fe2cd12049d2c0f8683373aeb98b6470464254ed
9 Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
 sulthan
|
3303a55b20 |
feat: one Poll Lane per Site, replacing the shared pace (#100) (#106)
Closes #100. Each Site runs its own Poll Lane: an independent goroutine with its own rest and pace from the registry (`backend/internal/latest/sites.go`), replacing the shared cooldown/interval/stagger/batch configuration. Rest (1h, all six Sites including the browser trio) is enforced by the due query's WHERE clause; the Lane sleeps its effective gap between fetches — the registry 10s, or rest/eligible when a Site holds enough Series, floored at 1s with a Site-naming warning when the floor engages. Lane-local failure handling: - Two challenge-held results stop that Site's Lane for 15m; the probes keep their stamp, untried Series stay due. - A lost browser sets a shared Poller flag: the other browser Lanes skip their passes for the same 15m (no stamp-per-pass-per-Lane on a dead tab), then decay and probe again. - Browser wake gate preserved (5 due, or one waiting 15m, ADR-0005); one tab shared by the three browser Sites; "browser lane behind by X" logged every pass. - Cover work (healing a stored source URL and filling a blank from the series page) runs in the background so a slow CDN cannot consume a Lane's gap. Removed: `LATEST_CHAPTER_POLL_{COOLDOWN,BROWSER_COOLDOWN,INTERVAL,BATCH,STAGGER}` and the 6h browser rest. Only `LATEST_CHAPTER_POLL_ENABLED` remains; DEPLOY.md documents the exact `.env` edit. ADR-0010 records the decisions. Reviewed-on: #106 Co-authored-by: Sulthan Zaki <sultankiki05@gmail.com> Co-committed-by: Sulthan Zaki <sultankiki05@gmail.com> |
||
|
sulthan
|
e8d1cba6c5 |
Move cover bytes to content-addressed filesystem storage (#65)
Refs #56 ## Summary Moves Kagane cover bytes out of Postgres bytea storage into an immutable, content-addressed filesystem store. Reader-visible behavior remains unchanged: the existing session-gated route serves stored bytes, missing bytes use the existing browser fetch path, and no browser still returns a missing cover. ## Changes - Added migration 0008, which drops the legacy `covers` table and recreates it with only `address`, `path`, and `content_type`. Existing byte rows are intentionally dropped. - Added SHA-256 source-URL addressing with two-level sharding (`ab/cd/<sha256>`). Writes use a temp file plus atomic link; reads validate the stored relative path before opening it. - Made `COVER_DIR` required in runtime config and Compose. Compose passes it as a Docker build argument and volume target, so custom durable paths keep image ownership, runtime config, and the named `cover-data` volume aligned. - Updated every `store.Open` caller and documented configuration, deployment, backup, and troubleshooting behavior. - Added filesystem, restart, migration-drop, no-browser, and content-addressing coverage. ## Verification - `go test ./...` - `CGO_ENABLED=0 go build ./...` - `docker build --build-arg COVER_DIR=/data/covers -t manga-bookmark-cover-check-custom ./backend` - `docker compose config --format json` confirms custom `COVER_DIR` is the volume target - `git diff --check origin/main` - LSP diagnostics clean for touched Go files Parents #47 and #55 remain open as required by #56. Reviewed-on: #65 Co-authored-by: Sulthan Zaki <sultankiki05@gmail.com> Co-committed-by: Sulthan Zaki <sultankiki05@gmail.com> |
||
|
sulthan
|
cd3a7e3d01 |
feat(latest): split browser poll cooldown (#48)
## Summary Split latest-chapter polling cooldowns by fetch cost. Browser-backed kagane and novelfull series now rest longer without changing the cadence of plain-TLS sites. ## Behavior - Plain-TLS series keep the 1h default cooldown. - Browser-backed series use `LATEST_CHAPTER_POLL_BROWSER_COOLDOWN`, defaulting to 6h. - Both cooldowns share the existing 15m minimum floor; invalid values retain the existing fallback behavior. - The poller still selects both classes in one due query per cycle. - Existing ordering and exclusions remain unchanged: reader-count precedence, least-recently-checked ordering, finished exclusion, archived polling, and orphan exclusion. ## Implementation - Added the browser cooldown to backend configuration and passed it through production poller construction. - Added the browser-site list as the single routing source used for both due-query cutoff selection and fetcher choice. - Kept all query values parameterized; the site list is passed as a bound PostgreSQL array parameter. - Updated startup logging to report interval, plain cooldown, browser cooldown, batch, and stagger. - Documented the variable, default, and floor in `README.md`, `.env.example`, `backend/AGENTS.md`, and `docker-compose.yml`. ## Review findings addressed The first review found that configuration parsing was correct but `startLatestPoller` did not pass `BrowserCooldown` into `latest.Poller`; every browser-backed row would therefore have been due immediately. Production construction now goes through `newLatestPoller`, with a regression test covering both cooldown fields. The review also identified duplicated browser-site knowledge in fetch routing. `slices.Contains(browserBackedSites, site)` now reuses the same list already supplied to the store query. ## Verification - Focused backend tests pass: `go test ./internal/latest ./internal/store .`. - Full suite passes: `go test ./...`. - `graphify update .` completed. - Issue #42 was updated and closed. Reviewed-on: #48 Co-authored-by: Sulthan Zaki <sultankiki05@gmail.com> Co-committed-by: Sulthan Zaki <sultankiki05@gmail.com> |
||
|
sulthan
|
2ef769d421 |
Open registration to guild members (#27) (#36)
Closes #27. Guild membership is now the whole gate. `discordCallback` checks membership (and `DISCORD_REQUIRED_ROLE` when set), then `Store.EnsureReader` creates the Reader on first sight and returns the same row on every later login. The refusal returns before `EnsureReader`, so a turned-away sign-in leaves no row behind. `OWNER_DISCORD_ID` still seeds the owner, but only as the administrator — it no longer gates login. The cutover grace path goes with it: `API_TOKEN`, `API_TOKEN_GRACE_UNTIL` and the legacy branch in `httpmw.ResolveReader` are deleted, so a credential authenticates exactly one Reader or nothing. `userscript.Handler` drops its re-derivation too — the resolved path segment is already the credential. New surfaces: an empty library offers both install links (behind the tab-specific empty states, so "No favourites yet" still wins), and the owner alone gets a Readers panel with `POST /readers/{id}/revoke`. The owner's own row is not revocable — 404, not a self-logout. Isolation is asserted from both directions for read, modify and delete, and the shared-series invariant is pinned: two Readers on one series produce one series row, two independent progresses, one poll per due cycle, and one Reader's delete leaves the other's bookmark and the poll intact. Verified: `go test ./...` green; live smoke against a throwaway Postgres — empty-library state in both colour branches, roster rendering, a real revoke through the panel (target 401s next request, owner untouched), owner self-revoke refused 404, per-Reader `/u/<cred>` and bearer auth both 200 with 404 for an unknown credential. Reviewed-on: #36 Co-authored-by: Sulthan Zaki <sultankiki05@gmail.com> Co-committed-by: Sulthan Zaki <sultankiki05@gmail.com> |
||
|
sulthan
|
4229c179b0 |
rebrand: MangaBM → BookmarkManager, add novel library support (#15)
Two intertwined changes — the rebrand and the novel library were developed on
the same branch because the novel UI plumbing is part of the new "Bookmark
Manager" wordmark in the web shell.
## What it does
- **Rebrand**: MangaBM → BookmarkManager across the Go module, compose stack,
env vars, Traefik hostnames, container/image names, userscript storage
prefixes (`mangabm:cache` → `bmgr:manga:cache`, `mangabm:queue` → `bmgr:manga:queue`),
and docs.
- **Novel library**: same backend, two libraries. New `kind` column splits
bookmarks into `manga` / `novel`; PUT validates it. Two userscripts:
- `manga-bookmark.user.js` — unchanged behaviour, just stamps its own `kind`.
- `novel-bookmark.user.js` — separate Violentmonkey install with adapters
for **novelfull.com** (polled via headless browser — Cloudflare JS
challenge) and **lightnovelworld.net** (polled via plain TLS).
- **Web UI**: library switch on the app shell. Login art, libswitch, and
novel-site colours from the Cinder design snapshot.
## Plumbing
- `addedColumns` ALTER for `kind` runs on first start after upgrade; every
pre-existing row is backfilled to `'manga'`. No manual SQL, no down-time.
- `ALLOWED_ORIGINS` gains the two novel sites.
- New `NOVEL_USERSCRIPT_PATH` env (default `/userscript/novel-bookmark.user.js`),
bindmounted alongside the manga script.
- Traefik router names `mangabm*` → `bmapi*` / `bmweb*`.
## Test status
- `go test ./...` — green
- `node --test userscript/test/logic.test.js` — 34 pass
- `node --test userscript/test/novel-logic.test.js` — 11 pass
- `node --check` on both userscripts — clean
## Notes for the redeploy
.env keys were renamed (`MANGA_API_HOST` → `BOOKMARK_API_HOST`,
`MANGA_WEB_HOST` → `BOOKMARK_WEB_HOST`). Update DNS / Traefik labels on the
prod override before pulling, otherwise the public hostnames go dark.
See the redeploy instructions I'll post next to this PR.
Reviewed-on: #15
Co-authored-by: Sulthan Zaki <sultankiki05@gmail.com>
Co-committed-by: Sulthan Zaki <sultankiki05@gmail.com>
|
||
|
sulthan
|
180ee78b1f |
Add comix.to and kagane.to support (#13)
Tracks read progress on comix.to and kagane.to alongside asura and demonic, in both the userscript and the backend. Implements `docs/superpowers/plans/2026-08-03-comix-kagane-support.md`. ## Userscript - `comix` adapter — `/title/<id>-<slug>`; only the id prefix is identity (the slug follows the title). No `og:image`, so the cover is matched by `alt`. - `kagane` adapter — reader URLs are uuids with no chapter number, so it comes out of `og:title`; anchor scanning is structurally impossible, replaced by `latestChapterFromApi` against kagane's same-origin JSON API. - `seriesId` threaded through `latestChapterFromAnchors` so comix can scope its scan to its own series and a recommendation strip cannot win the maximum. - `@match` for both hosts, panel chips, v1.6.0. ## Backend - `latestChapterFrom` cases: comix parses the SSR JSON state blob (`latestChapterUrl`, scoped to the series id); kagane parses API JSON (`chapter_no`). - Poller allowlist extended; `Poller.BrowserFetch` with `fetcherFor(site)` routes kagane to a browser fetcher. Nil means kagane is not polled at all — never a fallback to the TLS fetcher, which would only ever retrieve a challenge page. - `BrowserFetcher`: chromedp against a `headless-shell` sidecar. kagane sits behind a Cloudflare JS challenge that no TLS fingerprint clears, and the request is made inside the page rather than by replaying `cf_clearance`. - `BROWSER_WS_URL` wiring, sidecar in both compose files (no `ports:`, dedicated non-external network), Dockerfile on `golang:1.26-alpine` — chromedp requires go 1.26. - Web UI `--comix` / `--kagane` tokens in both colour branches. ## Notes for review - `series_url` is client-supplied and a headless browser is a strong SSRF primitive, so kagane's host is pinned twice: in `fetchableSeriesURL` and again in `kaganeAPIURL`. - Three chained defects found during verification made the browser path dead under Compose (sidecar flag collision, Chrome's Host-header DNS-rebinding check, the wrong chromedp option). Fixed; the compose comments record the wrong configurations too, so they don't get "simplified" back. - `ALLOWED_ORIGINS` now includes both new origins. Without it every write from comix/kagane silently fails CORS preflight, parks in the retry queue, and drops at the cap. ## Verification 221 backend tests, 32 userscript tests, static `CGO_ENABLED=0` build, both compose configs. Two gaps, both real: 1. The userscript on live pages via Violentmonkey needs a human browser profile — not run. Check: comix series page (title/cover, no chapter), comix chapter page (records the number; an *older* chapter must not regress it), comix SPA navigation without reload, kagane series page (og:image cover), kagane reader (number from `og:title`), both chips opening the right sites. 2. The kagane browser path has not completed end-to-end anywhere. Dial/navigate/fetch is confirmed, but Cloudflare 403'd headless-shell's Chrome on every attempt from the dev sandbox, and comix's poll-through-Docker was blocked by that environment's TLS interception. Both environment-dependent rather than branch defects — the first real deploy is the actual verification. Reviewed-on: #13 Co-authored-by: Sulthan Zaki <sultankiki05@gmail.com> Co-committed-by: Sulthan Zaki <sultankiki05@gmail.com> |
||
|
sulthan
|
f3b55fd883 |
Work the design critique down: chapter format, colour law, search, strip, a11y (#11)
Two rounds of design-critique fixes on the web UI. Every visual change was verified at 390x844 and 1280x900 in both dark and light with screenshots; `go test ./...` is green throughout; no new dependencies. ## Earlier commits on this branch The two oldest commits predate this session and were never opened as their own PR, so they are under review here too: - Confirm-gate the lifecycle actions, cluster the action strip by consequence. - Fix the accessibility findings from the audit: contrast, focus, reduced motion. ## The rest **Chapter format.** The userscript and the poller both write `"Chapter N"`, and the templates prefixed `Ch ` again, so every real Asura row read `Ch Chapter 250` — while a manual edit stored a bare `250`, leaving two formats in one list. `DisplayChapter`/`DisplayLatest` on `Bookmark` now strip the lead-in and re-add exactly one `Ch `. **Zero-result search.** The client filter only toggled `card.hidden`, so a query matching nothing left a blank list under a fully populated, unfiltered "Continue reading" strip. There is now a no-match state with a Clear-search button, and the strip goes down while a filter is active. **The colour law.** `--ember` is documented as meaning "new chapter" and was spent on eight things, including setting "Nothing new." in the colour reserved for new chapters. Destruction moves to a new `--danger` token; text-input focus follows the searchbar idiom and turns `--paper`. Contrast, both themes: `--danger` on the page 4.82 / 6.65, the solid Remove button 4.94 / 7.30, the confirm question 9.00 / 7.98. **The remove confirm.** Buttons 40px 8px apart became 46px 12px apart, and the question names the series and the loss instead of asking "Remove this?". It opens with **Cancel** focused, not Remove — the two reversible rows still open on their affirmative. **The recent strip.** It was the head of the same `updated_at DESC` list rendered directly below it, on every tab, costing ~240px of the first phone screen. It is now scoped to series with a chapter waiting, and only on All. With nothing new anywhere it does not render — deliberate. **Stale chrome.** The strip and the Updated badge describe the whole library but live outside the swapped `#list`, so archiving a series left it under "Continue reading" with the badge still counting it, and `/?tab=all` reached by htmx differed from the same URL reloaded. Both regions move into `chrome.html` and refresh out of band on every mutation and every tab switch. **Accessibility and touch.** Esc closes any open panel and returns focus to the cell that owns it; opening a confirm moves focus into it; the inline error scrolls into view and no longer self-destructs after 5s; every tab and desktop action cell clears 44px; `role="alert"` on the login error; the card monogram is no longer announced; the busy bar is clipped by its own travel rather than by `overflow: hidden` on the card. **Chapter form label.** The panel's only visible text named the published chapter while the field held your progress. The field gets a real label; "Latest known" moves below it. **gzip.** Nothing was compressed. A stdlib middleware handles the four text types and leaves woff2 alone: style.css 21.8 -> 6.5 KB, htmx 50.9 -> 16.4, filter.js 7.6 -> 2.9. ## Not addressed The `role="status"` error slot is still mutated while hidden and then revealed, which is the non-announcing pattern the confirm rows were fixed for. Delete is still a silent vanish. Both are flagged in the critique snapshot under `.impeccable/critique/`. Design health went 24/36 (66.7%) to 29/40 (72.5%) between snapshots; the two P1s that survived were found and fixed after that run. Reviewed-on: #11 Co-authored-by: Sulthan Zaki <sultankiki05@gmail.com> Co-committed-by: Sulthan Zaki <sultankiki05@gmail.com> |
||
|
sulthan
|
6af49e6790 |
Archived and finished buckets, userscript nav chips (#4)
Gives every bookmark a lifecycle bucket — `reading`, `archived`, or `finished` — so on-hold series leave the main list while still being polled for new chapters, completed series get a web-only bucket, and the userscript panel gains quick links to the web UI and both manga sites.
Design: `docs/superpowers/specs/2026-07-27-status-buckets-design.md`
## Data model
One additive column through the existing `addedColumns` migration list:
```sql
ALTER TABLE bookmarks ADD COLUMN status TEXT NOT NULL DEFAULT 'reading'
```
The `DEFAULT` backfills every pre-existing row as `reading`, so there is no separate migration step. `favorite` is unchanged and orthogonal — a series can be an archived favourite.
Rollback is safe: an old binary against the new database omits `status` from its INSERT (it gets the DEFAULT) and never mentions it in the conflict clause, so buckets survive.
## The write rule
`PUT /bookmarks/{key}` decodes a whole `Bookmark` and `Upsert` writes every column it knows about. `latest_checked_at` escaped this by staying out of `bookmarkColumns` entirely — `status` cannot, because the userscript must be able to archive and restore.
So an empty incoming status means **"no opinion"**, not a value, and resolves on the `VALUES` side of the upsert:
```sql
COALESCE(NULLIF(?, ''), (SELECT status FROM bookmarks WHERE key = ?), 'reading')
```
with `DO UPDATE SET status = excluded.status`.
It has to be this way round. `excluded.*` is the row *after* the `VALUES` expressions are evaluated, so applying the default there and then reading `excluded.status` in the conflict clause would see `'reading'` rather than the empty string — and would overwrite an archived row on every progress PUT from a client that knows nothing about the column. One expression, evaluated once, covers insert and update alike. The subquery runs inside the transaction, so it sees the row the statement is about to conflict with.
`TestUpsertEmptyStatusPreservesStored` is the guard on this.
`updated_at` behaviour is unchanged: it moves only when `last_chapter_num` changes, so archiving, finishing, restoring, and favouriting never reorder the list.
## Validation
`PUT /bookmarks/{key}` returns 400 for any status outside `{"", "reading", "archived", "finished"}`, and for `"finished"` specifically. Finishing a series is a web-UI decision, enforced server-side rather than by trusting every client to leave the value alone. The `/ui/*` endpoints have their own session-guarded route and are unaffected.
## Visibility
| Surface | All | Updated | Favourites | Archived | Finished |
|---|---|---|---|---|---|
| Web | reading | reading | reading | archived | finished |
| Userscript | reading | — | reading | archived | not shown |
Archived and finished appear in their own tab and nowhere else — including the web UI's "Continue reading" strip, which is now built from reading-only rows before tab filtering. An archived favourite shows up under Archived only: Favourites means "favourites I am currently reading".
## Backend
- **`store.go`** — `Bookmark.Status`, the column in `schema` / `addedColumns` / `bookmarkColumns` / `scanBookmark` / `Upsert`. `scanBookmark` normalises anything outside the three known buckets to `reading`, so no row can land in no list at all.
- **`store.go`** — `DueForLatestCheck` gains `AND status IS NOT 'finished'`. Archived series keep being polled; that is the whole point of archiving rather than deleting. Finished ones have nothing coming, so polling them only burns fetches and risks a spurious "new chapter" badge. `IS NOT` is null-safe, so a hand-edited NULL still qualifies.
- **`handlers.go`** — status validation on `PUT`, before any write.
- **`web.go`** — `buildListView` filters the new tabs and excludes both buckets from `all` / `new` / `fav` and the recent strip; new `POST /ui/bookmarks/{key}/status`, session-guarded like its siblings, read-modify-writing through `Store.Get` + `Store.Upsert` so the `updated_at` rule stays in one place.
- **`templates/`** — two more tabs; per-card controls (reading → Archive + Finish, archived → Restore + Finish, finished → Restore); empty-state copy for both new tabs.
- **`static/style.css`** — five tabs no longer divide a phone's width legibly, so the row scrolls sideways instead of squeezing.
## Userscript (1.3.0)
- Third tab **Archived** beside All and Favourites. A missing `status` reads as `reading`, so a list cached by the previous version still renders. `finished` matches no tab and is invisible everywhere.
- Per-item **Archive / Unarchive** button on the existing optimistic path: mutate local state and cache, `apiPut`, adopt the server's returned row.
- Header chip row linking the web UI and both manga sites, each `target="_blank" rel="noopener"`.
- `apiPut` now omits `status` unless the caller opts in — see below.
Still free of every `GM_*` API: plain `fetch`, page `localStorage`, on-page UI only.
## One bug worth calling out
The userscript's other mutations (`updateToCurrentChapter`, `setChapterManual`, `toggleFavorite`, `applyLatestChapterIfChanged`) build their payload with `Object.assign({}, existing, …)`, so they echoed the cached `status` back to the server. `GET /bookmarks` has no status filter — finished rows are in `state.list` and only hidden at render time — which made two failures reachable:
1. Reading a chapter of a series marked finished sent `"status":"finished"`, which the API rejects with 400. Progress never synced, behind a misleading "Offline — saved locally, will retry" toast, permanently.
2. Archiving on desktop and then reading on a phone whose cache predated the archive sent `"status":"reading"` and silently un-archived the series — contradicting the README's "reading an archived series leaves it archived".
Fixed at the single choke point: `apiPut(key, obj, { sendStatus = false })` strips `status` from a copy of the body unless the caller opts in, and only `toggleArchive` opts in. Only an explicit archive/restore has an opinion about the bucket; everything else omits the field so the server's keep-on-empty rule applies. Stripping merely the *invalid* values would not have been enough — a stale cached `"reading"` still clobbers a remote archive.
Also: the userscript's `backgroundRefreshLatest` now skips finished series, matching the server poller, instead of spending batch slots fetching pages for a series that has nothing coming.
## Known limitations, deliberate
Both are marked in-code with `ponytail:` comments naming the ceiling and the upgrade path:
- The poller's `Store.Get` + `Store.Upsert` is not wrapped in a transaction, so a client PUT that commits between the two is lost to the stale re-read. Already documented for read progress in `CLAUDE.md`; it now costs a status change too. Accepted for a single-user deployment.
- A card whose new status no longer matches the active tab stays on screen until the next list load. The alternative is an out-of-band swap or a full list refresh per toggle, and the card visibly showing its new state is enough feedback.
## Testing
`go test ./...` passes; `CGO_ENABLED=0 go build ./...` clean.
- **`store_test.go`** — a fresh row defaults to `reading`; a legacy database gains the column with every row `reading`; an `Upsert` carrying `""` preserves the stored bucket while a value replaces it; a status change does not move `updated_at`; `DueForLatestCheck` returns archived and skips finished; the poller's `Get` → `Upsert` round trip preserves `archived`.
- **`main_test.go`** — `PUT` with `finished` or garbage is 400, `""` / `reading` / `archived` round-trip; a PUT that omits the `status` key entirely (what a pre-1.3.0 userscript sends) preserves an archived bucket *and* applies the chapter progress in the same request.
- **`web_test.go`** — each tab returns only its bucket; the recent strip excludes archived and finished; the status endpoint requires a session, rejects unknown values, and does not move `updated_at`; the card renders the right controls per bucket.
Userscript has no automated harness, so it was checked against a live `https://asurascans.com` page: the chips resolve, Archive moves a series out of All and Favourites into Archived, the state survives a full reload (so it came from the server, not local optimism), Unarchive returns it, a series marked finished in the web UI appears in no tab, and — captured on the wire — the archive PUT carries `"status":"archived"` while a favourite toggle on that same archived series carries no `status` key at all.
Reviewed-on: #4
Co-authored-by: Sulthan Zaki <sultankiki05@gmail.com>
Co-committed-by: Sulthan Zaki <sultankiki05@gmail.com>
|
||
|
sulthan
|
62772e1eaa |
feat: server-side latest-chapter polling (#2)
Adds a background goroutine to the backend that re-checks each bookmarked series' newest published chapter on its own schedule, so `latest_chapter` stays fresh even when the manga sites are never opened in a browser.
This is a *second, parallel* signal, not a replacement: the userscript keeps its own `maybeCaptureLatestOnSeriesPage` / `backgroundRefreshLatest` logic, unchanged. `userscript/manga-bookmark.user.js` is byte-identical to `main`.
## How it works
One ticker goroutine in the same binary. Each wake it asks SQLite for bookmarks whose `latest_checked_at` has aged past a per-bookmark cooldown, fetches those series pages through a Chrome-fingerprinted HTTP client, extracts the max chapter number with a per-site regex, and writes it back through `Store.Get` + `Store.Upsert`. Every failure path logs and moves on.
Two independent clocks:
- **cooldown** — how long one bookmark rests between checks, enforced by the `WHERE` clause in `Store.DueForLatestCheck`, not by a timer.
- **interval** — how often the goroutine wakes and looks.
Shortening the interval therefore cannot shorten anyone's cooldown; it only makes the poller wake and find nothing due more often.
The row is stamped **before** the fetch, so an error, a timeout, or a shutdown mid-request still consumes the cooldown — a renamed or challenged series waits out a full cooldown instead of being retried every tick.
## Design decisions worth reviewing
**`latest_checked_at` is deliberately absent from the `Bookmark` struct and from `bookmarkColumns`.** `PUT /bookmarks/{key}` decodes a whole `Bookmark` and `Upsert` writes every column it knows about, so a userscript PUT — which has no idea this field exists — would write a zero and reset the cooldown, making the poller re-fetch that series on every tick for as long as the user kept reading it. Two tests guard this: `TestUpsertPreservesLatestCheckedAt` and `TestPutDoesNotClobberLatestCheckedAt`, the latter driving a real router PUT with a userscript-shaped body.
**`updated_at` never moves on a latest-chapter bump.** All chapter writes go through `Store.Get` + `Store.Upsert`, so the existing `CASE` keeps the stored timestamp when only `latest_chapter_num` changes and the bookmark list does not reorder. `TestRunOnceDoesNotReorderList` asserts both the timestamp and the `List()` head position.
**Fetches use `bogdanfinn/tls-client` with a Chrome profile.** Plain `net/http` was verified working against both sites on 2026-07-26, so this is not fixing an observed block — it is deliberate defence-in-depth against a future fingerprint-based one. The library is pure Go, so `CGO_ENABLED=0`, the static binary, and the distroless image are all unaffected. It does require the Go floor to move 1.23 → 1.24.
**`checkOne` validates before spending a request.** `series_url` is entirely client-supplied through `PUT /bookmarks/{key}`, so without a guard the poller would issue GETs from the server's own network position to any URL a token holder writes. The check requires a known site and an `https` URL with a non-empty host, and sits *after* the cooldown stamp so an unfetchable row is retried at cooldown pace rather than hot-looping.
## Config
Five new env vars, all with defaults sized for this deployment, all wired through `docker-compose.yml`:
| Variable | Default | Meaning |
| --- | --- | --- |
| `LATEST_CHAPTER_POLL_ENABLED` | `1` | Kill switch |
| `LATEST_CHAPTER_POLL_COOLDOWN` | `1h` | Per series, floored at `15m` |
| `LATEST_CHAPTER_POLL_INTERVAL` | `10m` | How often to wake |
| `LATEST_CHAPTER_POLL_BATCH` | `14` | Series per wake |
| `LATEST_CHAPTER_POLL_STAGGER` | `20s` | Delay between fetches in a batch |
`batch × (cooldown / interval)` = 84 series hold a true cooldown cadence at these defaults. Past that nothing breaks: the cadence stretches uniformly and the oldest-checked-first ordering keeps it fair. Bad values log and fall back rather than failing startup — the poller is an enhancement, and a typo in one of its knobs must not stop bookmark sync.
## Known limitation (accepted, documented)
The poller's `Store.Get` + `Store.Upsert` is not wrapped in a single transaction. If a userscript `PUT` commits in the sub-millisecond window between the two, the poller writes back its stale re-read — reverting that progress and, since the stored `last_chapter_num` now differs, tripping the `updated_at` `CASE` and reordering the list.
Accepted rather than fixed for a single-user deployment: the window is one SELECT wide, the poller only writes when a chapter number actually changed, and the next read self-heals it. The alternative — a transactional read-modify-write — means moving or duplicating the `updated_at` `CASE` that four tests and the whole list-ordering invariant depend on. Recorded in `CLAUDE.md` next to the poller's architecture bullet so it is not a silent trap.
## Testing
- Full suite green, including `-race`; `go vet` clean; `CGO_ENABLED=0` static build and `docker compose build` both pass on the bumped `golang:1.24-alpine`.
- No test touches the network: the `fetcher` interface exists so tests inject a fake, and no test imports `tls-client` or reaches either manga site.
- Extraction is fixture-driven against markup trimmed from real pages (2026-07-26), including a Cloudflare challenge page, cross-series chapter links, decimal chapters, and both raw `&` and `&` forms.
- Poller tests cover the no-reorder invariant, cooldown enforcement across passes, batch limiting, one bad series not stalling a batch, downward correction on a retracted chapter, cancelled contexts, and all four failure shapes still consuming the cooldown.
- Migration from a pre-column database has its own test — `newTestStore` takes the `CREATE TABLE` path, so the `ALTER TABLE` path would otherwise be untested.
- **Live smoke test:** real server, real fetch of asurascans.com. Log showed `latest is now Chapter 181` and `due=1 checked=1`; `GET /bookmarks` returned `latest_chapter_num: 181` with `updated_at` byte-identical to the PUT that created the row — the no-reorder invariant confirmed against a live site, not just a fake.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Reviewed-on: #2
Co-authored-by: Sulthan Zaki <sultankiki05@gmail.com>
Co-committed-by: Sulthan Zaki <sultankiki05@gmail.com>
|