ESM resolver Upstream added join to the path imports in resolve.js, which overlaps the import line this patch rewrites to pull in toNamespacedPath. Ref: nodejs/node#62239 * chore: remove upstreamed patches Both landed in v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271 * chore: update patches (trivial only) * fix(patch): preserve RSA-PSS legacy pubkey DER Upstream added ReadASN1Element to ncrypto.cc but only calls it from the OpenSSL 3 provider branch, so under BoringSSL it is an unused function and -Werror fails the build. Guard the definition with the same NCRYPTO_USE_OPENSSL3_PROVIDER condition as its callers. Ref: nodejs/node#64547 * fix(patch): update sqlite to 3.53.4 The sqlite update introduced an unsigned/signed comparison in rbuDeltaApply that trips -Wsign-compare under -Werror. Extend the existing sign-compare patch to cast lenDelta, which is never negative at that point, to unsigned int. Ref: nodejs/node#64745 * fix: do not pass removed JSPI V8 flag Node.js v24.20.0 passes --experimental-wasm-jspi unconditionally, but Chromium's V8 shipped JSPI and removed the flag, so node_mksnapshot aborts with 'bad option'. Add the same patch as the main-branch roll. Ref: nodejs/node#59941 Ref: https://chromium-review.googlesource.com/c/v8/v8/+/6937664 * chore: update filenames.auto.gni --------- Co-authored-by: electron-roller[bot] <84116207+electron-roller[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

* chore: bump node in DEPS to v24.20.0 * chore: remove upstreamed patches Both patches are backports that landed in Node.js v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271 * fix(patch): implement package maps in the ESM resolver Upstream rewrote packageResolve() to branch on hasPackageMap() and added join to the path imports, which shifted the context around the electron: specifier shortcut and the legacyMainResolve imports. Ref: nodejs/node#62239 * chore: update patches (trivial only) * fix(patch): split OpenSSL 3, BoringSSL, and legacy backends Upstream added an ASN.1 helper, ReadASN1Element, whose only callers live under NCRYPTO_USE_OPENSSL3_PROVIDER. With BoringSSL that macro is 0, so the helper is unused and fails -Werror=unused-function. Guard the definition with the same macro as its callers. Ref: nodejs/node#64211 * chore: update filenames.auto.gni Node.js v24.20.0 replaced src/quic/logstream.h with src/quic/session_manager.h. * fix(patch): update sqlite to 3.53.4 The new sqlite amalgamation compares 'unsigned int cnt' against 'int lenDelta' in rbuDeltaApply, which fails -Werror=sign-compare. Extend the existing sqlite sign-compare patch to cast that operand too. Ref: nodejs/node#64745 * fix(patch): do not pass removed JSPI V8 flag Node.js v24.20.0 backported JSPI enablement, which unconditionally passes --experimental-wasm-jspi to V8. Chromium's V8 removed that flag (JSPI is always on), so node_mksnapshot aborts with "bad option". Drop the argument in a new patch, matching main. Ref: nodejs/node#59941 Ref: https://chromium-review.googlesource.com/c/v8/v8/+/6937664 --------- Co-authored-by: electron-roller[bot] <84116207+electron-roller[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

joyeecheung · 2026-07-03T17:23:22Z

Background: this is flaking in the canary as https://chromium-review.googlesource.com/c/v8/v8/+/7761796 made the function eligible for maglev tier-up. Locally this made the flake disappear for me. Opening against the main branch since it makes sense here anyway.

The depth of the stack depends not only on the stack size, but also on the size of each stack frame, which in turn depends on which tier the recursive function happens to be running at when the overflow occurs. Under load the background tier-up can land at a non-deterministic point in the recursion and flake the test. Keep the recursive function in the interpreter with %NeverOptimizeFunction() so the frame size - and thus the depth - is deterministic.

The depth of the stack depends not only on the stack size, but also on the size of each stack frame, which in turn depends on which tier the recursive function happens to be running at when the overflow occurs. Under load the background tier-up can land at a non-deterministic point in the recursion and flake the test. Keep the recursive function in the interpreter with %NeverOptimizeFunction() so the frame size - and thus the depth - is deterministic. Signed-off-by: Joyee Cheung <joyeec9h3@gmail.com>

nodejs-github-bot · 2026-07-06T20:51:44Z

CI: https://ci.nodejs.org/job/node-test-pull-request/74610/ 💚

nodejs-github-bot · 2026-07-07T11:53:30Z

Landed in 3257a36

The depth of the stack depends not only on the stack size, but also on the size of each stack frame, which in turn depends on which tier the recursive function happens to be running at when the overflow occurs. Under load the background tier-up can land at a non-deterministic point in the recursion and flake the test. Keep the recursive function in the interpreter with %NeverOptimizeFunction() so the frame size - and thus the depth - is deterministic. Signed-off-by: Joyee Cheung <joyeec9h3@gmail.com> PR-URL: #64271 Reviewed-By: Michaël Zasso <targos@protonmail.com> Reviewed-By: James M Snell <jasnell@gmail.com>

…ite (#53062) Backports nodejs/node#64271. The test compares the recursion depth two Workers reach before overflowing, but depth also depends on frame size, which changes when the recursive function tiers up mid-recursion. On a busy CI runner that lands non-deterministically and the depths end up >10% apart. Pin the function to the interpreter so the depth is stable. The fix is only in Node 26.6.0+, so it is carried as a patch until the Node roll picks it up.

…ite (#53075) Backports nodejs/node#64271. The test compares the recursion depth two Workers reach before overflowing, but depth also depends on frame size, which changes when the recursive function tiers up mid-recursion. On a busy CI runner that lands non-deterministically and the depths end up >10% apart. Pin the function to the interpreter so the depth is stable. The fix is only in Node 26.6.0+, so it is carried as a patch until the Node roll picks it up. Co-authored-by: trop[bot] <37223003+trop[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

…ite (#53076) Backports nodejs/node#64271. The test compares the recursion depth two Workers reach before overflowing, but depth also depends on frame size, which changes when the recursive function tiers up mid-recursion. On a busy CI runner that lands non-deterministically and the depths end up >10% apart. Pin the function to the interpreter so the depth is stable. The fix is only in Node 26.6.0+, so it is carried as a patch until the Node roll picks it up. Co-authored-by: trop[bot] <37223003+trop[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

…ite (#53077) Backports nodejs/node#64271. The test compares the recursion depth two Workers reach before overflowing, but depth also depends on frame size, which changes when the recursive function tiers up mid-recursion. On a busy CI runner that lands non-deterministically and the depths end up >10% apart. Pin the function to the interpreter so the depth is stable. The fix is only in Node 26.6.0+, so it is carried as a patch until the Node roll picks it up. Co-authored-by: trop[bot] <37223003+trop[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

…ite (#53078) Backports nodejs/node#64271. The test compares the recursion depth two Workers reach before overflowing, but depth also depends on frame size, which changes when the recursive function tiers up mid-recursion. On a busy CI runner that lands non-deterministically and the depths end up >10% apart. Pin the function to the interpreter so the depth is stable. The fix is only in Node 26.6.0+, so it is carried as a patch until the Node roll picks it up. Co-authored-by: trop[bot] <37223003+trop[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

src_fix_libuv_assertion_on_windows.patch (nodejs/node#61999) and test_remove_impact_of_tier-up_changes_in_worker_stack_size_test.patch (nodejs/node#64271) are both included in Node.js v24.20.0.

Both landed in v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271

Both patches are backports that landed in Node.js v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271

Both landed in v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271

Both patches are backports that landed in Node.js v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271

src_fix_libuv_assertion_on_windows.patch (nodejs/node#61999) and test_remove_impact_of_tier-up_changes_in_worker_stack_size_test.patch (nodejs/node#64271) are both included in Node.js v24.20.0.

* chore: bump node in DEPS to v24.20.0 * fix(patch): implement package maps in the ESM resolver Upstream added join to the path imports in resolve.js, which overlaps the import line this patch rewrites to pull in toNamespacedPath. Ref: nodejs/node#62239 * chore: remove upstreamed patches Both landed in v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271 * chore: update patches (trivial only) * fix(patch): preserve RSA-PSS legacy pubkey DER Upstream added ReadASN1Element to ncrypto.cc but only calls it from the OpenSSL 3 provider branch, so under BoringSSL it is an unused function and -Werror fails the build. Guard the definition with the same NCRYPTO_USE_OPENSSL3_PROVIDER condition as its callers. Ref: nodejs/node#64547 * fix(patch): update sqlite to 3.53.4 The sqlite update introduced an unsigned/signed comparison in rbuDeltaApply that trips -Wsign-compare under -Werror. Extend the existing sign-compare patch to cast lenDelta, which is never negative at that point, to unsigned int. Ref: nodejs/node#64745 * fix: do not pass removed JSPI V8 flag Node.js v24.20.0 passes --experimental-wasm-jspi unconditionally, but Chromium's V8 shipped JSPI and removed the flag, so node_mksnapshot aborts with 'bad option'. Add the same patch as the main-branch roll. Ref: nodejs/node#59941 Ref: https://chromium-review.googlesource.com/c/v8/v8/+/6937664 * chore: update filenames.auto.gni --------- Co-authored-by: electron-roller[bot] <84116207+electron-roller[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

* chore: bump node in DEPS to v24.20.0 * chore: remove upstreamed patches Both patches are backports that landed in Node.js v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271 * fix(patch): implement package maps in the ESM resolver Upstream rewrote packageResolve() to branch on hasPackageMap() and added join to the path imports, which shifted the context around the electron: specifier shortcut and the legacyMainResolve imports. Ref: nodejs/node#62239 * chore: update patches (trivial only) * fix(patch): split OpenSSL 3, BoringSSL, and legacy backends Upstream added an ASN.1 helper, ReadASN1Element, whose only callers live under NCRYPTO_USE_OPENSSL3_PROVIDER. With BoringSSL that macro is 0, so the helper is unused and fails -Werror=unused-function. Guard the definition with the same macro as its callers. Ref: nodejs/node#64211 * chore: update filenames.auto.gni Node.js v24.20.0 replaced src/quic/logstream.h with src/quic/session_manager.h. * fix(patch): update sqlite to 3.53.4 The new sqlite amalgamation compares 'unsigned int cnt' against 'int lenDelta' in rbuDeltaApply, which fails -Werror=sign-compare. Extend the existing sqlite sign-compare patch to cast that operand too. Ref: nodejs/node#64745 * fix(patch): do not pass removed JSPI V8 flag Node.js v24.20.0 backported JSPI enablement, which unconditionally passes --experimental-wasm-jspi to V8. Chromium's V8 removed that flag (JSPI is always on), so node_mksnapshot aborts with "bad option". Drop the argument in a new patch, matching main. Ref: nodejs/node#59941 Ref: https://chromium-review.googlesource.com/c/v8/v8/+/6937664 --------- Co-authored-by: electron-roller[bot] <84116207+electron-roller[bot]@users.noreply.github.com> Co-authored-by: Sam Attard <sattard@anthropic.com>

nodejs-github-bot added needs-ci PRs that need a full CI run. test Issues and PRs related to Node.js core tests and test infrastructure. labels Jul 3, 2026

targos approved these changes Jul 3, 2026
View reviewed changes

joyeecheung added the request-ci Add this label to start a Jenkins CI on a PR. label Jul 3, 2026

github-actionsBot removed the request-ci Add this label to start a Jenkins CI on a PR. label Jul 3, 2026

This comment was marked as outdated.
Sign in to view

jasnell approved these changes Jul 7, 2026
View reviewed changes

joyeecheung added the commit-queue PRs queued for automated landing through the Commit Queue. label Jul 7, 2026

nodejs-github-bot removed the commit-queue PRs queued for automated landing through the Commit Queue. label Jul 7, 2026

nodejs-github-bot merged commit 3257a36 into nodejs:mainJul 7, 2026
79 of 81 checks passed

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

MarshallOfSound mentioned this pull request Aug 21, 2026
test: deflake test-worker-stack-overflow-stack-size in the Node.js suite electron/electron#53062
Merged
1 task

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

MarshallOfSound added a commit to electron/electron that referenced this pull request Aug 29, 2026
chore: remove upstreamed patches
5bd25fa
Both landed in v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271

MarshallOfSound added a commit to electron/electron that referenced this pull request Aug 29, 2026
chore: remove upstreamed patches
8882847
Both patches are backports that landed in Node.js v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271

ckerr pushed a commit to electron/electron that referenced this pull request Sep 1, 2026
chore: remove upstreamed patches
fafa531
Both landed in v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271

ckerr pushed a commit to electron/electron that referenced this pull request Sep 1, 2026
chore: remove upstreamed patches
8cf4b56
Both patches are backports that landed in Node.js v24.20.0. Ref: nodejs/node#61999 Ref: nodejs/node#64271