DGR-024: Implement real generated-gRPC protocol harness #8

New Issue

popov · 2026-07-16T20:36:01Z

popov commented

2026-07-16 20:36:01 +00:00

DGR-024: Implement real generated-gRPC protocol harness

Status / triage: completed; passes: true
Execution mode: AFK
Milestone: M1
Dependencies: DGR-022, DGR-023
Blocks (derived): DGR-033, DGR-042
Labels: area:protocol, area:testing, type:vertical-slice, priority:p0, ready-for-agent
Evidence class: model-free
Hardware: none
Model: none
Upstream: no

Objective / description

Build a real generated-gRPC protocol harness around the versioned shard_runtime.proto contract. Use generated Python and C++ stubs over an actual localhost transport and real process lifecycle; exercise captured deterministic protocol vectors and serialized protobuf bytes before a real model worker exists. Do not implement an in-memory fake transport, synthetic model outputs, or a production-looking stub/demo service.

Acceptance criteria

Start a real localhost gRPC server process using generated bindings and connect to it with a generated client; no in-memory fake channel or direct method-only seam.
Exercise prefill fragments, decode frames, release, cancel, flow-control, deadlines, malformed input, checksum failure, duplicates, and stale epochs using serialized protocol messages and captured deterministic vectors.
Prove direct and opaque-relay paths preserve identical protobuf bytes by recording and comparing actual wire frames at both boundaries.
Use real process/socket lifecycle and fail closed on transport, schema, epoch, size, cache, and deadline violations; do not claim model or accelerator behavior that is not exercised.
Applicable shared quality gates pass, and evidence records exact commands, raw outputs, generated artifact identities, wire-frame hashes, changed files, limitations, and dependency handoff.

Shared quality gates

Targeted deterministic tests pass; Python changes also pass python -m compileall packages tests.
git diff --check passes.
Default tests are model-download-free, API-credit-free, and GPU-free.
Evidence README records exact changed files, commands/results, limitations, and dependency handoff; no fabricated evidence or inherited completion credit.

Evidence handoff

Verified evidence: .scratch/distributed-gguf-runtime/evidence/DGR-024/README.md. Legacy evidence remains provenance only and grants no implementation completion credit.

# DGR-024: Implement real generated-gRPC protocol harness - **Status / triage:** completed; `passes: true` - **Execution mode:** `AFK` - **Milestone:** `M1` - **Dependencies:** `DGR-022`, `DGR-023` - **Blocks (derived):** `DGR-033`, `DGR-042` - **Labels:** `area:protocol`, `area:testing`, `type:vertical-slice`, `priority:p0`, `ready-for-agent` - **Evidence class:** `model-free` - **Hardware:** `none` - **Model:** `none` - **Upstream:** `no` ## Objective / description Build a real generated-gRPC protocol harness around the versioned shard_runtime.proto contract. Use generated Python and C++ stubs over an actual localhost transport and real process lifecycle; exercise captured deterministic protocol vectors and serialized protobuf bytes before a real model worker exists. Do not implement an in-memory fake transport, synthetic model outputs, or a production-looking stub/demo service. ## Acceptance criteria - [x] Start a real localhost gRPC server process using generated bindings and connect to it with a generated client; no in-memory fake channel or direct method-only seam. - [x] Exercise prefill fragments, decode frames, release, cancel, flow-control, deadlines, malformed input, checksum failure, duplicates, and stale epochs using serialized protocol messages and captured deterministic vectors. - [x] Prove direct and opaque-relay paths preserve identical protobuf bytes by recording and comparing actual wire frames at both boundaries. - [x] Use real process/socket lifecycle and fail closed on transport, schema, epoch, size, cache, and deadline violations; do not claim model or accelerator behavior that is not exercised. - [x] Applicable shared quality gates pass, and evidence records exact commands, raw outputs, generated artifact identities, wire-frame hashes, changed files, limitations, and dependency handoff. ## Shared quality gates - Targeted deterministic tests pass; Python changes also pass `python -m compileall packages tests`. - `git diff --check` passes. - Default tests are model-download-free, API-credit-free, and GPU-free. - Evidence README records exact changed files, commands/results, limitations, and dependency handoff; no fabricated evidence or inherited completion credit. ## Evidence handoff Verified evidence: `.scratch/distributed-gguf-runtime/evidence/DGR-024/README.md`. Legacy evidence remains provenance only and grants no implementation completion credit.

popov added this to the M1 milestone 2026-07-16 20:36:01 +00:00

popov added the area:testing type:vertical-slice area:protocol status:blocked ready-for-agent priority:p0 labels 2026-07-16 20:36:01 +00:00

deploy added status:ready and removed status:blocked labels 2026-07-17 20:44:11 +00:00

deploy commented

2026-07-17 20:44:11 +00:00

Dependency reconciliation: DGR-022 and DGR-023 are now closed/completed. Moving to status:ready.

Dependency reconciliation: DGR-022 and DGR-023 are now closed/completed. Moving to `status:ready`.

deploy added status:in-progress and removed status:ready ready-for-agent labels 2026-07-17 20:51:33 +00:00

deploy commented

2026-07-17 20:51:33 +00:00

Claimed by the GPT-5.6 Terra protocol/testing lane after verified completion of DGR-023. One isolated Ralph iteration; controller review required before integration or closure.

popov commented

2026-07-17 22:05:56 +00:00

Reclaimed after restart for independent-review repair. P1s: stale epoch isolation, prefill/decode flow-control plus in-band deadlines, bidirectional activation responses, and generated-stub-compatible in-memory gRPC lifecycle. Existing partial implementation is preserved; no completion until fresh review.

popov added status:ready and removed status:in-progress labels 2026-07-18 13:00:33 +00:00

popov commented

2026-07-18 13:00:34 +00:00

Controller tick ended with no worker process active. The fixed Terra worktree preserves the DGR-024 implementation; 76 focused/shared tests passed with 2 required skips, PRD validation/compileall/ruff/diff-check passed, and final independent no-P0/P1 review is still pending. Claim released to status:ready; no work was reset or discarded.

Controller tick ended with no worker process active. The fixed Terra worktree preserves the DGR-024 implementation; 76 focused/shared tests passed with 2 required skips, PRD validation/compileall/ruff/diff-check passed, and final independent no-P0/P1 review is still pending. Claim released to `status:ready`; no work was reset or discarded.

popov added status:in-progress and removed status:ready labels 2026-07-18 13:19:28 +00:00

popov changed title from ~~DGR-024: Implement in-memory fake gRPC seam transport~~ to DGR-024: Implement real generated-gRPC protocol harness

2026-07-18 21:21:03 +00:00

popov added status:ready and removed status:in-progress labels 2026-07-18 21:23:44 +00:00

popov commented

2026-07-18 21:23:44 +00:00

Policy audit completed: the former fake/in-memory seam specification was replaced in canonical PRD commit 54d19f9 with a real generated Python/C++ gRPC localhost-process harness. The old dirty fake-seam files remain preserved but receive zero credit and must not be integrated.

Policy audit completed: the former fake/in-memory seam specification was replaced in canonical PRD commit 54d19f9 with a real generated Python/C++ gRPC localhost-process harness. The old dirty fake-seam files remain preserved but receive zero credit and must not be integrated.

popov commented

2026-07-19 14:44:33 +00:00

REAL-IMPLEMENTATION CONTINUATION (2026-07-19): the durable workers are now executing repair/verify passes under the hard no-fake/stub/synthetic policy. Fake/in-memory artifacts are being removed and replaced with real generated-gRPC / upstream-patch / runtime-identity behavior. No completion credit is granted until independent review and integrated-branch gates pass.

popov added the ready-for-agent label 2026-07-21 18:54:23 +00:00

popov added status:in-progress and removed status:ready labels 2026-07-22 05:30:33 +00:00

popov closed this issue

2026-07-22 06:44:23 +00:00

popov added status:completed and removed status:in-progress labels 2026-07-22 06:44:23 +00:00

Sign in to join this conversation.

2 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: popov/neuron-tai#8