feat(native): MYCELIUM_DATA_DIR root + pgdata/models/wire-cert layout — #176 sub-task 3 prep#202
Merged
Merged
Conversation
…#176 sub-task 3 prep) Implements ticket 2 of `docs/native-tauri-shell-spike.md` "Suggested follow-up issues" — the headless prerequisite for the Tauri shell PR. The pre-existing `MYCELIUM_PGLITE_DATA_DIR` (factory.ts) and `MYCELIUM_LLAMA_MODELS_DIR` (embeddings.ts) defaulted to two unrelated per-OS paths (`…/mycelium/data` and `…/mycelium/models`). The Tauri shell needs to pass ONE root via `MYCELIUM_DATA_DIR` and have PGlite, llama.cpp, and the Wave-3 wire-cert all live underneath it. Changes: - `mcp-server/src/native/data-dir.ts` (new): single source of truth. - `defaultRootDir()` per platform (macOS/Linux/Windows). - `getDataDirLayout()` resolves `${root}/pgdata`, `${root}/models`, `${root}/wire-cert`. Legacy env vars still override their own subpath. - `ensureDataDirs()` mkdir's all four (root + 3 subdirs), idempotent. - `factory.ts`: `defaultDataDir()` delegates to the layout. `createNativeDb()` calls `ensureDataDirs()` before PGlite touches disk so `wire-cert/` exists empty for Wave-2/3 to populate (per spike §"Wave-3 forward-compatibility" row 4 — "reserve the path so the migration wizard does not have to special-case it later"). - `embeddings.ts`: `defaultLlamaModelsDir()` delegates to the layout. Drops the platform branch duplicated from factory.ts. Backward compatibility: - `MYCELIUM_PGLITE_DATA_DIR` still wins over `${root}/pgdata` (Spike-1 installs that pinned the path keep working). - `MYCELIUM_LLAMA_MODELS_DIR` still wins over `${root}/models`. - New default for the PGlite subdir flips from `…/mycelium/data` to `…/mycelium/pgdata` — no production users yet (PGlite landed in #185 this week and is dual-build, not the primary backend), but flagged for the migration wizard (#176 sub-task 7) which now writes to `pgdata/`. Tests: 6 new, all green. Full suite 1028/1029 green (1 pre-existing skip). No regression in pglite-adapter, llama-cpp-embedding, or embeddings tests. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This was referenced May 3, 2026
Merged
Merged
Dewinator
added a commit
that referenced
this pull request
May 4, 2026
The previous "0 PRs offen" snapshot is stale: the agent itself opened a 4-PR queue in the last 4 ticks (#208 standalone + #209→#210→#211 stack). Refresh the Aktiver-Code-Bestand block to credit landed sub-tasks (#202 DATA_DIR, #203 Tauri shell, #204 CI gate, #205 migration helpers) and add a Queue-Stand note that names the open stack, the linear merge constraint, and the ~19 still-supabase-direct services that the DbClient migration sub-story will work through atomically. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Dewinator
added a commit
that referenced
this pull request
May 4, 2026
…ture) + 26h Reed-lag Last refresh (ec8a088, tick 173) captured the 4-PR native-app cluster (#208/#209/#210/#211) right after Reed drained #202-#207. Tick 174 then opened PR #212 (W4.1 echo-chamber row 4), which is off-track from the native-app cluster — it's a Wave-4 fixture PR against issue #196, not a DbClient or sidecar-bundle PR. CLAUDE.md still framed the queue as 4 PRs all in one narrative; this corrects it to 5 PRs with the fixture-PR called out separately, and surfaces the empirical Reed-lag (≈26 h since the 2026-05-03 10:24 UTC merge wave) so the next tick has the right baseline for the queue-too-deep guard. Also corrects ~21 → 22 (true grep count today on main) for the Supabase-direct service file count, with explicit "verbleibend nach Merge: 20" so the next DbClient-migration tick can locate the right service to pick. No production code touched; pure CLAUDE.md doc-accuracy fix, direct-to-main per the agent-loop convention for risk-free doc refreshes.
Dewinator
added a commit
that referenced
this pull request
May 4, 2026
…sks 1/2/3-most/4/5/7-core landed The 09024a2 (tick 103) snapshot is two days stale: 9 of the 9 PRs in its "recommended merge order" have merged, plus #197/#198/#201/#202/#203/#204/ #205/#206/#207 landed in the same window. Refresh the doc that "is the canonical entry point for the native-app initiative" so a fresh reader does not get a 79-tick-old picture. Concrete changes: - Phase line: "spike phase complete · gated on PR-queue drain" → "implementation underway · sub-tasks 1, 2, 3 (most), 4, 5, 7 (core) landed; 5 PRs open, ~26 h Reed-lag since #202–#207". - End-state installer list: ".dmg / .msi / .AppImage" → ".pkg / .msi / .AppImage", with the 2026-05-04 epic-body decision (.pkg, not .dmg) cited. - Sub-task table: legend (✅ landed, 🔄 PR open, ⏳ spike-only) plus per-row PR status reflecting the current main: - Row 1 picks up #204 (79-migration walk as CI gate). - Row 3 picks up #202, #203, #206, #207 as landed and #208–#211 as open. Row was "_none yet_". - Row 4 / 5 reframed as "inherent" (PGlite is one WASM binary; node-llama-cpp ships per-platform bindings). - Row 7 picks up #205 (Docker → PGlite migration helpers core). - New "DbClient sub-story" subsection: 22 services touch Supabase directly on main; 2 in flight (Skills, SwarmPin); 20 remain after the open stack lands. Pattern: one service per PR. - "Recommended merge order — open PR queue (9 deep)" section replaced with "Open PR queue (5 deep, all CLEAN+MERGEABLE)" — the 9 it documented have all merged. - "Post-drain implementation gate" → "Post-stack implementation gate": parallel tracks (DbClient long-tail of 20 services + the sub-tasks 6/8/9 release-pipeline triangle on tauri-plugin-updater per #207). Doc-only, zero risk, direct-to-main per the standing rule for this file shape. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Dewinator
added a commit
that referenced
this pull request
May 4, 2026
The bird's-eye doc lagged the 2026-05-03 merge wave (PRs #202–#207 landing sub-tasks 1, 2, 3-scaffold, 4, 5, 7-core) and the W3 spike chain that settled the empirical mDNS budgets on `main`. Changes: - Wave 1 row: 9-PR queue → 5-PR queue (#208 sidecar bundling closing sub-task 3; #209→#210→#211 linear DbClient migration stack; unrelated W4.1 fixture #212). Note the 22-services DbClient sub-story with 2 already in flight. - Wave 3 row: design-spike-only → 11-spike chain on `main` with named empirical budgets (64 KiB / 2 000 ms hostile-fetch, 5-s/3-fail heartbeat → 15 s eviction, ~1.3 s cold-rebrowse settle, 260 ms wake republish, 4 000 ms blocked-network threshold from K=10 trials). Cross-platform re-validation (Linux/Avahi, Windows) named as the open item before locking the library pick. - Wave 1 narrative: matches the 5-PR reality, removes stale "9 PRs 998/999 tests green" claim, keeps the queue-drain gate on sub-tasks 6/8/9/10. Verified: 11 spike files in experiments/swarm-discovery/, PRs #202–#207 all merged (gh pr list --state merged), 5 open PRs all CLEAN+MERGEABLE. Refresh date bumped to 2026-05-04. 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
Implements ticket 2 of
docs/native-tauri-shell-spike.md"Suggested follow-up issues" — the headless prerequisite for the Tauri shell PR. Reed wants to ship a native standalone app (#176, Wave 1, prio 0.98); this PR teaches the mcp-server side to honour a singleMYCELIUM_DATA_DIRroot that the future Tauri sidecar passes in, instead of two unrelated per-OS paths.Why
Pre-existing state:
factory.tsdefaulted PGlite to…/mycelium/data,embeddings.tsdefaulted llama.cpp models to…/mycelium/models. Two unrelated per-OS branches, both duplicated. The Tauri shell needs to pass ONE env (MYCELIUM_DATA_DIR) and have PGlite, llama.cpp, AND the Wave-3 wire-cert all live underneath it (per shell-spike §3 lifecycle, §6 data dir, and the Wave-3 forward-compat table row 4).Changes
mcp-server/src/native/data-dir.ts(new): single source of truth.defaultRootDir()per platform (macOS/Linux/Windows).getDataDirLayout()resolves${root}/pgdata,${root}/models,${root}/wire-cert. Legacy env vars still override their own subpath.ensureDataDirs()mkdir's all four (root + 3 subdirs), idempotent.factory.ts:defaultDataDir()delegates to the layout.createNativeDb()callsensureDataDirs()before PGlite touches disk sowire-cert/exists empty for Wave-2/3 to populate (per spike: "reserve the path so the migration wizard does not have to special-case it later").embeddings.ts:defaultLlamaModelsDir()delegates to the layout. Drops the platform branch duplicated from factory.ts.Backward compatibility
MYCELIUM_PGLITE_DATA_DIRstill wins over${root}/pgdata(Spike-1 installs that pinned the path keep working).MYCELIUM_LLAMA_MODELS_DIRstill wins over${root}/models.…/mycelium/datato…/mycelium/pgdata— no production users yet (PGlite landed in feat(native): PGlite adapter foundation — issue #184 part 1 #185 this week and is dual-build, not the primary backend), but flagged for the migration wizard (epic: native standalone app (no Docker) — macOS / Windows / Linux #176 sub-task 7) which now writes topgdata/.Tests
mcp-server/src/__tests__/native/data-dir.test.ts: defaults, root-override propagation, both legacy overrides, mkdir creation, idempotence.Test plan
cd mcp-server && npm test— full suite greenMYCELIUM_DATA_DIR=/tmp/x npm teststill passes (env propagation)experiments/native-pg/spike-pglite.mjsstill works (uses default path)What this unblocks
Ticket 1 (Tauri 2 shell scaffold) —
tauri.conf.jsonwill spawn the sidecar withMYCELIUM_DATA_DIR=BaseDirectory::AppLocalData/myceliumand the mcp-server now knows what to do with it. Tickets 3 + 4 follow.🤖 Generated with Claude Code