ci(engines): widen engines.node to >=18 in CI to keep Node 18/20 jobs running#9145
Merged
Conversation
|
Contributor
Overall package sizeSelf size: 6.43 MB Dependency sizes| name | version | self size | total size | |------|---------|-----------|------------| | import-in-the-middle | 3.2.0 | 104.26 kB | 843.44 kB | | opentracing | 0.14.7 | 194.81 kB | 194.81 kB | | dc-polyfill | 0.1.11 | 25.74 kB | 25.74 kB |🤖 This report was automatically generated by heaviest-objects-in-the-universe |
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## master #9145 +/- ##
==========================================
- Coverage 93.73% 93.71% -0.03%
==========================================
Files 892 892
Lines 51236 51238 +2
Branches 11924 11923 -1
==========================================
- Hits 48025 48016 -9
- Misses 3211 3222 +11 Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Harness. 🚀 New features to boost your workflow:
|
rochdev
previously approved these changes
Jun 29, 2026
… running engines.node is bumped to >=22 to match the supported runtime range. The runtime guard (packages/dd-trace/src/guardrails/index.js) and the withVersions test helper both read engines.node and bail when the running major is below it, so on Node 18/20 every suite would silently skip (withVersions returns early, the init guard short-circuits, mocha reports 0 passing). CI widens engines.node back to >=18 at the action level (right after actions/setup-node) via scripts/ci/widen-engines-for-ci.js, so Node 18/20 jobs keep exercising real tests. The on-disk package.json that ships is unchanged (engines.node stays >=22). Co-authored-by: Cursor <cursoragent@cursor.com>
The widen step lives in the shared node/setup action, so it ran on every CI job, including integration-guardrails-unsupported, which installs Node 0.8 through 14 on purpose to verify the runtime guard aborts. Two problems: the script used a node:-prefixed require (unsupported before Node 14.18) so it crashed and failed those jobs, and even had it run, rewriting engines.node there would undermine a test that relies on the shipped >=22. Gate the step in the action so it only runs when the installed major is 18 or 20 (the supported majors the >=22 bump newly excludes). Every other version, including the unsupported matrix, keeps the shipped >=22 and the script is never parsed by a runtime too old to run it. Co-authored-by: Cursor <cursoragent@cursor.com>
…s.node Commit 1a79e10 (prepare to drop Node 18/20) changed the `oldest` version alias to read the minimum major out of package.json engines.node. With engines.node now bumped to >=22, `oldest` resolved to Node 22, which dropped Node 18 from the matrix and shifted the child_process job's legs to {22,20,24,26}. The resulting leg order (Node 22 first, then 20) destabilized the Bluebird global-Promise tests on the Node 20 leg, which pass on master's {18,20,24,26}. Pin `oldest` back to a literal 18 so CI keeps testing the oldest version we still exercise, decoupled from the shipped engines floor. This restores the matrix and leg order to match master; the widen step still restores >=18 on the 18/20 legs. Co-authored-by: Cursor <cursoragent@cursor.com>
pabloerhard
force-pushed
the
pabloerhard/rewrite-engines
branch
from
97be674 to
b13d145
Compare
BenchmarksBenchmark execution time: 2026-06-30 18:13:53 Comparing candidate commit 14ae973 in PR branch Found 0 performance improvements and 0 performance regressions! Performance is the same for 2253 metrics, 33 unstable metrics.
|
pabloerhard
added
the
only-land-on-next
pabloerhard
removed
the
only-land-on-next
rochdev
approved these changes
Jun 30, 2026
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 14ae9730b6
ℹ️ About Codex in GitHub
Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".
dd-octo-sts Bot
pushed a commit
that referenced
this pull request
Jul 1, 2026
… running (#9145) * ci(engines): widen engines.node to >=18 in CI to keep Node 18/20 jobs running engines.node is bumped to >=22 to match the supported runtime range. The runtime guard (packages/dd-trace/src/guardrails/index.js) and the withVersions test helper both read engines.node and bail when the running major is below it, so on Node 18/20 every suite would silently skip (withVersions returns early, the init guard short-circuits, mocha reports 0 passing). CI widens engines.node back to >=18 at the action level (right after actions/setup-node) via scripts/ci/widen-engines-for-ci.js, so Node 18/20 jobs keep exercising real tests. The on-disk package.json that ships is unchanged (engines.node stays >=22). Co-authored-by: Cursor <cursoragent@cursor.com> * ci(engines): only widen engines.node on Node 18/20 jobs The widen step lives in the shared node/setup action, so it ran on every CI job, including integration-guardrails-unsupported, which installs Node 0.8 through 14 on purpose to verify the runtime guard aborts. Two problems: the script used a node:-prefixed require (unsupported before Node 14.18) so it crashed and failed those jobs, and even had it run, rewriting engines.node there would undermine a test that relies on the shipped >=22. Gate the step in the action so it only runs when the installed major is 18 or 20 (the supported majors the >=22 bump newly excludes). Every other version, including the unsupported matrix, keeps the shipped >=22 and the script is never parsed by a runtime too old to run it. Co-authored-by: Cursor <cursoragent@cursor.com> * ci(node): pin oldest alias to Node 18 instead of deriving from engines.node Commit 1a79e10 (prepare to drop Node 18/20) changed the `oldest` version alias to read the minimum major out of package.json engines.node. With engines.node now bumped to >=22, `oldest` resolved to Node 22, which dropped Node 18 from the matrix and shifted the child_process job's legs to {22,20,24,26}. The resulting leg order (Node 22 first, then 20) destabilized the Bluebird global-Promise tests on the Node 20 leg, which pass on master's {18,20,24,26}. Pin `oldest` back to a literal 18 so CI keeps testing the oldest version we still exercise, decoupled from the shipped engines floor. This restores the matrix and leg order to match master; the widen step still restores >=18 on the 18/20 legs. Co-authored-by: Cursor <cursoragent@cursor.com> * add early return --------- Co-authored-by: Cursor <cursoragent@cursor.com>
rochdev
pushed a commit
that referenced
this pull request
Jul 2, 2026
… running (#9145) * ci(engines): widen engines.node to >=18 in CI to keep Node 18/20 jobs running engines.node is bumped to >=22 to match the supported runtime range. The runtime guard (packages/dd-trace/src/guardrails/index.js) and the withVersions test helper both read engines.node and bail when the running major is below it, so on Node 18/20 every suite would silently skip (withVersions returns early, the init guard short-circuits, mocha reports 0 passing). CI widens engines.node back to >=18 at the action level (right after actions/setup-node) via scripts/ci/widen-engines-for-ci.js, so Node 18/20 jobs keep exercising real tests. The on-disk package.json that ships is unchanged (engines.node stays >=22). Co-authored-by: Cursor <cursoragent@cursor.com> * ci(engines): only widen engines.node on Node 18/20 jobs The widen step lives in the shared node/setup action, so it ran on every CI job, including integration-guardrails-unsupported, which installs Node 0.8 through 14 on purpose to verify the runtime guard aborts. Two problems: the script used a node:-prefixed require (unsupported before Node 14.18) so it crashed and failed those jobs, and even had it run, rewriting engines.node there would undermine a test that relies on the shipped >=22. Gate the step in the action so it only runs when the installed major is 18 or 20 (the supported majors the >=22 bump newly excludes). Every other version, including the unsupported matrix, keeps the shipped >=22 and the script is never parsed by a runtime too old to run it. Co-authored-by: Cursor <cursoragent@cursor.com> * ci(node): pin oldest alias to Node 18 instead of deriving from engines.node Commit 1a79e10 (prepare to drop Node 18/20) changed the `oldest` version alias to read the minimum major out of package.json engines.node. With engines.node now bumped to >=22, `oldest` resolved to Node 22, which dropped Node 18 from the matrix and shifted the child_process job's legs to {22,20,24,26}. The resulting leg order (Node 22 first, then 20) destabilized the Bluebird global-Promise tests on the Node 20 leg, which pass on master's {18,20,24,26}. Pin `oldest` back to a literal 18 so CI keeps testing the oldest version we still exercise, decoupled from the shipped engines floor. This restores the matrix and leg order to match master; the widen step still restores >=18 on the 18/20 legs. Co-authored-by: Cursor <cursoragent@cursor.com> * add early return --------- Co-authored-by: Cursor <cursoragent@cursor.com>
rochdev
pushed a commit
that referenced
this pull request
Jul 2, 2026
… running (#9145) * ci(engines): widen engines.node to >=18 in CI to keep Node 18/20 jobs running engines.node is bumped to >=22 to match the supported runtime range. The runtime guard (packages/dd-trace/src/guardrails/index.js) and the withVersions test helper both read engines.node and bail when the running major is below it, so on Node 18/20 every suite would silently skip (withVersions returns early, the init guard short-circuits, mocha reports 0 passing). CI widens engines.node back to >=18 at the action level (right after actions/setup-node) via scripts/ci/widen-engines-for-ci.js, so Node 18/20 jobs keep exercising real tests. The on-disk package.json that ships is unchanged (engines.node stays >=22). Co-authored-by: Cursor <cursoragent@cursor.com> * ci(engines): only widen engines.node on Node 18/20 jobs The widen step lives in the shared node/setup action, so it ran on every CI job, including integration-guardrails-unsupported, which installs Node 0.8 through 14 on purpose to verify the runtime guard aborts. Two problems: the script used a node:-prefixed require (unsupported before Node 14.18) so it crashed and failed those jobs, and even had it run, rewriting engines.node there would undermine a test that relies on the shipped >=22. Gate the step in the action so it only runs when the installed major is 18 or 20 (the supported majors the >=22 bump newly excludes). Every other version, including the unsupported matrix, keeps the shipped >=22 and the script is never parsed by a runtime too old to run it. Co-authored-by: Cursor <cursoragent@cursor.com> * ci(node): pin oldest alias to Node 18 instead of deriving from engines.node Commit 1a79e10 (prepare to drop Node 18/20) changed the `oldest` version alias to read the minimum major out of package.json engines.node. With engines.node now bumped to >=22, `oldest` resolved to Node 22, which dropped Node 18 from the matrix and shifted the child_process job's legs to {22,20,24,26}. The resulting leg order (Node 22 first, then 20) destabilized the Bluebird global-Promise tests on the Node 20 leg, which pass on master's {18,20,24,26}. Pin `oldest` back to a literal 18 so CI keeps testing the oldest version we still exercise, decoupled from the shipped engines floor. This restores the matrix and leg order to match master; the widen step still restores >=18 on the 18/20 legs. Co-authored-by: Cursor <cursoragent@cursor.com> * add early return --------- Co-authored-by: Cursor <cursoragent@cursor.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What does this PR do?
Bumps the shipped
package.json#engines.nodefrom>=18to>=22, while keeping CI exercising real tests on Node 18 and 20 (rather than letting them silently skip). Concretely:package.json#engines.nodeto>=22(nodeMaxMajorunchanged at27). This is the value that ships.scripts/ci/widen-engines-for-ci.js, which readspackage.json, setsengines.nodeto>=18, writes it back, and logs the before/after..github/actions/node/setup/action.ymlas a new step right afteractions/setup-node. The step is gated to Node majors 18 and 20: on those legs it widensengines.nodeto>=18; on every other version it leaves the shipped>=22untouched and never even invokes the script.oldestversion alias in the same action back to a literal Node 18, instead of deriving the minimum major fromengines.node.Motivation
With
engines.nodebumped to>=22, the runtime guard inpackages/dd-trace/src/guardrails/index.jsand thewithVersionstest helper both readengines.nodeand bail when the running major is below it. On Node 18/20 that means every suite would silently skip:withVersionsreturns early, theinit.jsguard short-circuits, and mocha reports 0 passing — green CI that exercised nothing.Widening
engines.nodeto>=18at the action level keeps the Node 18/20 jobs exercising real tests. The widen step is the only new CI logic — there are no per-spec or per-helper workarounds. The on-diskpackage.jsonthat ships is unchanged:engines.nodestays>=22; the widening only happens in the CI checkout at runtime.Two follow-up fixes were needed so the rest of CI keeps working with the bumped floor:
Gate the widen step to Node 18/20. The step lives in the shared
node/setupaction, so it ran on every job — includingintegration-guardrails-unsupported, which installs ancient Node (0.8–14) on purpose to verify the guard aborts. The script used anode:-prefixed require (unsupported before Node 14.18) and crashed there; even without the crash, rewritingengines.nodeon those runtimes would undermine that job. Gating to majors 18/20 fixes both, and means the script is only ever parsed by a runtime new enough to run it. The unsupported matrix keeps seeing the shipped>=22.Pin the
oldestalias to Node 18. A prior commit (prepare infrastructure for dropping Node 18/20 in v6, ci: prepare infrastructure for dropping Node.js 18/20 in v6 #9033) had changed theoldestalias to derive the minimum major fromengines.node. Onceengines.nodebecame>=22,oldestresolved to Node 22, which dropped Node 18 from the matrix and shifted thechild_processjob's legs to{22, 20, 24, 26}. That leg order (Node 22 first, then 20) destabilized the Bluebird global-Promisetests on the Node 20 leg, which pass on master's{18, 20, 24, 26}. Pinningoldestback to a literal 18 restores the matrix and leg order to match master; the widen step still restores>=18on the 18/20 legs.Additional Notes
CI on the latest commit: the previously red
child_process,instrumentation-child_process, and allintegration-guardrails/integration-guardrails-unsupported/unit-guardrailsjobs are green. (ThePerformance and correctness testsfailure is unrelated to this change and is being handled separately.)Verified locally with engines widened to
>=18:./node_modules/.bin/mocha --timeout 30000 packages/dd-trace/test/plugins/with-versions.spec.js→ 2 passing./node_modules/.bin/mocha --timeout 60000 packages/dd-trace/test/profiling/exporters/agent.spec.js→ 7 passingNo production code (
packages/*/src/) is touched.Made with Cursor