Add node 18 checkout on github workflow - #2
Merged
Conversation
Signed-off-by: Paul Schultz <pschultz@pobox.com>
Signed-off-by: Paul Schultz <pschultz@pobox.com>
Signed-off-by: Paul Schultz <pschultz@pobox.com>
Signed-off-by: Paul Schultz <pschultz@pobox.com>
Signed-off-by: Paul Schultz <pschultz@pobox.com>
sanketpathak
pushed a commit
to sanketpathak/rhdh
that referenced
this pull request
Sep 30, 2025
…ce-update-wrapper e2e test update for marketplace wrapper pr
openshift-merge-bot Bot
pushed a commit
that referenced
this pull request
Nov 14, 2025
* fix(auto-approve): add checkout step to fix 'not a git repository' error The gh CLI commands were failing because the repository wasn't checked out. Added actions/checkout@v4 step to ensure git context is available. * test: temporarily enable workflow for polasudo to test functionality * test: add fix/auto-approve-bot to eligible patterns for testing * fix(auto-approve): add issues:write permission for label management * fix(auto-approve): use pull_request_target for proper permissions Using pull_request_target instead of pull_request allows the workflow to run with write permissions even for PRs from forks or bot accounts. This is necessary for adding labels and approving PRs. * Fix auto-approve workflow: use REST API for adding labels instead of gh CLI * test: trigger workflow re-run * test: remove trigger file * test: temporarily disable user condition to test workflow * fix(auto-approve): use pull_request_target and add issues:write permission - Change trigger from pull_request to pull_request_target for proper permissions - Add issues:write permission required for label management - Add --repo flag to gh commands for explicit repository context - Remove checkout step (not needed with pull_request_target) - Fixes GraphQL: Resource not accessible by integration error * fix(auto-approve): add fix/auto-approve-bot pattern for testing * feat(auto-approve): add push trigger for faster testing feedback - Add push trigger on fix/auto-approve-bot branch for immediate testing - Handle both push and pull_request_target events in workflow logic - Add logging step for push events to show eligibility check results - Only run label/approve actions on actual PR events * test: trigger workflow rerun * test: simplify push logging to debug workflow triggering * test: add temporary PR labeling simulation for fork testing * test: enable real PR labeling on push events for testing * test: target PR #1 in fork for labeling test * test: target PR #2 for labeling demonstration * feat: clean up auto-approve workflow for production use - Remove all testing code (push triggers, test PR numbers, debug logging) - Simplify workflow to only run on pull_request_target events - Only run for rhdh-bot PRs - Keep fix/auto-approve-bot pattern for final testing before merge - Ready for production deployment * test: temporarily allow polasudo for final workflow testing * feat: finalize auto-approve workflow for production - Remove temporary polasudo testing allowance - Remove fix/auto-approve-bot test pattern - Clean, production-ready workflow for rhdh-bot PRs - Ready to merge and deploy * test: trigger workflow to demonstrate PR #2 labeling * Delete test-workflow-final.md * chore: reorganize auto-approve workflow flags for better readability
nickboldt
pushed a commit
that referenced
this pull request
Nov 24, 2025
* fix(auto-approve): add checkout step to fix 'not a git repository' error The gh CLI commands were failing because the repository wasn't checked out. Added actions/checkout@v4 step to ensure git context is available. * test: temporarily enable workflow for polasudo to test functionality * test: add fix/auto-approve-bot to eligible patterns for testing * fix(auto-approve): add issues:write permission for label management * fix(auto-approve): use pull_request_target for proper permissions Using pull_request_target instead of pull_request allows the workflow to run with write permissions even for PRs from forks or bot accounts. This is necessary for adding labels and approving PRs. * Fix auto-approve workflow: use REST API for adding labels instead of gh CLI * test: trigger workflow re-run * test: remove trigger file * test: temporarily disable user condition to test workflow * fix(auto-approve): use pull_request_target and add issues:write permission - Change trigger from pull_request to pull_request_target for proper permissions - Add issues:write permission required for label management - Add --repo flag to gh commands for explicit repository context - Remove checkout step (not needed with pull_request_target) - Fixes GraphQL: Resource not accessible by integration error * fix(auto-approve): add fix/auto-approve-bot pattern for testing * feat(auto-approve): add push trigger for faster testing feedback - Add push trigger on fix/auto-approve-bot branch for immediate testing - Handle both push and pull_request_target events in workflow logic - Add logging step for push events to show eligibility check results - Only run label/approve actions on actual PR events * test: trigger workflow rerun * test: simplify push logging to debug workflow triggering * test: add temporary PR labeling simulation for fork testing * test: enable real PR labeling on push events for testing * test: target PR #1 in fork for labeling test * test: target PR #2 for labeling demonstration * feat: clean up auto-approve workflow for production use - Remove all testing code (push triggers, test PR numbers, debug logging) - Simplify workflow to only run on pull_request_target events - Only run for rhdh-bot PRs - Keep fix/auto-approve-bot pattern for final testing before merge - Ready for production deployment * test: temporarily allow polasudo for final workflow testing * feat: finalize auto-approve workflow for production - Remove temporary polasudo testing allowance - Remove fix/auto-approve-bot test pattern - Clean, production-ready workflow for rhdh-bot PRs - Ready to merge and deploy * test: trigger workflow to demonstrate PR #2 labeling * Delete test-workflow-final.md * chore: reorganize auto-approve workflow flags for better readability * fix(ci): add checkout step to auto-approve-bot-prs workflow to fix git repository error - Adds actions/checkout@v4 to provide git context for gh CLI commands - Fixes 'fatal: not a git repository' error in auto-approve workflow
Fortune-Ndlovu
pushed a commit
to Fortune-Ndlovu/rhdh
that referenced
this pull request
Dec 1, 2025
…eveloper#3717) * fix(auto-approve): add checkout step to fix 'not a git repository' error The gh CLI commands were failing because the repository wasn't checked out. Added actions/checkout@v4 step to ensure git context is available. * test: temporarily enable workflow for polasudo to test functionality * test: add fix/auto-approve-bot to eligible patterns for testing * fix(auto-approve): add issues:write permission for label management * fix(auto-approve): use pull_request_target for proper permissions Using pull_request_target instead of pull_request allows the workflow to run with write permissions even for PRs from forks or bot accounts. This is necessary for adding labels and approving PRs. * Fix auto-approve workflow: use REST API for adding labels instead of gh CLI * test: trigger workflow re-run * test: remove trigger file * test: temporarily disable user condition to test workflow * fix(auto-approve): use pull_request_target and add issues:write permission - Change trigger from pull_request to pull_request_target for proper permissions - Add issues:write permission required for label management - Add --repo flag to gh commands for explicit repository context - Remove checkout step (not needed with pull_request_target) - Fixes GraphQL: Resource not accessible by integration error * fix(auto-approve): add fix/auto-approve-bot pattern for testing * feat(auto-approve): add push trigger for faster testing feedback - Add push trigger on fix/auto-approve-bot branch for immediate testing - Handle both push and pull_request_target events in workflow logic - Add logging step for push events to show eligibility check results - Only run label/approve actions on actual PR events * test: trigger workflow rerun * test: simplify push logging to debug workflow triggering * test: add temporary PR labeling simulation for fork testing * test: enable real PR labeling on push events for testing * test: target PR redhat-developer#1 in fork for labeling test * test: target PR redhat-developer#2 for labeling demonstration * feat: clean up auto-approve workflow for production use - Remove all testing code (push triggers, test PR numbers, debug logging) - Simplify workflow to only run on pull_request_target events - Only run for rhdh-bot PRs - Keep fix/auto-approve-bot pattern for final testing before merge - Ready for production deployment * test: temporarily allow polasudo for final workflow testing * feat: finalize auto-approve workflow for production - Remove temporary polasudo testing allowance - Remove fix/auto-approve-bot test pattern - Clean, production-ready workflow for rhdh-bot PRs - Ready to merge and deploy * test: trigger workflow to demonstrate PR redhat-developer#2 labeling * Delete test-workflow-final.md * chore: reorganize auto-approve workflow flags for better readability * fix(ci): add checkout step to auto-approve-bot-prs workflow to fix git repository error - Adds actions/checkout@v4 to provide git context for gh CLI commands - Fixes 'fatal: not a git repository' error in auto-approve workflow Assisted-by: Cursor Signed-off-by: Fortune Ndlovu <fndlovu@redhat.com>
5 tasks
openshift-merge-bot Bot
pushed a commit
that referenced
this pull request
Apr 24, 2026
* feat(e2e-coverage): scaffold Playwright page.coverage collection (RHIDP-13243) Scaffolding for E2E frontend coverage on the rhdh repo (part of RHDHPLAN-851, Epic RHIDP-13242). Opt-in via COLLECT_COVERAGE=true; default behavior unchanged so existing E2E runs are unaffected. Infrastructure only - Extended test fixture at e2e-tests/playwright/support/coverage/test.ts that wraps page.coverage.startJSCoverage / stopJSCoverage around each test when COLLECT_COVERAGE=true. - Playwright reporter at e2e-tests/playwright/support/coverage/reporter.ts that aggregates raw V8 output into a merged Istanbul LCOV + HTML report via monocart-coverage-reports. - Reporter registered conditionally in playwright.config.ts so it is a no-op unless COLLECT_COVERAGE=true. - monocart-coverage-reports@2.12.11 added to e2e-tests devDependencies. - Docs at docs/coverage/e2e-rhdh.md covering usage, migration pattern, env vars, known limitations. - .gitignore tweaks so source code under paths containing "coverage" (docs/coverage, e2e-tests/playwright/support/coverage) is tracked while the generated "coverage/" output dirs remain ignored. No specs are migrated in this PR. Specs opt in via: import { test, expect } from "../support/coverage/test"; Follow-ups (separate PRs) - Migrate one spec (e.g. smoke-test) to validate the pipeline end-to-end. - Wire CI step to upload lcov.info to Codecov with flag rhdh-e2e-frontend. - Evaluate instrumented showcase image variant (RHIDP-13244 spike). Refs - https://issues.redhat.com/browse/RHIDP-13243 - https://issues.redhat.com/browse/RHIDP-13242 - https://issues.redhat.com/browse/RHDHPLAN-851 Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * chore(e2e-coverage): update yarn.lock for monocart-coverage-reports Previous commit added monocart-coverage-reports@2.12.11 to e2e-tests/package.json but did not update e2e-tests/yarn.lock. Yarn's hardened mode on public PRs forbids lockfile modifications during install, which blocks the CI lint/prettier/shellcheck step. Regenerated the lockfile locally with `yarn install`. Only new entries for monocart-coverage-reports and its transitive deps. Refs - #4680 - https://issues.redhat.com/browse/RHIDP-13243 Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(e2e-coverage): satisfy ESLint naming and no-unused-vars rules CI lint step rejected the new files under e2e-tests/playwright/support/coverage/ because of the project eslint-config naming rules. Fixes: - Rename top-level `const` variables to camelCase (COVERAGE_RAW_DIR/COVERAGE_REPORT_DIR/COVERAGE_OUTPUT_DIR → camelCase) - Drop unused `_config` and `_result` parameters on Reporter.onBegin / onEnd (the Reporter interface allows narrower implementations). - Replace destructuring `const { CoverageReport } = await import(...)` with `const monocart = await import(...)` + `new monocart.CoverageReport(...)` so the PascalCase class name does not trip the variable naming rule. - Keep the Playwright-idiomatic `export const test` and `export const expect` names, guarded by `eslint-disable-next-line @typescript-eslint/naming-convention` with a comment explaining the rationale (renaming would force every consumer to alias on import, degrading the DX). Verified locally: - `npx eslint playwright/support/coverage/test.ts playwright/support/coverage/reporter.ts` clean - `yarn prettier:check playwright/support/coverage/` clean - `yarn tsc:check` clean Refs - #4680 - https://issues.redhat.com/browse/RHIDP-13243 Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * refactor(e2e-coverage): address review feedback on scaffolding Harden the Playwright coverage scaffolding based on review notes so the first COLLECT_COVERAGE=true runs do not surface papercuts in CI. Correctness - Filename collision under parallel workers: raw V8 coverage files now include workerIndex and retry in the filename, and the title is taken from testInfo.titlePath (describe chain included) rather than just title. Two specs with identical titles in different describe blocks no longer collide. - Silent error swallow in reporter.readdir: the catch now only swallows ENOENT (directory missing on first run). Any other I/O error rethrows so CI logs show the real cause instead of the misleading "no coverage collected" warning. - JSON shape guard: entries loaded from raw files are verified to be arrays before handing to monocart. Corrupted or unexpected files are skipped with a clear warning instead of failing deep inside the library. Structure - Extracted the env-var-to-path resolution into a shared paths.ts, so the fixture and reporter can never drift on COVERAGE_OUTPUT_DIR / COVERAGE_REPORT_DIR. Exports follow the project UPPER_CASE convention for exported const. Hardening - Timeout wrapper (default 2 min, COVERAGE_GENERATE_TIMEOUT_MS override) around monocart.CoverageReport.generate(). A hung aggregation can no longer hold the Playwright run open in CI. - Parallel file reads via Promise.all. Scales better as the full 64-spec suite starts emitting coverage. Readability - playwright.config.ts uses `satisfies ReporterDescription[]` instead of the earlier `as [string]` cast — removes the type gymnastics while keeping the conditional registration. - Inline comments on startJSCoverage options explain the rationale for resetOnNavigation: false and reportAnonymousScripts: false. - Reworded the dynamic-import comment to match what it actually does (defers monocart load until generation time; the dep is installed regardless because it's in devDependencies). Docs - docs/coverage/e2e-rhdh.md updated with: - new env var COVERAGE_GENERATE_TIMEOUT_MS - note on the new filename scheme (workerIndex + retry) so the parallel collision question is explicit - import path examples for specs at depths 2, 3, and 4 (there is no tsconfig path alias configured) Local validation - npx eslint playwright/support/coverage/*.ts playwright.config.ts: clean - yarn tsc: clean - yarn prettier:check (coverage files + playwright.config.ts + docs): clean Refs - #4680 - https://issues.redhat.com/browse/RHIDP-13243 Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(e2e-coverage): address qodo review findings (bugs 2, 5, 7) Addresses actionable findings from the Qodo automated review. Three of the seven flagged items are real bugs that are fixed here; the other four are either false positives or scope decisions covered in the PR reply. Fixed - Stale raw files across runs (qodo #2). The reporter's onBegin now does `fs.rm COVERAGE_RAW_DIR` followed by `mkdir`, so the merged report only reflects the current Playwright run. Without this the raw dir accumulated *.json across runs and produced an incorrect LCOV. - Coverage teardown could fail tests (qodo #5). Both startJSCoverage and the file-write teardown are now wrapped in try/catch with a console.warn. Coverage collection is best-effort and can never fail a test run, including scenarios like unsupported browsers, pages closed before teardown, or transient I/O errors. - Custom-context specs silently skipped (qodo #7). Exported two helpers — startCoverageForPage(page) and stopCoverageForPage(page, testInfo) — so specs that manage their own BrowserContext/Page via browser.newContext() (e.g. plugins/adoption-insights, plugins/ scorecard) can opt in explicitly. Helpers are no-ops when COLLECT_COVERAGE is unset and share the same error-safe wrappers as the auto fixture. Not changed (with reason) - Fixture typing (qodo #1) — false positive. Reproduced locally: a migrated spec destructuring `{ page }` from the extended test compiles clean under `yarn tsc`. Kept `<NonNullable<unknown>>`. - Global fixture for every spec (qodo #3) — intentional scope. A per-spec import migration is phased so each batch lands in its own reviewable PR. Playwright does not offer a mechanism to override the test import globally without editing every spec file. - Codecov upload step (qodo #4) — intentional scope, depends on RHIDP-13230 (Codecov GitHub App + Vault token) landing first. The upload step is tracked for the follow-up CI wiring PR. - Source-map validation (qodo #6) — validation requires a running RHDH instance against which to run the instrumented suite. Planned as part of the first real COLLECT_COVERAGE=true run rather than a scaffolding PR. Docs updated - docs/coverage/e2e-rhdh.md now documents the custom-context pattern with a full example using startCoverageForPage / stopCoverageForPage, plus a "Specs that have not migrated" note making the phased migration explicit. Local validation - npx eslint playwright/support/coverage/*.ts: clean - yarn tsc: clean - yarn prettier:check (coverage + docs): clean Refs - #4680 - https://issues.redhat.com/browse/RHIDP-13243 Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(e2e-coverage): address PR review feedback on gitignore patterns Replace broad **/coverage ignore with scoped rules per zdrapela's review: - Root .gitignore: target dynamic-plugins/packages/plugins coverage dirs - e2e-tests/.gitignore: use /coverage/ instead of exclusion patterns - Add coverage to e2e-tests/.prettierignore and eslint ignores Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * docs(e2e-coverage): add AI assistant rule for instrumented imports Adds a rulesync rule that instructs AI assistants (Cursor, Claude Code, OpenCode) to use the coverage-instrumented test/expect imports instead of @playwright/test when creating or modifying spec files. Suggested-by: zdrapela Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * refactor(e2e-coverage): add @support path alias for coverage imports Add a TypeScript path alias @support/* → playwright/support/* in e2e-tests/tsconfig.json so specs can use a clean, depth-independent import instead of fragile relative paths. Update docs, code comments, and AI assistant rules accordingly. Suggested-by: zdrapela Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
nickboldt
pushed a commit
that referenced
this pull request
Apr 27, 2026
* feat(ci): parallelize base and RBAC helm deployments Run base_deployment and rbac_deployment concurrently in initiate_deployments() and initiate_deployments_osd_gcp(). Both target disjoint namespaces (NAME_SPACE vs NAME_SPACE_RBAC + NAME_SPACE_POSTGRES_DB), so there is no resource conflict. Overlapping the two helm upgrades and their internal readiness waits saves ~3-5 min of wall-clock time on CI. The _run_parallel_deployments helper captures each background job's exit code via `wait "$pid" || rc=$?` and returns 1 if either fails, so failures always propagate and are individually reported in the logs. Tests continue to run sequentially downstream (parallel Playwright would OOM given CI memory limits). Resolves: RHIDP-12296 * refactor(ci): add defensive checks and improved logging to parallel deployments Address code review findings from PR #4679: 1. **Namespace validation**: Add defensive check to ensure NAME_SPACE and NAME_SPACE_RBAC are different and non-empty before starting parallel deployments. Prevents subtle race conditions if misconfigured. 2. **Improved logging**: - Log both namespace names at the start for debugging - Add synchronization point logs ("Waiting for parallel deployments..." and "Parallel deployments finished") to make log flow clearer when output from both background jobs is interleaved 3. **Documentation**: Add "Requires" section to function header documenting the namespace disjointness requirement These changes improve debuggability and fail-fast behavior without changing the core parallelization logic. * fix(ci): eliminate race conditions for parallel deployments Address 3 critical race conditions identified by Qodo code review: **1. Shared temp files in helm::merge_values (Bug #1)** - BEFORE: Used fixed paths /tmp/step-without-plugins.yaml and /tmp/step-only-plugins.yaml - AFTER: Generate unique temp files per invocation via mktemp - WHY: Concurrent helm::merge_values calls would overwrite each other's intermediate files, corrupting merged values **2. Kubeconfig global state mutation (Bug #2)** - BEFORE: namespace::configure and apply_yaml_files both ran `oc config set-context --current --namespace=...` - AFTER: Removed all oc config set-context calls; added explicit --namespace flag to all oc/kubectl apply commands that lacked it - WHY: Parallel deployments racing on current context would apply resources to nondeterministic namespaces **3. In-place YAML file edits (Bug #3)** - BEFORE: apply_yaml_files used sed -i to modify shared YAML templates directly in the repo - AFTER: Copy templates to per-deployment tmpdir, patch copies, apply from tmpdir - WHY: Concurrent sed -i edits would clobber each other and apply manifests with wrong namespace fields All three bugs would have caused nondeterministic deployment failures when base_deployment and rbac_deployment run in parallel. Changes: - .ci/pipelines/lib/helm.sh: mktemp + trap cleanup for intermediate files - .ci/pipelines/lib/namespace.sh: Remove oc config set-context - .ci/pipelines/utils.sh: Tmpdir for YAML patches, explicit --namespace on all oc apply calls * fix(ci): address code review findings for parallel deployments 1. postgres-cred.yaml in-place edit race: configure_external_postgres_db now patches a temp copy instead of editing the shared template via sed_inplace. Same class of bug fixed in apply_yaml_files. 2. Unquoted ${NAME_SPACE} in namespace::configure calls: quote both occurrences (lines 553, 721) to prevent word-splitting. 3. mktemp in helm::merge_values overwrite path: move mktemp + trap inside the "merge" branch so the "overwrite" path doesn't create and immediately discard two unused temp files. 4. Remove redundant "Waiting for parallel deployments" log::section that fires right before the blocking wait calls — the bookend sections are sufficient. * fix(ci): replace trap RETURN with explicit cleanup trap RETURN leaks to the calling function in bash < 4.4, causing unbound variable errors when step_1_file/tmpdir are expanded in the caller's scope under set -o nounset. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
openshift-merge-bot Bot
pushed a commit
that referenced
this pull request
Jul 20, 2026
* docs(e2e): add E2E layer migration matrix (RHIDP-15076) Phase-1 deliverable of RHIDP-15076 (E2E Test Optimization epic RHIDP-13501): classifies all 30 e2e specs by target layer (L1-L4b), maps which are supplementable by Layer 3 component tests or the cluster-free L4a harness, and includes the companion analysis of the rhdh-plugin-export-overlays test system. Referenced from the Jira story's closing comment; updated 2026-07-02 with the L4a harness validation results from PR #5005 (4 tests green cluster-free). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * docs(e2e): address review — reconcile counts, durable references, README link - Fix the Tally L3 count (10, not 9) so the column totals sum to the 30-spec heading; reconcile the heading itself (29 on main + #19 pending in PR #4967). - "runs 4 tests" -> "runs 2 specs (4 test cases)" — the doc's accounting unit is the spec; note that spec numbers refer to the matrix below. - Soften "fully covers the 12 pure-backend workspaces" to load + API surface: scaffolder-backend-module-kubernetes also has a UI e2e that needs the render harness, so "fully" overstated the native-smoke scope. - Replace short commit hashes and the fork-only RHIDP-13235-layer3-component-tests branch name with the durable PR reference (rhdh#4864, closed) — hashes on a mutable branch dangle after a rebase or branch deletion. - Give DRAFT a promotion condition (groomed into RHIDP-13528/13529). - Link the matrix from docs/e2e-tests/README.md ("Adding a Test") so the doc is discoverable outside the Jira comment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * docs(e2e): correct licensed-users-info blocker — internal backend plugin, not OCI * docs(e2e): address review — fix spec count, dangling reference, dated wording - 2026-07-07 update: 10 specs (not 9) — the list itself names ten files. - Drop the positional "queue above" reference (the queue list no longer precedes it) and past-tense the 2026-07-02 "now runs" so the two dated updates stop contradicting each other. - Self-date the PR #4967 "open" claim and add PR #5057 to References. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * docs(e2e): record the merged overlays native smoke harness (#2714/#2731) The companion analysis recommended rebuilding PR #2231's in-process idea on the published CLI — that landed in the overlay repo on 2026-07-06/07 as smoke-tests-native (overlays#2714, ~20x vs the Docker smoke) plus workspace mode (overlays#2731), running as a dedicated native-smoke.yaml workflow alongside the Docker one. Mark recommendation 1 as landed and add both PRs to References. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * docs(e2e): address review — reconcile landed-vs-target claims and stale count - "Every remaining candidate landed" contradicted the #2 bullet in the same section — now "every candidate except #2". - The native-harness recommendation read as if the full Docker-smoke replacement had shipped; scope it to what overlays#2714 landed (backend boots), with catalog-extending modules and the frontend bundle probe still pending (same qualifier in the bucket table). - ROI paragraph still said 9 L3 candidates after the tally fix made it 10. - Date-mark the pre-#2714 two-tier takeaway so it points at the update. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * docs(e2e): correct #4864 status and refresh matrix against main Re-verified the document against origin/main (2026-07-20): - rhdh#4864 was **merged** on 2026-06-01, not "closed, not merged". The six Layer 3 tests are on main under packages/app/src — replace the prose descriptions with the actual file paths so the reference stays durable. - Batch 1 no longer asks to "close out" #5/#6/#8/#9; that work landed. - #18 github-happy-path is no longer a showcase `.fixme`; #5022 parked it at e2e-tests/playwright/blocked/github-happy-path.blocked.ts. - Spell out how the "29 specs on main" count is composed (28 *.spec.ts + the blocked one) so it can be re-checked mechanically. - Refresh dates; #4967 is still open as of 2026-07-20. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.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 / why we need it
Which issue(s) does this PR fix
Fixes #?
PR acceptance criteria
How to test changes / Special notes to the reviewer