test(installer): preregister P0-P9 greenfield RED
This commit is contained in:
+33
@@ -1368,3 +1368,36 @@ All work is **alpha** (< 0.1.0) until Jason approves 0.1.0 beta release.
|
||||
10. ASSUMPTION: **Conversations and messages get their own PG tables** (not stored in brain's entity model). They follow a chat-specific schema with proper foreign keys to users and projects. Rationale: Chat has different access patterns (streaming, pagination, search) than brain entities.
|
||||
|
||||
11. RESOLVED: **Pi handles all target LLM providers natively.** Anthropic, OpenAI/Codex, Z.ai, Ollama, LM Studio, and llama.cpp are all supported via Pi's built-in providers or `models.json` configuration with `openai-completions` API type. No custom provider adapters needed in @mosaicstack/agent — only configuration management.
|
||||
|
||||
---
|
||||
|
||||
## Greenfield install correctness — C1 (#1050)
|
||||
|
||||
### Problem and objective
|
||||
|
||||
A from-zero install can report success while leaving the target host unusable because the installer has no transactional state machine capable of certifying its own postconditions. C1 supplies the structural spine and red-first fixture; later cards repair the individual failed postconditions.
|
||||
|
||||
### Normative requirements
|
||||
|
||||
1. The installer SHALL implement the canonical P0–P9 numbering from the greenfield-install PRD v2: P0 Resolve context; P1 Preflight; P2 Acquire artifacts; P3 Install CLI; P4 Install framework + skills; P5 Identity; P6 Runtime linking / activation; P7 Services; P8 Shell discoverability; P9 Verify + commit.
|
||||
2. Every phase SHALL declare preconditions, action, committed postconditions, and rollback. An unverifiable postcondition SHALL fail the install non-zero with the named phase and a remediation line; no best-effort failure may still certify success. P1's required-tool closure includes tools invoked by later phases, including `git`; a downstream prerequisite may not remain undeclared and degrade silently.
|
||||
3. A durable mutation journal SHALL open before the first mutation and commit at P9. Fallible command output needed to diagnose a phase SHALL be journaled and surfaced, never discarded.
|
||||
4. `--check` SHALL run exactly the P0–P8 postcondition predicates without mutation, report each phase PASS/FAIL, and exit non-zero if any predicate fails.
|
||||
5. P4 SHALL consume a checkout-free, lane/versioned shipped-set declaration published by the installer. C1 SHALL NOT select among the currently disagreeing framework-payload, repository-root, sync-source, and W-jarvis populations; while no declaration exists, P4 reports `NOT-MEASURED / UNDECLARED` and remains blocking rather than fabricating a count. C5 owns the declaration's contents and containment/loadability fulfillment.
|
||||
6. The from-zero fixture SHALL be lane-parametric, use Debian/glibc, run the documented install command as a non-root target user with an isolated HOME, and inherit no host credentials, npm cache, home directory, or runtime configuration.
|
||||
7. The fixture SHALL select `next` with `--next` or `MOSAIC_NEXT=1` and assert the resolved lane version. Internal predicates use P3's absolute CLI path; shell discoverability is tested only at P8.
|
||||
8. Fault injection after each P2–P8 phase SHALL prove either clean rollback or a durable, honestly reported resumable partial state, with no journal incorrectly left in progress.
|
||||
9. Unsupported musl/Alpine and unavailable Docker SHALL fail loudly rather than skip as pass.
|
||||
|
||||
### C1 acceptance criteria
|
||||
|
||||
1. The pre-C1 from-zero matrix records both discriminating controls: with `git` absent, the legacy installer still exits zero while P1 fails and skill sync degrades; with `git` present, P1 passes and the observed sync store/runtime links are 101/101. The C1 installer must fail at P1 before mutation when `git` is absent.
|
||||
2. The discriminating P3 row passes: the binary exists at the expected absolute path and reports exactly the resolved `next` lane version, while P4, P5, and P8 fail.
|
||||
3. The `--check` mutation negative control proves host fingerprints are byte-identical before and after observation.
|
||||
4. Woodpecker executes and validates the expected RED fixture; C1 does not repair P4/P5/P8 or activate #869.
|
||||
|
||||
### Explicit exclusions and dependencies
|
||||
|
||||
- C2 owns P8/PATH, C3 owns P5/headless identity, C4 owns P6 activation policy, and C5 owns P4/skills.
|
||||
- Main-lane execution is a promotion precondition owned by #1037; C1 only makes the fixture lane-parametric.
|
||||
- RM-02 and #869 activation are out of scope.
|
||||
|
||||
Reference in New Issue
Block a user