You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
ci(e2e): start the managed-SSH lane for remote product changes no route names - #30210
Our end-to-end test for SSH hosts running a managed Orca server (the "managed-SSH Docker lane") only starts when a PR touches a file on a hand-written list. Two recent remote bugs (#30130, #30120) changed files that weren't on the list, so the lane never ran and the bugs merged. This adds a catch-all: any product change in the remote-host code now starts the lane.
What Changed
Before: a change to, for example, src/main/ipc/ssh.ts, src/cli/runtime/serve-orcad-launch.ts, src/main/orcad/orcad-entry.ts, src/shared/protocol-version.ts, the sidebar sleep flow or store/repos/repo-removal.ts selected no managed-SSH spec, so the lane was skipped. Six ssh-orcad-* specs had no route at all and ran only when the spec file itself changed.
After: a new route, ssh.orcad-remote-fallback, matches product (non-test) files under:
project removal in store/repos, plus the sidebar sleep and host-removal files.
It lists the six unrouted specs. The lane already runs every ssh-orcad-* spec once any of them is selected, so this route only decides whether the lane runs.
The lane stays advisory. It is not added to verify's needs (decision D1).
Breadth chosen, and its cost. Replayed over the last 600 commits on main, the lane would have started for:
10.3% of commits today;
21.0% with this route;
36.7% with the broader list in the round-2 proposal, which also adds runtime/rpc/, daemon/, renderer/src/runtime and store/terminals.
The lane costs 2 runners × about 17 minutes. Those broad folders mostly hold local-only code, so I left them out to roughly halve the added runner minutes. Measure the real minutes for a week before widening (decision D2).
Why
Hand-kept routes always miss the next new file. A fallback by folder is the only mechanism that covers files nobody has listed yet. A follow-up PR adds an import-graph contract for exact routes.
Linked Issue
Round 2 SSH hardening, slice S1 (internal). Motivated by #30130 and #30120.
Visual Proof
N/A: CI routing only.
Testing
New config/scripts/pr-e2e-remote-fallback-route.test.mjs:
every tests/e2e/ssh-orcad-*.spec.ts has a source route;
Included review availability: This review used your included allowance. Your plan provides up to 10 included reviews per hour; 3 remain after this review.
📝 Walkthrough
The change adds the ssh.orcad-remote-fallback route to select six managed-SSH E2E specs for matching product-source paths. It adds tests for route coverage, selected remote product-source paths, and unit-test and local-only paths that must not select the managed-SSH lane.
Priority: ⬇️ Low
Merge Risk:⚪ Minimal · up to 8e7ad
The new routing is ready to merge after normal checks; no actionable issue remains identified.
Pre-merge checks: 4 · 1
Failed checks: 1 warning
Check name
Status
Explanation
Resolution
Docstring Coverage
Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff. Analyzed 1 functions across 2 files.
Write docstrings for the functions missing them to satisfy the coverage threshold.
Passed checks: 4 passed
Check name
Status
Explanation
Title check
The title clearly identifies the main change: starting the managed-SSH E2E lane for remote product changes without existing routes. The wording is slightly awkward but remains specific and relevant.
Description check
The description is complete and aligned with the template. It explains the problem, scope, rationale, linked issues, testing, CI impact, and checklist status. It also documents the known baseline test…
Linked Issues check
Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check
Check skipped because no linked issues were found for this pull request.
Full details: Description check
Explanation
The description is complete and aligned with the template. It explains the problem, scope, rationale, linked issues, testing, CI impact, and checklist status. It also documents the known baseline test failures and the affected validation results.
Fix all pre-merge checks with AI
✨ Finishing Touches: 💡 1
📝 Generate docstrings: 💡
Commit to this branch
Create a new PR
Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.
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
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.
ELI5
Our end-to-end test for SSH hosts running a managed Orca server (the "managed-SSH Docker lane") only starts when a PR touches a file on a hand-written list. Two recent remote bugs (#30130, #30120) changed files that weren't on the list, so the lane never ran and the bugs merged. This adds a catch-all: any product change in the remote-host code now starts the lane.
What Changed
Before: a change to, for example,
src/main/ipc/ssh.ts,src/cli/runtime/serve-orcad-launch.ts,src/main/orcad/orcad-entry.ts,src/shared/protocol-version.ts, the sidebar sleep flow orstore/repos/repo-removal.tsselected no managed-SSH spec, so the lane was skipped. Sixssh-orcad-*specs had no route at all and ran only when the spec file itself changed.After: a new route,
ssh.orcad-remote-fallback, matches product (non-test) files under:src/main/{ssh,orcad,managed-server}/;src/main/ipc/{ssh*,runtime-environment*,orcad-*};src/cli/runtime/{serve*,launch-serve*};src/shared/{remote-runtime*,orcad-*,protocol-version.ts,runtime-capability*};store/repos, plus the sidebar sleep and host-removal files.It lists the six unrouted specs. The lane already runs every
ssh-orcad-*spec once any of them is selected, so this route only decides whether the lane runs.The lane stays advisory. It is not added to
verify's needs (decision D1).Breadth chosen, and its cost. Replayed over the last 600 commits on main, the lane would have started for:
runtime/rpc/,daemon/,renderer/src/runtimeandstore/terminals.The lane costs 2 runners × about 17 minutes. Those broad folders mostly hold local-only code, so I left them out to roughly halve the added runner minutes. Measure the real minutes for a week before widening (decision D2).
Why
Hand-kept routes always miss the next new file. A fallback by folder is the only mechanism that covers files nobody has listed yet. A follow-up PR adds an import-graph contract for exact routes.
Linked Issue
Round 2 SSH hardening, slice S1 (internal). Motivated by #30130 and #30120.
Visual Proof
N/A: CI routing only.
Testing
New
config/scripts/pr-e2e-remote-fallback-route.test.mjs:tests/e2e/ssh-orcad-*.spec.tshas a source route;Two of its three tests fail on main.
pr-e2e-route-contracts.test.mjsandci-e2e-job-selection.test.mjspass.The full static-analysis job ran locally.
Automated tests added/updated
Review
Codex review (gpt-6.1-sol, high): clean.
Agent skill upstream boundary
Notes
No product, wire or platform impact. CI cost: about 10% more PRs run the advisory Docker lane.
Checklist
N/Avisual proof with reason