Rename, links and the bulk-edit job

kayg commented

2026-10-13 15:01:15 +00:00

Owner

Rename, links and the bulk-edit job (ADR 0015, 0024).

Links on disk are [title](<relative/path.md>) only; [[ and @ are input sugar.
One journaled bulk-edit job (plan, batches, resume after crash, undo), very fast. Rename = one bulk edit that rewrites links.
Collisions get " 2", case-only renames via temp name, Unicode-normalised names.

Rename, links and the bulk-edit job (ADR 0015, 0024). - Links on disk are `[title](<relative/path.md>)` only; `[[` and `@` are input sugar. - One journaled bulk-edit job (plan, batches, resume after crash, undo), very fast. Rename = one bulk edit that rewrites links. - Collisions get " 2", case-only renames via temp name, Unicode-normalised names.

kayg added the

backend

reliability

labels

2026-10-13 15:01:15 +00:00

kayg referenced this issue

2026-10-13 15:19:05 +00:00

Backend core follow-ups #17

kayg referenced this issue

2026-10-13 16:05:32 +00:00

Notes storage and live rooms #24

kayg referenced this issue

2026-10-13 16:09:05 +00:00

Notes storage and live rooms #27

kayg referenced this issue

2026-10-13 16:14:29 +00:00

Notes storage and live rooms #28

kayg referenced this issue

2026-10-13 19:18:43 +00:00

Search foundation: provider trait, per-user FTS5 index, live from the journal #53

kayg referenced this issue

2026-10-13 19:20:59 +00:00

Search foundation: provider trait, per-user FTS5 index, live from the journal #53

kayg referenced this issue

2026-10-13 19:25:56 +00:00

Real UI: the mockup design on the real engine, the whole flow end to end #55

kayg referenced this issue

2026-10-13 19:47:19 +00:00

Real UI: the mockup design on the real engine, the whole flow end to end #60

kayg referenced this issue

2026-10-13 19:57:10 +00:00

Real UI: the mockup design on the real engine, the whole flow end to end #66

kayg referenced this issue

2026-10-13 20:02:02 +00:00

Real UI: the mockup design on the real engine, the whole flow end to end #67

kayg referenced this issue

2026-10-13 20:05:34 +00:00

Real UI: the mockup design on the real engine, the whole flow end to end #68

kayg referenced this issue

2026-10-13 20:11:55 +00:00

Real UI: the mockup design on the real engine, the whole flow end to end #69

kayg referenced this issue

2026-10-13 20:17:25 +00:00

Real UI: the mockup design on the real engine, the whole flow end to end #70

kayg referenced this issue from a pull request that will close it,

2026-10-13 20:45:59 +00:00

Notes: rename, links, backlinks and the bulk-edit job #74

kayg commented

2026-10-13 20:46:07 +00:00

Author

Owner

Progress: PR #74 is open. It covers the bulk-edit job (plan, batches, resume after a crash, undo), rename with link repointing, the backlinks API (linked plus unlinked through search), and make note / move to note as one job each. 254 workspace tests pass, including a crash mid-job that resumes exactly once.

Local perf at 10k notes with 1k links: rename takes 1.2 to 2.2 s, and backlinks p95 is under 1 ms. The perf VM was busy with #44, so a perf-VM run is still to come. The API shape for the UI is in the PR body.

Progress: PR #74 is open. It covers the bulk-edit job (plan, batches, resume after a crash, undo), rename with link repointing, the backlinks API (linked plus unlinked through search), and make note / move to note as one job each. 254 workspace tests pass, including a crash mid-job that resumes exactly once. Local perf at 10k notes with 1k links: rename takes 1.2 to 2.2 s, and backlinks p95 is under 1 ms. The perf VM was busy with #44, so a perf-VM run is still to come. The API shape for the UI is in the PR body.

kayg referenced this issue from a pull request that will close it,

2026-10-13 20:58:30 +00:00

Notes: rename, links, backlinks and the bulk-edit job #75

kayg referenced this issue from a pull request that will close it,

2026-10-13 21:11:01 +00:00

Notes: rename, links, backlinks and the bulk-edit job #76

kayg referenced this issue from a pull request that will close it,

2026-10-13 21:13:59 +00:00

Notes: rename, links, backlinks and the bulk-edit job #77

kayg referenced this issue from a pull request that will close it,

2026-10-13 21:19:57 +00:00

Notes: rename, links, backlinks and the bulk-edit job #78

kayg commented

2026-10-13 21:22:28 +00:00

Author

Owner

PR #78 is green on both jobs (it replaces #74–#77, which the forge stopped syncing). Since the first post:

Undo fix: undo now removes a made note even after the server's own first save has added its properties. Only someone else's typing keeps the note.
e2e fix: seeded notes are named after their H1, and outside reads and writes ask the server where the file is. The file now follows the title, so a test note was being renamed under the test.
Flaky test fixed: notes_live saves 20 ms after a change, and could save before the websocket join. The test now types at the end of whatever text it synced.

The perf-VM run is still to do; the VM was busy with #44's 100k run the whole time.

PR #78 is green on both jobs (it replaces #74–#77, which the forge stopped syncing). Since the first post: - **Undo fix:** undo now removes a made note even after the server's own first save has added its properties. Only someone else's typing keeps the note. - **e2e fix:** seeded notes are named after their H1, and outside reads and writes ask the server where the file is. The file now follows the title, so a test note was being renamed under the test. - **Flaky test fixed:** `notes_live` saves 20 ms after a change, and could save before the websocket join. The test now types at the end of whatever text it synced. The perf-VM run is still to do; the VM was busy with #44's 100k run the whole time.

kayg closed this issue

2026-10-13 21:23:06 +00:00

kayg referenced this issue

2026-10-13 21:30:16 +00:00

Web: make note, move to note, mentions and renames on the #9 jobs #79

kayg referenced this issue

2026-10-13 21:41:22 +00:00

Web: make note, move to note, mentions and renames on the #9 jobs #81

kayg referenced this issue

2026-10-13 21:50:59 +00:00

Rooms: typing reaches collaborators before the disk #82

kayg referenced this issue