Revert NEC-15's checkpoint writer until its stall is understood #152

Merged

nectenda-agent merged 1 commit from worktree-revert-nec-15 into main 2026-09-24 16:39:59 +01:00

Conversation 0 Commits 1 Files changed 11 +52 -709

nectenda-agent commented 2026-09-24 16:27:09 +01:00

Collaborator

Copy link

Task: NEC-15 — https://projectron.nerchure.com/tasks/15

Reverts #145 (206d1f6 and b7e3271). #151 (the NEC-87 fix) landed on top of NEC-15 meanwhile and is independent of it; its only shared file is the generated docs/v1/conformance.md. So the result is exactly cdda616 plus #151. That was checked: git diff cdda616 HEAD equals git diff b7e3271 b9366a9 outside the generated file, and the conformance check passes. cdda616 is the last main that passed CI in full, and it is what eu1 and accounts are running.

Why. main's run on b7e3271 failed in build: checkpoints.test.ts › "past the compaction threshold, through the WebSocket path › keeps the earlier state recoverable" timed out at 5 s. It was a stall, not slowness: the server started at 15:05:51.19 and the test logged nothing more. The same commit passed on the PR's run, and 120 of 120 runs on the Mac under parallel load, so it is intermittent and unexplained. It stalls where the test drives the real threaded writer through a compaction, as production does, so it does not ship until someone has instrumented it.

Nothing to migrate back. Promote was skipped on b7e3271, and eu1 and accounts are on cdda616, so no production database ever carried #145's schema change.

Changelog: #145's entry must be dropped at the next cut. #145 carried "The server now keeps earlier versions of each note when it compacts edit history, instead of discarding them." Users never received it, so this revert is NONE, but scripts/changelog.mjs has no revert handling and would otherwise list #145.

Changelog

NONE

Task: NEC-15 — https://projectron.nerchure.com/tasks/15 Reverts #145 (206d1f6 and b7e3271). #151 (the NEC-87 fix) landed on top of NEC-15 meanwhile and is independent of it; its only shared file is the generated `docs/v1/conformance.md`. So the result is **exactly cdda616 plus #151**. That was checked: `git diff cdda616 HEAD` equals `git diff b7e3271 b9366a9` outside the generated file, and the conformance check passes. cdda616 is the last main that passed CI in full, and it is what eu1 and accounts are running. **Why.** main's run on b7e3271 failed in `build`: `checkpoints.test.ts` › "past the compaction threshold, through the WebSocket path › keeps the earlier state recoverable" timed out at 5 s. It was a **stall**, not slowness: the server started at 15:05:51.19 and the test logged nothing more. The same commit passed on the PR's run, and 120 of 120 runs on the Mac under parallel load, so it is intermittent and unexplained. It stalls where the test drives the real threaded writer through a compaction, as production does, so it does not ship until someone has instrumented it. **Nothing to migrate back.** Promote was skipped on b7e3271, and eu1 and accounts are on cdda616, so no production database ever carried #145's schema change. **Changelog: #145's entry must be dropped at the next cut.** #145 carried "The server now keeps earlier versions of each note when it compacts edit history, instead of discarding them." Users never received it, so this revert is NONE, but `scripts/changelog.mjs` has no revert handling and would otherwise list #145. ## Changelog NONE

nectenda-agent added the

area:server

label 2026-09-24 16:27:09 +01:00

nectenda-agent added 1 commit 2026-09-24 16:27:09 +01:00

Revert NEC-15's checkpoint writer until its stall is understood ... 1bd34ce05c

This reverts 206d1f6 and b7e3271 (PR #145).

main's run on b7e3271 failed in build: checkpoints.test.ts, "past the
compaction threshold, through the WebSocket path > keeps the earlier
state recoverable", timed out at 5 s. It was not slow: the test logged
its server starting at 15:05:51.19 and then nothing for five seconds.
The same commit passed on the pull request's run, and 120 of 120 runs
on a Mac under parallel load. So it is intermittent, and not understood.

It stalls where the test drives the real threaded writer through a
compaction, as production does. Whether that is the test's harness or
the server's write path is not known, and a stalled writer is a stalled
sync server. So this does not ship until somebody has instrumented it.
The fleet never took it: promote was skipped, and eu1 and accounts were
still on cdda616, so no production database ever carried the new schema
and nothing needs migrating back.

main goes green on the code it had, the merge queue resumes, and NEC-15
goes back on the board with the evidence.

Task: NEC-15

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_015k9tkUtbU3wBpCHAHq7Qas

cruelacid force-pushed worktree-revert-nec-15 from 1bd34ce05c to b6b264de8e 2026-09-24 16:33:44 +01:00 Compare

nectenda-agent merged commit b6b264de8e into main 2026-09-24 16:39:59 +01:00

cruelacid referenced this pull request 2026-09-24 20:17:05 +01:00

Drop a reverted pull request’s entry from the changelog it never shipped in #156

cruelacid referenced this pull request from a commit 2026-09-24 20:17:06 +01:00

Drop a reverted pull request's entry from the changelog it never shipped in

Sign in to join this conversation.

Reviewers

No reviewers

Labels

Clear labels   

area:docs

docs/

  

area:identity

packages/identity

  

area:ops

deploy/, CI, hosting

  

area:plugin

packages/plugin

  

area:server

packages/server

  

channel:community

Obsidian forum, Obsidian Discord, Reddit

  

channel:direct

Cold outreach to named firms

  

channel:owned

nectenda.com pages, the feedback board

  

channel:press

Hacker News, newsletters, outlets, YouTube creators

  

channel:social

Bluesky, Mastodon, X, YouTube accounts

  

e2ee-constrained

End-to-end encryption changes the shape of the answer

  

gate:at-ga

Held until Gates A-C close and terms/privacy leave DRAFT

  

gate:pre-ga

Safe to do before the launch gates close

  

marketing

Reach and positioning; work published off nectenda.com

  

parity

Relay feature-parity programme

  

relay:absent

4 - Not supported, not requested, not planned at Relay

  

relay:planned

2 - On Relay's roadmap (Active/Planned) or has an open PR

  

relay:requested

3 - Asked for by Relay users or under 'Considering'; not committed

  

relay:supported

1 - Relay ships this today

  

risk:additive

New message type, table or column; backward compatible

  

risk:contract

Changes something existing accounts or stored data depend on

  

risk:none

Client-only; the server is untouched

  

usability

How the product is understood and learned, not what it does

No labels

area:docs

area:identity

area:ops

area:plugin

area:server

channel:community

channel:direct

channel:owned

channel:press

channel:social

e2ee-constrained

gate:at-ga

gate:pre-ga

marketing

parity

relay:absent

relay:planned

relay:requested

relay:supported

risk:additive

risk:contract

risk:none

usability

Milestone

Clear milestone

No items

No milestone

Projects

Clear projects

No items

No project

Assignees

Clear assignees

No assignees

2 participants

Notifications

Subscribe

Due date

The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set

Reference

Nectenda/nectenda!152

No description provided.