Repository navigation
One package, client-first homepage and the runner (#84, #85, #86 onto main) - #87
Merged
Merged
Conversation
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* docs: lead homepage and README with Database client * test(e2e): check the client-first homepage and refresh its screenshots Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* Add scoped durable derivation runner with acknowledged delivery snapshots * Exercise durable runner recovery and temporal replay across backend fixtures * Run the compliance host through durable handlers and document runner delivery * fix(runner): keep wakeups that arrive during head reads; settle tests without clock drift The coordinator captured its wake signal after reading the head, so a commit or finished delivery that fired during that read was lost until the next idle check. The corpus settle helper advanced the TestClock while waiting for backend I/O, which let slow backends (PostgreSQL) drift past the valid-time boundaries cases 6 and 7b assert on. It now waits for I/O in real time and moves the clock only to reach a scheduled retry or for one idle-check tick when unscanned commits or a due deadline sit unchanged. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * test(runner): wait for timer installation instead of reading it once Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * plan: W2.4 and W3 PRs open Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Case 7b placed its correction deadline 10 ms after the replacement. Under heavy machine load the settle helper's idle-check ticks drifted the injected clock past it, the deadline fired, and the wait for it never finished. Cases 6 and 7b now space boundaries seconds apart, and a settle timeout reports the runner's last status. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Contributor
|
Documentation preview removed (PR closed). |
This was referenced Nov 12, 2026
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.
Brings three already-reviewed changes to
main: #84 (one package), #85 (client-first homepage) and #86 (runner). Ben merged those into the stacked branchesergonomics/next-2andergonomics/one-package, so they never reachedmain. This branch replays the same commits onto the squashedmain(b184b37). The rebased commits produce a tree identical toergonomics/one-packageate92d440. One commit is added on top: a test-only fix for a runner suite flake found while checking the combined branch under load (below).What it carries
@triplex-build/triplexwith subpaths/sql,/sqlite,/postgres,/cloudflareand/foundationdb(both experimental),/host,/http,/testkitand/mcp. Onetriplexbin, includingtriplex mcpandtriplex dashboard. Drivers are optional peers.docs/releasing.mdlists the npm names to deprecate. Minor changeset.Databaseclient and the single install linenpm install @triplex-build/triplex effect. The homepage code is checked, and the e2e test and screenshots are updated.@triplex-build/triplex/runnerwith the conformance suite on KV, SQLite and PostgreSQL, the compliance host rewritten onto it, and a fix for a lost-wakeup race.ConsumerCheckpointnow survives a wall-clock rollback. New public API:KvTriples.layerWithClock. Minor changeset.Merge note
#83 (site) also edits
docs/current-state.mdanddocs/roadmap.md. Whichever of the two merges second will conflict there. I'll resolve it on that branch after the first one merges.Checks
Local, on the combined branch:
pnpm check: passed (at the rebased head, before the test-only commit; format and lint rerun after it).pnpm test:postgres:integration: 77/77 passed.pnpm pack:check: one tarball and a clean consumer, including/runner.Flake found and fixed: the first Postgres run failed runner case 7b once under a load average of ~44. Case 7b had boundaries only 10 ms apart, so the test helper's 1 ms clock ticks could drift past a deadline. The runner itself is correct. Cases 6 and 7b now space their boundaries seconds apart, and a settle timeout now reports the runner's last status. Afterwards the runner suite passed 10/10 runs, then 8/8 with every CPU saturated.
🤖 Generated with Claude Code