replace flock-based daemon liveness with socket-ping detection#4
Merged
rsnodgrass merged 1 commit intomainfrom Feb 17, 2026
Merged
Conversation
Contributor
Author
|
@rsnodgrass |
port8080
added a commit
that referenced
this pull request
Feb 18, 2026
3 tasks
jmf-pobox
added a commit
to punt-labs/quarry
that referenced
this pull request
Feb 28, 2026
Examined sageox/ox#4. The product surface is a conversation recording link and a session link in the PR body. The hosted web UI is for browsing. The capture mechanism is logging. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
jmf-pobox
added a commit
to punt-labs/quarry
that referenced
this pull request
Feb 28, 2026
* fix: hook matcher now matches standalone MCP tool names The PostToolUse suppress-output hook only matched plugin-namespaced tool names (mcp__plugin_quarry_quarry__*). Standalone MCP servers use mcp__quarry__* which the matcher missed, causing raw JSON output in non-plugin sessions. Closes quarry-fsx Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix: route /quarry sync through background CLI to avoid blocking The sync_all_registrations MCP tool blocks the conversation for 10-30s with large knowledge bases. Route through `quarry sync` via Bash run_in_background instead — user gets notified on completion and can keep working. Closes quarry-9cj Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * docs: ambient knowledge vision and README update Add research/vision.md documenting the full architecture for quarry as Claude Code's ambient memory layer — passive learning via hooks, active recall via session briefings and PreToolUse nudges, config-driven selectivity via /quarry learn, and staging queue for async ingestion. Update README Automagic Mode section to Ambient Knowledge, framing the distinction between quarry as a tool (projections) and quarry as a plugin (host integration that changes Claude Code's behavior). Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * docs: SageOx competitive analysis Hosted context platform solving the same core problem (knowledge evaporates between agent sessions). Quarry's position is open source, local-first, individual/small-team — not enterprise. Notes git-based team sharing as a potential future direction. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * docs: soften provenance distinction in competitive analysis Provenance is logging with correlation — save the conversation, save the commit, tag them together. Quarry's learning hooks already produce the raw material. The sharp line between knowledge tools and provenance tools was overstated. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * docs: add SageOx PR analysis — two URLs is the whole product Examined sageox/ox#4. The product surface is a conversation recording link and a session link in the PR body. The hosted web UI is for browsing. The capture mechanism is logging. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * docs: add shadow branch storage model to vision Following Entire.io's pattern (entire/checkpoints/v1), captured content could live on a git shadow branch — pushed alongside code via pre-push hook, pulled by teammates, vectorized locally. Simplest possible team sharing: no server, no accounts, git handles sync. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * docs: quarry as force multiplier for design tools PR/FAQ meeting transcripts, Z spec elaborations, risk assessments, and competitive analyses are high-value design artifacts locked in single conversations. Quarry's ambient capture makes them searchable across sessions — design tools become cumulative rather than starting from zero each time. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * docs: PR/FAQ for Quarry Ambient Knowledge vision New hypothesis-stage PR/FAQ exploring quarry as Claude Code's ambient memory layer. Two loops: passive learning captures web research and session transcripts; active recall nudges Claude before redundant web searches. Includes researcher-verified evidence (Qodo survey, Entire.io, SageOx, Limitless/Meta), competitive landscape with Windsurf Cascade Memories, and platform risk analysis. Peer reviewed: addressed TAM grounding, Usability risk upgrade to Medium, opportunity cost FAQ, scaling FAQ, Anthropic platform risk, and Qodo three-tier stat accuracy. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix: address CI and review feedback on PR #70 - Fix 15 markdownlint errors across README, research, and vision docs (code block language tags, blank lines around lists, bare URLs) - Replace absolute paths with relative links in research file - Qualify README ambient knowledge section as planned/future behavior - Add --db awareness note to /quarry sync command - Document PR workflow with gh pr checks/comments in CLAUDE.md Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.6 <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
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
.lock), platform-specificflock/LockFileExcode, and setsid/detach logicIsRunning()now pings the Unix socket (100ms timeout) instead of checking flockIsHealthy()distinguishes "no socket" (not running) from "socket exists but unresponsive" (hung)Discussion
https://www.sageox.ai/team/team_jihjpfkt8b/media/recordings/rec_019c688a-19c0-76f7-8763-597f072fc52f
Session
https://www.sageox.ai/repo/repo_019c5812-01e9-7b7d-b5b1-321c471c9777/sessions/2026-02-17T11-44-user-OxIOgz/view
Test plan
make lint— 0 issuesmake test— all passingox agent primestarts daemon,ox daemon statusconfirmsox statuswithout daemon — works, shows daemon not runningSAGEOX_DAEMON=false ox agent prime— no daemon starts