fix(updater,dash): surface apply errors + auto-recover stale install dir#386
Merged
Conversation
The Settings → Updates "Install update" button silently no-op'd from
the user's POV: the apply endpoint returned 202 with a job id, the UI
showed a "started" toast and never polled, and the background job
crashed with UpdateExtractError ("install dir already exists and is
non-empty") because a prior failed attempt left /usr/lib/hal0/hal0-<v>/
populated. Three fixes:
1. ui/src/api/hooks/useUpdates.ts:
- useUpdateCheck now GETs /api/updates/check (was POSTing to a
GET-only route → 405).
- useUpdateApply takes an optional `version` (was misnamed
`channel`); the channel is implicit server-side.
- New useUpdateJob(jobId) polls /api/updates/status/{id} until
terminal so the UI can show "queued/installing…" inline and
toast on applied/failed.
2. ui/src/dash/settings.jsx:
- "Install update" calls applyM.mutate(undefined, …) with onError
toast, captures the returned job id, and shows progress + the
backend's verdict via useUpdateJob.
- lemonade "Check" calls checkM.mutate(undefined) with onError;
channel arg was meaningless.
3. src/hal0/updater/updater.py:
- _extract_tarball now quarantines a prior hal0 extraction at the
same path to <dest>.stale-<unix-ts> instead of refusing, so a
retry after a half-failed apply isn't permanently wedged.
Foreign non-empty dirs are still refused — we identify hal0
installs by VERSION file or pyproject.toml name="hal0".
- Removed the duplicate non-empty check in Updater.apply(); the
extract step is the single source of truth.
Tests: updated test_apply_refuses_when_install_dir_exists_nonempty to
require a foreign payload; added test_apply_quarantines_stale_hal0_install
asserting the recovery path. 34/34 updater + 41/41 touched API tests
green.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2 tasks
thinmintdev
added a commit
that referenced
this pull request
May 28, 2026
…#387) PR #386 fixed the silent Install update bug; this records it in the CHANGELOG (Unreleased → Fixed) and codifies the underlying pattern in PLAN §9. CHANGELOG: new Fixed section under Unreleased explaining the three fixes (UI hook verbs/signatures + job poller, backend extract quarantine, dedupe). PLAN §9 (Update mechanism): - Note the new extract-time quarantine behavior for prior hal0 extractions (no longer blocks retry after a half-failed apply). - New "Async-job API contract" subsection: any 202+job_id endpoint (updater apply, model pull, future jobs) requires the client to poll GET /status/{id} until terminal state and toast the verdict. A 202 ack alone is silently wrong — this was the entire root cause of #386. Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This was referenced May 28, 2026
thinmintdev
added a commit
that referenced
this pull request
May 29, 2026
End-of-stream cut for v0.3. Bundles MCP-completion, memory-map redesign, Settings → Updates fix (#386), silent-eviction dispatcher recovery (#392), ADR-0020 OpenRouter callback skeleton (#409), persona spending-cap primitive (#411), δ-harness Hermes coverage (#410), and the docs/internal pin + dashboard-v3 walkthrough (#389/#390). After this tag, active scope rolls to v0.4 (install-mode reconciliation + UI polish + fully-implemented Agents/UI/Install bootstrapped) and v0.5 (MCP admin + memory wiring across UI and agents). CHANGELOG merged from two coexisting Unreleased blocks into a single [v0.3.2-alpha.1] section; added missing entries for #392 (dispatcher), #387 (async-job polling contract), and the docs PRs #389/#390. pyproject 0.3.1-alpha.1 → 0.3.2-alpha.1. uv.lock resynced (was stuck at 0.3.0a1 from prior drift). Co-authored-by: Claude Opus 4.7 (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
1 participant
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.
Summary
/api/updates/status/{id}until terminal and toasts the backend's verdict; theuseUpdateCheckhook stops POSTing to a GET-only route./usr/lib/hal0/hal0-<v>/extract dirs from prior failed applies. Foreign non-empty dirs are still refused — recognises a hal0 install byVERSIONfile orpyproject.tomlname="hal0", otherwise moves aside to<dest>.stale-<unix-ts>so the retry succeeds.onErrortoasts on both Install update and lemonade Check.What was broken
End-to-end repro on the LXC:
Update available 0.3.1-alpha.1→ click Install update → success toast → nothing happens.Root cause stack:
POST /api/updates/applyreturns 202 with{id, state: "queued"}./status/{id}.UpdateExtractError: install dir already exists and is non-empty: /usr/lib/hal0/hal0-0.3.1-alpha.1from a previous half-failed attempt, transitions tostate: "failed"./api/updates/checkwhich is GET-only → silent 405.Changes
ui/src/api/hooks/useUpdates.ts: fix verb + arg shapes, adduseUpdateJob(jobId)poller (1.5s, tolerates transient errors during self-restart, stops onapplied/failed).ui/src/dash/settings.jsx: wire applyM withonSuccess(setJobId)+onError(toast), renderqueued…/installing…inline, fireapplied/failedtoasts once.src/hal0/updater/updater.py:_extract_tarballquarantines prior hal0 extractions instead of refusing; remove duplicate non-empty check inapply().tests/updater/test_updater.py: refit the "refuses non-empty" test to require a foreign payload; add quarantine-and-retry test.Test plan
pytest tests/updater/— 34/34pytest tests/api/test_updater_routes.py tests/api/test_typed_errors.py tests/api/test_stubs_return_envelope.py— 75/77 (2 skipped)ruff format --check+ruff checkclean on edited filestsc --noEmit+npm run buildgreen🤖 Generated with Claude Code