Workplan to actually control the testrun

This commit is contained in:
2026-05-07 23:07:14 +02:00
parent cc43881565
commit cf02bae731

View File

@@ -0,0 +1,252 @@
---
id: OPEN-CMIS-TCK-WP-0002
type: extension-workplan
title: "Live OpenCMIS TCK Execution And Capability Maturity"
repo: open-cmis-tck
extension: open-cmis-tck
domain: markitect
status: active
owner: codex
planning_priority: high
planning_order: 3
created: "2026-05-07"
updated: "2026-05-07"
depends_on:
- "OPEN-CMIS-TCK-WP-0001"
state_hub_workstream_id: "da3f0d16-ba8e-4147-b0fc-ab3462e0b7b0"
---
# OPEN-CMIS-TCK-WP-0002: Live OpenCMIS TCK Execution And Capability Maturity
## Purpose
Move `open-cmis-tck` from a guide-board extension foundation to a locally usable
test facility that can actually run selected Apache Chemistry OpenCMIS TCK
checks against a CMIS Browser Binding system and generate evidence about what
works, what fails, what is unsupported by design, and what capability maturity
the target currently demonstrates.
The first delivery target is local execution on a developer workstation. Moving
the same setup to a dedicated test host or container image is deliberately
deferred until the local command path is proven.
## Background
WP-0001 established the extension boundary: guide-board can discover the
extension, preflight a CMIS Browser Binding endpoint, invoke a configured TCK
command, capture raw artifacts, normalize results, map requirement refs to CMIS
capability groups, and retain compact run summaries.
What remains is the real operational layer around the Apache Chemistry
OpenCMIS TCK:
- resolve the exact upstream TCK runtime and license/access posture,
- set up Java/Maven dependencies without vendoring restricted assets,
- map guide-board check groups to actual OpenCMIS TCK suites/classes/options,
- execute the TCK locally against a target repository,
- normalize real output, not only fixture output,
- produce a practical CMIS capability maturity scorecard.
## Target Architecture
```text
local workstation
-> Java/Maven/TCK dependency bootstrap
-> CMIS target profile + credential reference
-> guide-board plan
-> CMIS preflight
-> OpenCMIS TCK adapter command
-> raw TCK logs / reports / transcripts
-> normalized guide-board evidence
-> CMIS capability mapping
-> maturity scorecard
-> assessment package and Markdown report
```
## Boundary
This workplan prepares and runs local conformance checks. It does not claim
formal CMIS certification, redistribute restricted third-party artifacts,
replace OASIS/Apache authority material, modify the CMIS server under test, or
move execution to a container/test host yet.
## D2.1 - Resolve TCK Runtime And Access Model
```task
id: OPEN-CMIS-TCK-WP-0002-T001
status: todo
priority: high
state_hub_task_id: "f3144edb-9807-4fd6-ae29-b8db18939bc4"
```
Acceptance:
- Identify the current Apache Chemistry/OpenCMIS TCK source, Maven coordinates,
executable entry point, supported Java/Maven versions, and output formats.
- Document license/access posture and whether artifacts can be downloaded,
built from source, cached locally, or must be user-provided.
- Record exact version pins or version ranges for reproducible local runs.
- Add a local dependency manifest or lock note that guide-board can snapshot in
assessment evidence.
## D2.2 - Local Environment Bootstrap Command
```task
id: OPEN-CMIS-TCK-WP-0002-T002
status: todo
priority: high
state_hub_task_id: "f993c1ef-8e6f-4ad1-8375-4487887deb8b"
```
Acceptance:
- Provide a single local command or script that checks Java, Maven, network/cache
access, and OpenCMIS TCK runtime availability.
- The bootstrap does not commit downloaded artifacts or secrets to the repo.
- Missing prerequisites produce actionable diagnostics.
- Successful bootstrap writes a local runtime summary that can be referenced by
later runs.
## D2.3 - OpenCMIS TCK Adapter Invocation
```task
id: OPEN-CMIS-TCK-WP-0002-T003
status: todo
priority: high
state_hub_task_id: "a446a80f-fc63-4ea8-9720-9294db57ade9"
```
Acceptance:
- Implement an adapter command that can invoke selected real OpenCMIS TCK checks
from guide-board `check_group` values.
- The adapter accepts Browser Binding URL, repository ID, credentials reference
or anonymous mode, timeout, output directory, and selected check group.
- Raw stdout, stderr, TCK logs, machine-readable results, and invocation
metadata are captured under the run directory.
- The adapter can run a minimal repository/type group against a live target.
## D2.4 - Target Profiles And Credential References
```task
id: OPEN-CMIS-TCK-WP-0002-T004
status: todo
priority: high
state_hub_task_id: "c33a4d9a-a398-4a02-93c2-1d38e62a578f"
```
Acceptance:
- Provide local target profile templates for anonymous, basic-auth, and
externally referenced credential modes without storing secrets.
- Add clear examples for a `kontextual-engine` compatibility endpoint and a
generic CMIS Browser Binding endpoint.
- Profile diagnostics confirm endpoint URL, binding, repository ID, expected
capability groups, known gaps, and runtime timeout posture.
- Documentation explains how to run a profile validation before starting a TCK
run.
## D2.5 - Real Result Normalization
```task
id: OPEN-CMIS-TCK-WP-0002-T005
status: todo
priority: high
state_hub_task_id: "03ba9506-cce8-44fa-a036-2ab1ffc7a176"
```
Acceptance:
- Normalize actual OpenCMIS TCK output into guide-board evidence for each
selected check group.
- Preserve real TCK case IDs, class names, failure messages, skipped tests, and
raw artifact paths.
- Distinguish pass, fail, expected gap, unsupported by design, blocked, and
infrastructure error from real output.
- Add fixtures derived from real sanitized output so normalization remains
regression-tested.
## D2.6 - Live Pilot Run
```task
id: OPEN-CMIS-TCK-WP-0002-T006
status: todo
priority: high
state_hub_task_id: "d9eb9384-3352-4b71-9918-57282ee00411"
```
Acceptance:
- Run the local infrastructure against at least one reachable CMIS Browser
Binding target.
- Produce a guide-board run directory with evidence, findings, mappings,
assessment package, retention summary, and Markdown report.
- The report clearly states which CMIS capability groups passed, failed, were
blocked by infrastructure, or were unsupported by design.
- Known gaps do not hide unexpected failures in the same capability group.
## D2.7 - CMIS Capability Maturity Scorecard
```task
id: OPEN-CMIS-TCK-WP-0002-T007
status: todo
priority: high
state_hub_task_id: "7365052f-0d76-4fb2-b32c-236476b0f937"
```
Acceptance:
- Define a small maturity model for CMIS Browser Binding capability groups.
- Score repository/type, object/content, navigation, query, relationships,
ACL/policy, versioning, change log, and extension/known-gap posture.
- Separate conformance evidence from maturity interpretation.
- Generate a compact JSON scorecard and Markdown summary suitable for downstream
product-quality tracking.
## D2.8 - Operator Documentation
```task
id: OPEN-CMIS-TCK-WP-0002-T008
status: todo
priority: medium
state_hub_task_id: "da7022de-3d05-43bf-9c06-91f00c159bdc"
```
Acceptance:
- Document the local setup flow from clean checkout to first live TCK run.
- Include troubleshooting for Java/Maven, TCK artifact resolution, endpoint
reachability, repository ID mismatch, credentials, timeouts, and failed tests.
- Provide copy-pasteable commands for local execution through guide-board.
- Clearly mark the formal-certification boundary.
## D2.9 - Container/Test Host Handoff Notes
```task
id: OPEN-CMIS-TCK-WP-0002-T009
status: todo
priority: low
state_hub_task_id: "254d099a-5406-43b0-9497-9af594a9b911"
```
Acceptance:
- Capture what would need to move into a container or dedicated test host after
local execution is stable.
- Identify mounts for profiles, credentials, cached TCK assets, and run output.
- Identify environment variables and network assumptions.
- Do not implement the container/test-host move in this workplan unless local
execution is already proven.
## Definition Of Done
- A developer can run one documented local bootstrap command.
- A developer can point a profile at a CMIS Browser Binding endpoint and run at
least one real OpenCMIS TCK check group through guide-board.
- The run produces normalized evidence, raw artifacts, findings, mappings,
retention summary, assessment package, and Markdown report.
- The output explains what is working, what is failing, what is blocked by
infrastructure, and what is unsupported by design.
- A first CMIS capability maturity scorecard is generated from evidence.
- Container/test-host migration is documented as a follow-on path, not a hidden
prerequisite for local use.