Skip to content

feat(vcs): add persisted sparse worktree materialization - #1

Open
maxibotstef wants to merge 31 commits into
review-base/sparse-worktree-materialization-07891e9from
feat/sparse-worktree-materialization-upstream-20260904
Open

feat(vcs): add persisted sparse worktree materialization#1
maxibotstef wants to merge 31 commits into
review-base/sparse-worktree-materialization-07891e9from
feat/sparse-worktree-materialization-upstream-20260904

Conversation

@maxibotstef

@maxibotstef maxibotstef commented Sep 5, 2026

Copy link
Copy Markdown
Owner

Clawd worktrees can be created from an explicit, persisted sparse profile, with full checkout as the default and fallback. The UI reads the repository's versioned contract and structured task card; it never derives a profile from conversation, title, or branch text.

The server creates and verifies the worktree before starting a provider, persists requested and effective identity through thread events and migration 048, and recreates missing worktrees from the persisted effective profile. Clean expansion to full is the only transition. Corrupt identity and later-turn mismatches block with recovery guidance. A selected task card still loading prevents Send until its result is available.

Verification covers driver, reactor, contract, projection, migration, review and UI behavior; real creator comparisons against full twins for all three profiles; interruption tests; and isolated source-built event/projection readback, clean expansion and omitted-profile full behavior. Final reactor suite: 69 passing. The projector regression uses the Effect test harness and passes with the other ten projector tests.

CI passed on the final branch head: lint, typechecks, desktop build/preload verification, tests and all three server shards, Rust, release smoke, and mobile native static analysis. Native fingerprint also passed. CI runner routing changes only this named fork to GitHub-hosted workers; upstream runner labels and all existing checks are preserved. One green run is not a claim of long-term runner reliability.

Fork-only, ready for review. The final exact Claude Fable 5.1/high review returned GO_WITH_CAVEATS, no blockers. Its caveats concern historical evidence wording and runner-performance uncertainty, not a demonstrated product-code blocker. Production code is unchanged since 4b99442; the final CI/test head is f6dda9b.

This PR targets the dedicated reviewed base at 07891e9, not fork main or upstream main. No upstream submission is requested. Migration 048 must be checked against any different landing target. The PR is not merged; no installed bundle changes, live-database writes, service restart or rollout occurred.

Implemented with Codex in isolated worktrees; reviewed with exact Claude Fable 5.1 at high effort through ACP. Source and CI evidence does not claim a deployment or completed landing.

Stef added 29 commits September 4, 2026 06:20
@github-actions github-actions Bot added the vouch:trusted PR author is trusted by repo permissions or the VOUCHED list. label Sep 5, 2026
@github-actions

github-actions Bot commented Sep 5, 2026

Copy link
Copy Markdown

Thread transfer impact

✅ Thread transfer remains within every enforced ceiling.

ℹ️ No successful main baseline artifact is available yet. This run establishes the initial measurement.

Provider Metric Main baseline This PR Impact PR ceiling
Codex Total thread wire 13.6 KiB 15.1 KiB
Codex Thread snapshot wire 7.2 KiB 7.3 KiB
Codex Live turn WebSocket wire 6.4 KiB 7.8 KiB
Codex Live turn WebSocket decoded 56.2 KiB 66.4 KiB
Codex Live turn messages 8 21
Claude Total thread wire 13.7 KiB 15.1 KiB
Claude Thread snapshot wire 7.2 KiB 7.3 KiB
Claude Live turn WebSocket wire 6.5 KiB 7.8 KiB
Claude Live turn WebSocket decoded 57.8 KiB 66.4 KiB
Claude Live turn messages 9 21

Baseline: unavailable · PR result: f6dda9b · Source CI: success

Scenario and decoded snapshot size

10 historical turns, 5 command tools per turn, 878.9 KiB retained MCP result per historical turn, and a 1.05 MiB retained result in the measured turn.

  • Codex decoded thread snapshot: 114.1 KiB
  • Claude decoded thread snapshot: 114.8 KiB

Updated in place by a trusted workflow. PR artifacts are strictly validated and never executed.

@maxibotstef
maxibotstef marked this pull request as ready for review September 5, 2026 12:19
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

size:XXL vouch:trusted PR author is trusted by repo permissions or the VOUCHED list.

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant