Repository navigation
Description
Crate split step 4 of 7 — from the merged design
Design: doc/2026-11-01T16-53-00Z_design_id_crate_split/2026-11-01T16-53-00Z_design_id_crate_split.md (merged as PR #3824, 2026-11-01). Tracking issue: #3447. Program: #3195 (build-time review; Drew's priority order — shipped-binary runtime performance first, build time second).
Scope (design, verbatim intent)
Design step 4: id-groups. Groups/membership/encryption extract; seam 5 (encryption's session leak) is cut, and seam 6's bridge file leaves for its future home in id-worlds.
Current state
Not started.
Dependency — do not start early
Step 3 (id-sync) merged. If this step's queue task is claimed before the dependency lands, park it with that reason.
Fuel canary (non-negotiable, every step): the FUEL_BUDGETS tables in pkgs/id/src/sandbox.rs must be byte-identical after this step; the fuel-tables diff is committed in the step PR. Any drift means behaviour changed and the step stops. Full verification protocol: design note §"Verification protocol".
Verification protocol (design, every step): one PR; main green throughout; the shipped binary is built by the same Nix derivations at every point (all three entry points: the sub-flake's id-web and id-lib, and default.nix); test count matches or exceeds the step-0 baseline with any delta named; cold check-build timing recorded on #3447 and #3195 with the laptop load noted. Heavy builds run under ~/tools/fleet-build-lock on the laptop (max 2 concurrent, CARGO_BUILD_JOBS = half the cores) — there is no load gate (Drew, 2026-11-02).
Acceptance criteria
- The extraction lands as its own PR off current main, sequenced after its predecessor step (see Related).
- The id facade re-exports the extracted crate so existing call sites compile unchanged in behavior.
- CI green on the PR (workspace build + the id test suites), including the F11.10 production-line guard.
- This issue closes with the merge; the next step's issue is unblocked in sequence.
Activity
dezren39 commentedon Nov 2, 2026
Pickup by repo queue worker lane C for task 6301 (crate split step 4 of #3447). Sequence revalidation at 2026-11-02T15:52Z: step 3 (#4406, extract id-sync) is still OPEN and no merged PR for it was found, so step 4 does not start early. Task 6301 parks until #4406 lands.
dezren39 commentedon Nov 2, 2026
Parked by repo queue worker lane C (task 6301): sequence gate not met at 2026-11-02T15:52Z — step 3 (#4406) still OPEN, not merged. Resume condition: #4406 closed/merged; then start the id-groups extraction per task 6301 (canary: FUEL_BUDGETS byte-identical; builds under ~/tools/fleet-build-lock per the packet's build policy).
76 remaining items
dezren39 commentedon Nov 4, 2026
🤖 Agent note · handed off
dezren39 commentedon Nov 4, 2026
🤖 Agent note · handed off
dezren39 commentedon Nov 5, 2026
agent-note:v1
event: handed_off
task: 6301
worker: repo-worker-code (lane code-f)
stage: handed_off
summary: Task 6301 re-claimed by lane code-f; execution-lane assignment again returned assigned_lane=opencode (session_id null), so this lane hands off per the runbook. Standing resume condition from prior checkpoints: the step-4 sequence gate — step 3 (issue #4406) must close/merge before extraction work on #4407 starts.
dezren39 commentedon Nov 5, 2026
🤖 Agent note · handed off
opencode; handing off per the worker runbook §1b — I do not work this task directly and created no worktree.handed_offrecorded on task 6301 at 2026-11-05T01:30Z; labels already queue/claimed + status/in-progress + work/assigned (verified fresh this shift).dezren39 commentedon Nov 5, 2026
🤖 Agent note · handed off
dezren39 commentedon Nov 5, 2026
🤖 Agent note · handed off
opencodedezren39 commentedon Nov 5, 2026
🤖 Agent note · handed off
dezren39 commentedon Nov 5, 2026
🤖 Agent note · handed off
dezren39 commentedon Nov 5, 2026
🤖 Agent note · handed off
dezren39 commentedon Nov 5, 2026
🤖 Agent note · handed off