Skip to content

test: delete obsolete presentation and helper suites - #2478

Merged
Astro-Han merged 22 commits into
mainfrom
test/cleanup-obsolete-tests-2
Aug 8, 2026
Merged

Astro-Han merged 22 commits into
mainfrom
test/cleanup-obsolete-tests-2

Conversation

@Astro-Han

@Astro-Han Astro-Han commented Aug 8, 2026

Copy link
Copy Markdown
Contributor

Summary

Continue the aggressive cleanup from #2462 by deleting obsolete tests, orphaned prompt fixtures, and dead test-only APIs across Desktop, Core, Headless, UI, CLI, Computer Use, Runtime, and Runtime Host.

The branch changes 110 files with 13 additions and 8,343 deletions (net -8,330). Removed coverage is dominated by:

  • CSS, geometry, markup, localization, copy, animation, and terminal-display assertions
  • duplicated parser, renderer alias, provider-matrix, source-string, and fixture-wiring pins
  • obsolete Headless prompt experiments and two test-only prompt assets
  • small schema/default-value smokes already covered by higher-level contracts or workflows
  • 55 entire test files and 2 obsolete prompt assets, plus further partial-suite trimming

Focused validation found one runtime-loaded Harbor config that the static scan had misclassified as orphaned. That active asset was restored; only its source-string mirror test remains deleted.

Knip exposed and this PR also removes:

  • CLI _detectColorLevelForTesting
  • three unused Headless A/B fixture builders
  • unused Headless isTaskRunLocator

Why this is safe

Unique coverage alone was not treated as a retention reason. The retained suite still protects the active product contracts whose failure cost matches test maintenance: authorization, secrets, persistence, recovery, fail-closed protocols, dependency patches, and external-process lifecycle.

Explicitly retained guards

  • every dependency patch guard named by patches/README.md: AI SDK tool-call tracking, Astryx List accessible naming, grouped diff counts, new-message reset behavior, and blank UA-CH platform fallback
  • credential and secret redaction, including streamed output and persisted provider errors
  • permission escalation, authorization URLs, prompt-injection/authority boundaries, and sandbox takeover
  • persistence, deletion, recovery, storage-root authority, and session lifecycle gates
  • Linux sandbox capability enforcement used by CI
  • fail-closed protocol validation and lifecycle gates that prevent irreversible external actions, process leaks, or stale ownership

Verification

  • git diff --check origin/main...HEAD — passed
  • npm run format:check / npm run format — passed
  • npm run lint — passed
  • affected workspace typechecks for Desktop, Core, Headless, UI, CLI, Runtime, Computer Use, and Runtime Host — passed
  • npx knip compared with the clean root checkout — no new findings; four existing unused Headless exports are gone
  • focused Core tests — 7 passed
  • focused Headless tests — 89 passed, 1 intentionally skipped; Harbor adapter rerun after restoring its active config — 26 passed, 1 intentionally skipped
  • focused UI tests — 37 passed
  • focused CLI tests — 270 passed
  • focused Runtime tests — 19 passed
  • focused Computer Use tests — 100 passed
  • focused Runtime Host protocol tests — 46 passed
  • focused Desktop main tests — 58 passed in the latest pass
  • full repository tests and full Playwright E2E were intentionally not run locally; GitHub CI owns full coverage

CI

All triggered checks passed on the expanded third-pass diff:

  • typecheck
  • test_workspaces
  • test_headless
  • test_runtime_host
  • e2e_shard (1/2) and e2e_shard (2/2)
  • aggregate test and e2e gates
  • alignment_audit
  • windows_baseline (non-blocking)

storybook was correctly skipped because this PR does not change stories.

Review focus

Residual risk is intentionally concentrated in presentation regressions, exact helper formatting, provider-list drift, and small cross-platform/toolchain display behavior. Active dependency configs and all explicit security, authority, persistence, recovery, protocol, sandbox, and process-lifecycle guards remain.

@Astro-Han
Astro-Han marked this pull request as ready for review August 8, 2026 05:32
@Astro-Han
Astro-Han merged commit 2e6261d into main Aug 8, 2026
12 checks passed
@Astro-Han
Astro-Han deleted the test/cleanup-obsolete-tests-2 branch August 8, 2026 05:33
ggbdpq added a commit to ggbdpq/maka that referenced this pull request Sep 11, 2026
Four follow-ups from the review of the A10 audit:

- foundation-contract layer 3 now says what apache#4803 actually did: the
  broad route inventory was retired rather than relocated, nothing
  enumerates routes against a running Electron app, and layer 2's
  ax-tree-audit.mjs runs Storybook-only. The old text claimed
  equivalent coverage that apache#4803 explicitly declined to make, and the
  adjacent "Storybook 与 Electron E2E 共用" line said the same thing.
- ui-coverage: the three prose passages still describing the retired
  Electron spec (dynamic settings-navigation enumeration, named-dialog
  overlay sequencing, running-UI inventory) are corrected the same way,
  so the two documents no longer contradict each other.
- translation_status is source-only on all nine documents: none has a
  counterpart, and main's convention reserves synced for pairs.
- provenance §3 marks the binary-recovered cursor facts as removed by
  apache#3293 (hotspot is now the glyph tip); codex-pip notes that the
  pip-motion test it cited as exact coverage was deleted by apache#2478;
  provider-evidence says four evidence classes; the lease-class list
  matches computer-use.ts's four classes; the window-action lesson is
  annotated with its date.

Refs apache#3522

Generated-by: GLM-5.3-Flash (ZCode)
ggbdpq added a commit to ggbdpq/maka that referenced this pull request Sep 11, 2026
Four follow-ups from the review of the A10 audit:

- foundation-contract layer 3 now says what apache#4803 actually did: the
  broad route inventory was retired rather than relocated, nothing
  enumerates routes against a running Electron app, and layer 2's
  ax-tree-audit.mjs runs Storybook-only. The old text claimed
  equivalent coverage that apache#4803 explicitly declined to make, and the
  adjacent "Storybook 与 Electron E2E 共用" line said the same thing.
- ui-coverage: the three prose passages still describing the retired
  Electron spec (dynamic settings-navigation enumeration, named-dialog
  overlay sequencing, running-UI inventory) are corrected the same way,
  so the two documents no longer contradict each other.
- translation_status is source-only on all nine documents: none has a
  counterpart, and main's convention reserves synced for pairs.
- provenance §3 marks the binary-recovered cursor facts as removed by
  apache#3293 (hotspot is now the glyph tip); codex-pip notes that the
  pip-motion test it cited as exact coverage was deleted by apache#2478;
  provider-evidence says four evidence classes; the lease-class list
  matches computer-use.ts's four classes; the window-action lesson is
  annotated with its date.

Refs apache#3522

Generated-by: GLM-5.3-Flash (ZCode)
ggbdpq added a commit to ggbdpq/maka that referenced this pull request Sep 12, 2026
Review follow-up on apache#5172. The audit paragraph attributed the PiP window's
motion constants to the apache#3293 cursor replacement, but `pip-motion.ts` is
byte-identical before apache#3456, after it, and at this head — the replacement
rebuilt the agent cursor engine and glyphs, not the PiP springs or anchor
scoring. The paragraph now says the apache#2478 deletion only ends the retired
test's coverage, records that the motion module is untouched, and scopes
the cursor-provenance link to the agent cursor explicitly.

Generated-by: GLM-5.3-Flash (ZCode)
me2seeks pushed a commit that referenced this pull request Sep 12, 2026
* docs: audit the Computer Use documentation group (#3522)

Audit the nine Computer Use documents against current source:

- foundation-contract: the layer-3 verification description still
  named apps/desktop/e2e/accessibility-coverage.spec.ts, which #4803
  retired with the broad route inventories; the section now records
  that the coverage is carried by the ax-tree-audit rules and the
  Storybook smoke at their owning boundaries.
- The other eight documents verified clean: command names and lab
  fixture variables (evidence-classes), outcome types and the
  deterministic harness (host-events-contract), the maka_computer
  action surface and fail-closed coordinate policy (model-loop),
  provenance paths and bundled-tools pinning (provenance,
  cursor-provenance), report classes and sanitizer scripts
  (provider-evidence), and the story manifest (ui-coverage).
- All nine gain the standard frontmatter with last_verified dates.

Refs #3522

Generated-by: GLM-5.3-Flash (ZCode)

* docs: correct the Computer Use audit per review

Four follow-ups from the review of the A10 audit:

- foundation-contract layer 3 now says what #4803 actually did: the
  broad route inventory was retired rather than relocated, nothing
  enumerates routes against a running Electron app, and layer 2's
  ax-tree-audit.mjs runs Storybook-only. The old text claimed
  equivalent coverage that #4803 explicitly declined to make, and the
  adjacent "Storybook 与 Electron E2E 共用" line said the same thing.
- ui-coverage: the three prose passages still describing the retired
  Electron spec (dynamic settings-navigation enumeration, named-dialog
  overlay sequencing, running-UI inventory) are corrected the same way,
  so the two documents no longer contradict each other.
- translation_status is source-only on all nine documents: none has a
  counterpart, and main's convention reserves synced for pairs.
- provenance §3 marks the binary-recovered cursor facts as removed by
  #3293 (hotspot is now the glyph tip); codex-pip notes that the
  pip-motion test it cited as exact coverage was deleted by #2478;
  provider-evidence says four evidence classes; the lease-class list
  matches computer-use.ts's four classes; the window-action lesson is
  annotated with its date.

Refs #3522

Generated-by: GLM-5.3-Flash (ZCode)

* docs: separate PiP motion provenance from the cursor replacement

Review follow-up on #5172. The audit paragraph attributed the PiP window's
motion constants to the #3293 cursor replacement, but `pip-motion.ts` is
byte-identical before #3456, after it, and at this head — the replacement
rebuilt the agent cursor engine and glyphs, not the PiP springs or anchor
scoring. The paragraph now says the #2478 deletion only ends the retired
test's coverage, records that the motion module is untouched, and scopes
the cursor-provenance link to the agent cursor explicitly.

Generated-by: GLM-5.3-Flash (ZCode)
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant