Regression from the id-label resolution (09f8f087): opening an id-kind
value menu carried the STORED ids as the typed filter, and the new
id-exact-match narrowed the menu to exactly the already-picked rows —
nothing else was visible, so "can't add more than one" (and clicking
the lone visible row unpicked it). The ids now ride as the typed
lookup ONLY in quiet mode (label resolution); a real menu open browses
(blank) or filters by the typed NAME draft.
The multi-value affordance is now ARITY-AWARE, per user request: the
op decides how many values the cell takes, and both the hint and the
pick behavior follow it — "in" says "pick one or more" and toggles
ids into the comma list; "is" says "pick one" and REPLACES the value.
Changing the operator updates the hint in place.
Pins: menu-open after a pick must not carry the stored ids (the
regression), one-arity pick replaces instead of toggling, and the
placeholder follows the op's arity. 38 filter-builder vitest cases;
web suite 183/183; container filter + episode-tree slices green.
Comma-joined multi-picks matched ZERO: clauseFeed/clauseLibrary/
clauseCollection parsed each rule value as ONE uuid, but the builder's
pick loop stores several picks as a single comma-joined value
(["id1,id2"]) — uuid.Parse failed on the joined string, the id list
came back empty, and "feeds in [A, B]" matched no episodes (one pick
worked, which is why the briefing recipe survived). parseUUIDValues
now splits every value on commas (trimming) before parsing, accepting
both the builder's joined shape and hand-written separate elements —
same leniency the suggest-side id lookup already has.
The multi-value cells now say so: id-kind value inputs get the
placeholder "pick one or more" (the only boxes on the page that accept
several values; every other cell replaces its value).
Pins: comma_joined_multi_feed_matches_both — a tree collection built
with the exact comma-joined wire the client sends, asserting BOTH
feeds' episodes match. Filter-builder vitest 35/35, container filter +
episode-tree slices green.
The library cell reverted to the raw uuid after Apply: the apply flow
reloads the page, the SSR-preloaded tree re-ingests, and idLabels —
populated only by suggestion PICKS — starts empty, so valueCellText
falls back to the id. Same on any fresh load of a stored id-rule.
Two halves:
- Server: the id-source suggest query's typed filter (library/
collection/added_by/feed) now also exact-matches a comma-separated
ID list alongside the name ILIKE — so a suggest call with the stored
ids returns their labeled rows. No new endpoint; the menu and the
resolver share one surface.
- Client: fetchSuggestions gains a QUIET mode (fills idLabels, touches
no menu state), id rows fall back to their stored ids as the typed
lookup when no draft is present, and init() calls resolveIdLabels()
after tree ingest — one quiet fetch per id-kind rule resolves every
stored id to its name from the first paint after reload.
Pins: the Go side (suggest with a library's UUID returns the row with
the library NAME as display), the client side (resolveIdLabels fills
labels, the cell renders the name, no menu opens). 35 filter-builder
vitest cases green; container filter slice green.
The previous fix was too blunt: under OR it sent NO constraint tree at
all, so "library = ebooks" stopped narrowing the author menu the
moment an OR-connected author rule appeared. The user spec is subtler:
library = ebooks AND (author = Lewis Carroll OR author = ?)
→ the menu shows EBOOK authors, with Lewis Carroll OFF the list.
suggestTree now walks root → edited rule and builds one conjunctive
chain:
- ALL-group siblings stay (AND narrows: the ebook scope persists);
- ANY-group siblings drop — they are independent alternatives under
OR and must not narrow;
- EXCEPT a same-field ANY sibling, whose picked values NEGATE into
the chain (contains → not_contains, is → is_not, in → not_in): the
picked value leaves the menu, everything else in scope stays;
- an ALREADY-NEGATED sibling picks nothing to take off, so it drops
rather than flipping positive (a positive flip would narrow);
- a field without a negation twin falls back to dropping;
- incomplete rules never constrain (not filters yet), and the edited
rule itself never constrains its own menu.
Pins: the user's exact tree (ebook scope kept + Lewis Carroll negated
off), the root-OR variant (different-field sibling drops, same-field
negates), the twinless fallback, plus the existing AND/OR cases —
34 filter-builder vitest cases green, container filter slice green
(TestFilterSuggestionsUnderRules unaffected: the server contract is
unchanged, only which qf the client composes).
Pre-existing format drift in seven web/src files (blank-line and
indent normalization by vp fmt) plus templ generate normalization
across the generated _templ.go files; both swept up while committing
the filter-builder fix. No behavior changes.
Two user-reported builder defects:
1. Id-kind value cells (library/collection/feed/added_by) rendered the
stored UUIDs in the input box. The menu offered names (s.display)
and picking stored the id correctly — but the cell bound
rule.values.join(',') raw. The cell now renders through
valueCellText: idLabels give the NAMES; typing drafts a name filter
in idDrafts (the stored values stay ids — the clause matches ids —
and are updated only by picking). Text/number/days cells are
unchanged.
2. The value menu ignored the connective: fetchSuggestions always sent
the sibling rules as the qf constraint, so with "Author contains
Lewis Carroll" present, flipping to OR still starved the author
menu down to Lewis Carroll. Under OR the siblings are independent
alternatives — a new rule's candidates must not be narrowed by
them. pathUnderAll now checks the edited rule's group path: under
AND all the way up, siblings constrain as before (helpful
narrowing); under any OR on the path, NO tree rides and the menu
browses the whole scope.
Server contract unchanged — the suggest endpoint already treats a
missing qf as an unconstrained browse; the fix is client-side
mode-awareness. Pins: 6 new vitest cases (AND sends siblings, OR sends
none, id label display, draft typing never touches stored ids, pick
toggles + relabels, text passthrough).
Archaeology, prompted by the user asking when it broke: the builder's
save NEVER worked. The page was born (2026-02-19, a92bf99a) posting
auto_assign_rules as a JSON-encoded STRING into an endpoint that has
bound []Rule since its creation (2026-01-31, 6495cc2c), plus fields
(manual_book_ids, match_type, library_id, show_on_dashboard) the
endpoint never accepted. The preview button hit a real endpoint, so
the page looked alive while Save always failed at request binding.
Verified live before this change: the page's exact payload 400'd
"invalid request" against the unmodified server.
The fix adapts the builder's payload at write time — the legacy rules
vocabulary converts to a filter tree (RulesToFilterTree restored from
the retirement commit as the single write path, unit test restored
with it):
- match_type seeds the converted tree's root mode (any default).
- manual_book_ids attach to the created collection after create
(unparsable ids skipped — the AddBooks leniency).
- show_on_dashboard persists (new create field).
- Converted trees validate under ParseFilterTree: an unhonorable rule
is a 400 at write naming the problem, never a stored 500 at render.
Dashboard sections: smart user sections now evaluate their tree LIVE —
GetUserCollectionItems merges stored membership with the composed
engine's matches (deduped, library-scoped, capped at the section
limit). This exposed a REAL production wiring bug: every production
NewDashboardService call omitted the search service, so tree
evaluation through the dashboard could never have worked (the test
harness wired it correctly — the gap was production-only and latent,
because nothing in the dashboard evaluated trees before this
migration). The dashboard handler and both inline sites now receive
the search service, and evaluation failures log instead of silently
emptying a section.
Supersedes the same-day loud-400 stance for the builder's payload
(contract_pins migration entry); custom-section-builder.md documents
the save contract; the create_collection API doc gains the adapted
fields. New TestCustomSectionBuilderSave/BadRule pin the payload →
tree → dashboard-render path end to end. Full container suite green
(155s).
The pre-tree rules machinery is deleted: no row ever carried a rules
array or a non-default match mode (verified in the dev DB), the web
never read the envelope keys, and the Android app consumes none of it.
Dead code gets no documentation.
Removed:
- PUT /api/collections/:id/match-mode — endpoint, route, and the
UpdateCollectionMatchMode query. The tree's root mode owns any|all.
- auto_assign_rules / rule_match_mode from every envelope
(create/list/detail/update), from the create/update request shapes,
and from the schema: both columns DROPPED (schema.sql carries
idempotent DROP COLUMN IF EXISTS heals for existing DBs; sqlc
regenerated).
- The rules CRUD service layer (AddRule/SetRuleEnabled/RemoveRule +
validateRule + UpdateCollectionRulesJSON), ParseRules /
EnabledRules / IsSmartCollection, RulesToFilterTree (+ the SSR
preload branch), the legacy branch of EvaluateSmartMembership, and
SmartEvalOptions.MatchMode. IsCollectionSmart = tree presence;
create-path isSmart = collection_type=smart OR a filter_tree.
Kept (live consumers): POST /api/collections/preview still evaluates
rules payloads in memory (the custom-section builder's preview), and
include_auto stays on the list (the Android app sends it).
The custom-section builder's SAVE was already broken — its
stringified auto_assign_rules payload failed request binding (probed
live: 400 invalid request). Create now rejects that payload loudly
with a 400 naming the retirement, instead of silently creating a
hollow static section; custom-section-builder.md carries the flag.
The builder's tree migration remains the flagged follow-up.
Test-side (asymmetry rule — deleted behavior, deleted tests):
TestPodcastSmartCollections rewritten onto trees (mode lives in the
tree root), match-mode subtests and tree_wins_over_legacy_rules
deleted, TestRulesToFilterTree deleted, contract-pin envelope keys
retracted per the contract_pins 2026-10-31 retirement entry. Mirrors
hand-synced: openapi.yaml, bruno create/update, docs/api/*.md,
docs/schema.go, api.d.ts. Full container suite green (155s).
collections + collection_items join the RLS-governed tables, same
doctrine as the media_items family lens (2026-10-29). Policies route
through two PURE predicate functions (app_can_read_collection /
app_can_write_collection) that never query collections — no policy
recursion under FORCE RLS — so a future sharing model is one EXISTS
branch inside the functions, not a schema re-touch. Reads: owner or
admin; writes: owner, admin escape for SYSTEM collections only.
Why: the write path had no app-level ownership gate on Update/Delete/
AddBooks — any authenticated user could mutate another user's
collection by id. The DB is now the enforcement of record; existing
app-level gates (match-mode, reorder, view-settings) stay as
defense-in-depth.
Behavior changes (contract_pins 2026-10-31):
- cross-user GET detail: 200 -> 404 (row invisible)
- cross-user PUT update: silent success -> 404 (ErrNoRows mapping;
service wrap fixed %v -> %w so errors.Is sees pgx.ErrNoRows)
- cross-user POST books: silent success -> 404
- cross-user DELETE: 204, now a true DB no-op
- anonymous sees nothing; admin persona unchanged
Bootstrap: registration seeds system collections BEFORE the JWT
exists — CreateDefaultCollectionsForUser declares the new user's
identity (rls.WithUser), the same fix mirrored in the test harness
(createDefaultCollectionsForUser owner-scoped; collection sweeps run
under the owner's identity).
Tests: new TestCollectionsRLS pins the cross-user matrix end-to-end
(404 detail/update/add/view-settings, 204 no-op delete, admin escape,
owner positive control, registration seeding under RLS). Harness
fixture writes to non-system collections now run as the row's OWNER
(scopeExecAs/ownerSeedCtx) — the admin write-escape covers system
rows only, and bare-ctx pool reads of collections see nothing.
Filter/playlist/audiobook/scan-state fixtures converted. Full
container suite green (159s).
THE CLASS: the /api JWT middleware's TokenLookup is
"cookie:token,header:Authorization:Bearer" — the token cookie always
rides same-origin fetches, so a signed-in user never needs the
localStorage bearer. But a dozen call sites gated their fetch on
`localStorage.getItem("token")` and early-returned when absent. The
global 401 interceptor CLEARS that localStorage entry on any session
rotation/blip — after which every gated write silently did nothing:
theme changes reverted on reload (the reported "can no longer change
themes", on top of the interpolation bug d5dbaa48 fixed), chain-order
selects, collection deletion, Add-to-collection submits, filter
save/load/delete, conflict actions, linking actions, stats charts,
tag suggestions, device management, admin scanner actions — all dead
for perfectly authenticated cookie users.
Fix: storage.ts exports authHeaders() (bearer when present, cookie
carries otherwise) and every gate/header site now uses it — theme.ts,
collections.ts (htmx auth, doRemoveBook/doBulkRemove, setChainOrder),
bookPicker submit, filter_builder suggestions + match count,
collection-rules apply hook, bookshelf save/load/delete filter,
conflicts x4, linking x3, admin scanner actions + polls (requireToken
removed — its "Session expired" toast fired for cookie users too),
dashboard preferences/sections, custom-section-builder, header
scanner poll, stats x3, tag suggestions, profile delete-account
(which also wrongly redirected cookie users to /login), and
device-management x4. Reads keep their graceful-empty behavior; the
bearers still ride along when present.
Pinned: the knob-matrix theme spec now REMOVES localStorage.token and
asserts the theme PUT still fires (cookie-only) and the class flips —
the exact strand that started this.
Found by noticing the panel browser's Nord click switched the class
but produced no PUT: the persistence half was gated on a token the
401 interceptor had cleared.
The empty-tree grammar relaxation (create-modal episodes seed) must not
leak into the search deep-link contract: qf={"mode":"all","children":[]}
as a QUERY would silently mean "match everything". The seed's meaning —
a rules-less collection matching nothing — lives in the stored tree and
its composers; the query entry point now rejects a rule-less tree
explicitly, preserving the pinned 400 that TestFilterQFValidationAndPrecedence
guards (the test caught the leak during this round's full-suite run).
Six specs driving the user-reported controls the way a human does,
encoding the sweep rules COVERAGE.md now states: menus/modals assert
VISIBILITY (never DOM text — hidden menus' text is why the 2026-10-30
panel sweep certified dead controls), geometry asserts the z-ladder
via elementFromPoint and viewport containment via boundingBox, every
click must produce an observable outcome (PUT fired, class flipped,
chip text swapped) and zero new console problems — the silent
CSP-throw family (theme buttons, connective chip, invisible op menu)
now fails a test instead of surviving to a user's mouse.
Covers: theme picker (interpolation regression from eb57b720),
pencil-glyph vs card-body contract, builder field menu / op menu
(full non-empty vocabulary) / connective both directions, the
Smart→Episodes create-modal seed, the picker overlay's ladder and
height, and the chain-order select's persistence round-trip.
Fixtures self-create/self-delete under test-prefixed names; localStorage
token + selectedLibrary=__all__ are seeded because the API login
establishes only the cookie; run the folder with --workers=1 (the login
route rate-limits and all suites share testuser — a parallel worker's
login revokes another's mid-run).
Three mirrors had drifted:
- The rules-endpoint retirement (4f71eda6) deleted the routes but not
their mirrors: removed docs/developer/api/collections/
{add_auto_assign_rule,remove_auto_assign_rule,test_rule,update_rule}
.md + their bruno twins, and pruned /api/collections/test-rules,
/{id}/rules, /{id}/rules/{ruleId} from the generated openapi spec
(hand-edited — regenerating with -update would wipe hand-authored
bruno suites that predate the docs mirrors).
- New endpoint PUT /api/collections/:id/view-settings gets a doc page
+ schema-registry entry + bruno mirror (repo convention: the route
and its docs land together).
- 54 pre-existing TestDocsMatchRoutes failures: the media-item
endpoints emit the 16 episode/download fields (source_feed_id,
feed_position, download_state, enclosure_url, hosts/guests, …) and
libraries emit purpose/allow_member_feeds, but the doc examples
were never extended — the package only runs on the host, so the
container suite never caught it. Examples now carry every key the
handlers emit; internal/docs is fully green.
Round-4 console hygiene: fast navigation rejects with "AbortError:
Transition was skipped" (the browser's MPA view transitions —
@view-transition navigation:auto — when a navigation interrupts a
running crossfade) and Alpine's x-transition hide rejects with
{isFromCancelledTransition: true} when a re-show supersedes an
out-phase. Both are races the user cannot lose anything in, but they
print console errors on every quick click — and the knob matrix's
zero-console-error tripwire (the thing that catches silent CSP death)
depends on a clean console. A narrow global gate swallows exactly
these two shapes; every other rejection still reports.
PUT /api/collections/:id/view-settings writes the row; it now demands
UserOwnsCollection like its sibling config mutations (SetMatchMode,
ReorderItems) — otherwise any signed-in user could flip another
user's curated listening order. Discovered while documenting the
endpoint: the sibling pattern exists precisely because the main
PUT/DELETE routes lack it (their gap is noted for follow-up, not
expanded here).
closeCollectionModal picked document.querySelector(".fixed.inset-0") —
the FIRST overlay in DOM order, which is the sitewide #confirm-modal
that the header renders (hidden, display:none, but present). The real
create/edit modal stayed open: every Close and Cancel click on the
collection modals did nothing. Caught by the knob-matrix spec — the
pencil-glyph case clicked Close and the Edit modal persisted.
The buttons now pass $el and the overlay OWNING the button is closed
(closest(".fixed.inset-0")), falling back to #modal-container's child
for non-Alpine callers. Both the collection create/edit modal and the
restore-system modal take the argument.
The debug print served its purpose: the stale-binary native-order render
was an artifact of an incomplete rebuild, not code. Fresh container
resolves view_settings.chain_order and the chain plays in the curated
order (newest-first verified on a 23-episode feed with a date tie at the
tail, handled by the position tie-break).
The scope door's first live test showed the episodes editor leading its
field menu with People (Author/Narrator/Contributor) and Publishing —
book-only fields that can never match an episode row — while the fields
the scope exists for (Feed, Downloaded, Played) sat below the fold.
Root cause: the registry's book fields carry no Scopes restriction
(nil = all scopes), a leftover from before the episode fields existed.
35 unambiguous book-only fields (people, publishing metadata, topics,
series/comics, identifiers, cover, reading-specific statuses) are now
tagged Scopes:books. Title, description, published dates, language,
file facts, explicit, the my-reading group (listening progress is
real), and collection stay dual-scope. Parse continues to accept the
hidden fields everywhere (their composed clauses simply match nothing
on episode rows) — this governs the AUTHORING surface, not the
grammar.
TestFilterParseValidation's empty-root rejection updated to the new
contract (empty root legal — the seed; nested empty groups still
invalid), and TestEpisodesScopeHidesBookOnlyFields pins both scopes'
vocabulary.
The first real-browser create of an Episodes-scope collection failed
with "invalid request": the modal's hx-post submits
application/x-www-form-urlencoded, and the form value for filter_tree
had to bind into CreateCollectionRequest.FilterTree (any) — echo's
form binder errors on any-typed fields instead of setting the string,
so every form carrying a tree 400'd before validation even ran. JSON
bodies (the save-filter bridge, the editor's Apply) were unaffected,
which is why no API-level or Go test caught it: only a browser posting
the real form shape exercised the transport.
Fix: the form variant binds a dedicated string field
(FilterTreeForm, form:"filter_tree") on both request structs while
FilterTree keeps the JSON tag; resolveTreePayload picks whichever
transport arrived and both ride the same parseRequestFilterTree
validation. Pinned by TestEpisodeCollectionsFormSeedCreate (form-encoded
seed → 201 → scope round-trips).
"Newest episodes from my selected podcasts" is now buildable in the UI,
end to end — the user could not find ANY podcast-related filter surface
because none existed: the New Collection modal could only create
book-scope smart collections, and the engine's episodes scope (item 3.1)
had no authoring door.
The scope door:
- The create modal's Smart card gains a Rules-scope segmented control
(📚 Books / 🎧 Podcast episodes); the copy no longer claims smart
means "match books by rules". Choosing Episodes seeds an EMPTY
episodes tree ({"scope":"episodes","mode":"all","children":[]}) so the
collection is born episode-scoped: the editor opens with the episode
vocabulary (feed/downloaded/played/published date) and membership
runs the episode engine from the first rule.
- ParseFilterTree now accepts an EMPTY ROOT (nested empty groups stay
invalid) — the seed's legal shape. The root arrives at depth 1 by
call convention, hence the depth>1 exemption.
- The composers define the empty-tree semantics: zero children matches
NOTHING (ComposedLiveMedia/ComposedCount guards), mirroring
EvaluateSmartCollection's zero-enabled-rules convention. Without the
guard the composed query degrades to the bare scope head and a fresh
collection would swallow every episode in scope — pinned by
TestEpisodeCollectionsEmptySeedMatchesNothing (envelope book_count 0)
and TestEmptyRootTreeSeedsEpisodesScope.
The curated chain order (plan §10 R2, user need: the briefing plays
NEWEST-first):
- view_settings.chain_order ("native"/"newest"/"oldest"/"reverse")
stored on the collection — curated content, not a player-side knob.
- PUT /api/collections/:id/view-settings merges the key into the
stored JSONB (a blind write would clobber the playlist toggles).
- buildSmartCollectionSequenceContext routes OrderEpisodes through
handlers.CollectionChainOrder; the detail header's Play-as-Sequence
gains the order select next to it (PlayCollectionButton), which
PUTs and reloads. Pinned by chain_follows_the_curated_order
(reverse flips the chain indexes and hops).
Caught by the new test, a latent clobber fixed: UpdateCollection
treated a NIL view_settings payload as "write NULL" — any edit that
carried no settings dropped every stored key (chain_order, playlist
toggles from API clients, imported provenance). Nil now means
untouched, the same convention filterTree already had.
Composer guards also cover ComposedCount, so the builder's live count
reads the same semantics. Tests: collectionChainOrder unit matrix,
empty-root parse unit pin, two integration pins (seed matches nothing;
curated order drives the chain and survives whole-map edits).
Every one of these was silent at the code level and loud in a real
browser; each fix carries its own tripwire (render pin, unit pin, or
CSP-safe rewrite) so the class cannot quietly return.
1. Theme picker dead since the CSP migration (eb57b720): the Appearance
panel's theme buttons emitted `@click="selectTheme(opt.Name)"` as
LITERAL text (templ gotcha #1 — quoted attribute values are never
interpolated; `opt` is a Go loop var), so Alpine threw "Undefined
variable: opt" on every click and no PUT /api/auth/theme ever fired.
The wood-paneling sibling always used the correct expression form,
which is why the panel half-worked. Fixed to the interpolated form;
TestThemePickerHandlersAreInterpolated pins that real theme names
reach the HTML and the raw identifier never does.
2. Filter-builder op dropdown "does nothing": the field ingest
hardcoded `ops: []` into every field's meta while only mirroring the
ops into the global label registry — opsFor() returned [] for EVERY
field, so the op menu rendered zero items: an invisible menu, no
console error (user tried author/publisher/title/subtitle, all
equally dead). FieldMeta.ops is now the wire's op objects, opsFor()
is a standalone exported helper, and opsForField unit tests pin the
round-trip. Verified live: the field menu opens/closes while the op
menu stayed shut, both with identical wiring — the difference was
the data, not the wiring.
3. AND/OR connective chip dead: the chip carried an `if` STATEMENT in
the Alpine attribute (`@click="if (row.chipClickable) ..."`) which
the CSP build's evaluator rejects as a statement — the whole
expression silently never ran. Moved the clickability check into a
toggleConnective(row) component method; the attribute is now a plain
method call.
4. Pencil vs card navigation race: navigateToCollection guarded with
`event.target instanceof HTMLElement`, which is FALSE for the
pencil glyph's SVGElement — a click landing on the glyph sailed past
the guard and navigated while the htmx edit-modal request still
fired (the reported "modal flashes, then the builder page"). The
guard now uses target.closest("button") (works on SVGElement) and
takes $event explicitly instead of the deprecated window.event.
.icon-btn is already a 36px square, so no hit-area change needed.
5. Add Items modal under the sidebar + taller than the window: the
picker overlay sat at z-50 UNDER the app-sidebar's z-60 (its left
edge rendered beneath the nav), and the card had no height bound —
header + tabs + filters + grid + footer overflow short windows with
the footer unreachable (fixed overlay, no page scroll). The picker
and create-modal overlays move to z-[65], both confirm dialogs to
z-[70], and the picker card is now a flex column capped at
max-h-[85vh] whose grids flex-fill (they grow on tall windows,
shrink on short ones). The ladder is documented at .app-sidebar in
input.css: sidebar 60 / overlays 65 / confirm 70 / toasts 9999.
Found by clicking every control the way a user does (panel clicks +
computed-style reads + screenshots) after the previous sweep trusted
DOM-text snapshots, which show hidden menus' text and cannot see
occlusion, stacking, or geometry at all.
Marks the sweep-candidate list as executed (every knob driven in a real
browser), preserves the four defect classes the sweep fixed so the next
agent does not re-learn them the hard way, and pins the panel-vs-
Playwright limitation (value assignment without change events) that any
future automated spec of the htmx-triggered controls must work around.
The real-browser sweep (the gotcha classes' native habitat — CSP refusals
fail SILENTLY at build/test time) found four defects, all in the item
3.3/3.4 surfaces:
- Shared builder rows used Alpine attribute expressions that the CSP
parser rejects outright: backtick template literals in :style, the
?. optional chain in :placeholder/x-if, and an if-STATEMENT in a
@click.outside. Rewritten as string concat plus two component methods
(fieldIsDays, dismissSuggestions). Until this fix every drawer open
threw a wall of "CSP Parser Error" console noise and the days-value
affordance never rendered.
- "2 bookss match": the count noun is interpolated into "<n> <noun>s
match", so callers must pass the SINGULAR ("book"/"item").
- The rules editor page received its fields/tree payloads from the
route as already-JSON-encoded strings and wrapped them in JSONArg
AGAIN — double-encoded, so JSON.parse yields a string and the editor
loaded with NO vocabulary and NO prefill. The route now emits raw
JSON into the data attributes (single encode).
- Apply was silently inert: the hook store round-trip used
window.Alpine?.store (optional-chained, no-op when absent) and the
Cancel/Apply block sat outside the builder island scope for part of
the session. Both fixed via the imported Alpine instance and moving
the buttons inside the island; Apply now PUTs the tree end-to-end
(verified live: prefill → edit → apply → 201 → reload → membership).
Browser-verified clean (no fixes needed): picker scope tabs + episode
grid (?scope=episodes door), selection store across swaps, Add Selected
episode membership, library-switcher partial contract (episodes visible
under All Libraries, correctly filtered out under a book library — and
now never OFFERED as a switcher option), episode cards + reorder arrows
persisting position, collection-chain player (Ep N/M, prev/next carry
?queue=collection:, order control absent in collection mode and present
+ persistent in feed mode both in player and feed page), and the
empty-tree Apply guard (400 — a static collection can never be
promoted by accident).
Notes for the next reader of these lines: the OpenChamber panel's
browser.select sets select.value without dispatching change, so htmx
change-triggers must be exercised by real events (Playwright) rather
than the panel; and one 500 was observed on the player's leave-page
progress PUT — a request whose context the navigating browser cancels
mid-write, pre-existing and benign (the immediately following save
returned 200).
The D13 podcast exclusion (9a148c23) was written as a templ
`if !IsBookLibrary(...) { continue }` — but templ markup control flow
has NO continue STATEMENT: codegen emitted the word as a TEXT node
(visible junk between options on every page's library subbar) and
rendered the podcasts option anyway, defeating the exclusion. The real
browser sweep caught what every Go/API test structurally cannot:
generated markup that renders wrong yet compiles clean.
Rewritten as an inverted guard (the idiomatic templ skip). New render
pin TestLibrarySwitcherExcludesPodcasts fails on both halves of the
original leak — the literal text AND the presence of the podcasts
option — so this codegen gotcha class stays tripped-wired.
Findings from the systematic gotcha-class review before the browser
pass:
- The collection smart editor's Apply derived the collection NAME from
document.title — which parses to "Smart Collection", RENAMING the
collection on every apply. The server now renders the real name
(data-collection-name) and the write path refuses to proceed without
it.
- TestRenderingPartialsRender pinned the pre-episode-collections
empty-state copy ("No books in this collection yet.") — updated to
the shipped label ("Nothing in this collection yet."; collections
hold episodes now).
- The deleted handler-local rule comparators (evaluateRule/
compareValues — removed in item 3.4's single-evaluator refactor)
left behind two unit-test files testing the duplicates; deleted.
Their semantics remain covered at the service level
(TestEvaluateRules_* over FieldMatches/EvaluateRules) and
TestTextToString (which actually pins utils.PgTextString) stays.
Everything else reviewed clean against the gotcha classes: templ
whole-expression attrs, ?={ boolean attrs, JSONArg quoting in
x-init/data attrs, Alpine CSP store paths, htmx form-encoding, the
" assert escaping, and the aria-label selector contract on the
player's reload listener.
The bookshelf's save modal grows a second action: "Save as Collection"
POSTs the current (completed, scope-stamped) tree to
/api/collections — the same validated write path every tree
collection uses — and lands you on the new collection's page. The qf
engine then drives it everywhere: the shelf's live shelf-count, the
collections page, the replay chain for episode trees. The legacy
"Save Filter" (bookmark the wire params) stays as the sibling action.
This is the plan's bridge: a book search graduated into a live shelf
with one click, authored in the ONE filter system.
The smart-collection editor at /collections/:id/rules now hosts the
SAME builder island as All-Books (FilterBuilderRows — the rows markup
extracted into one shared templ both mounts render; the shared
filterBuilderExtra component). One vocabulary, one engine, one row UI;
the collection differences are exactly three: the payload is the
collection's scope (episode trees get the episode fields — books
trees hide them), Apply writes the tree to the collection (PUT
filter_tree through the validated write path), and a legacy rules
array preloads as its CONVERTED tree (RulesToFilterTree rendered
server-side) so the upgrade lands at the first Apply — read-compat
never mutates data silently.
Retired with the page: the per-rule CRUD endpoints and test-rules
(their handler bodies, routes, api-docs rows, and the dead collections.ts
island functions — the file's LIVE halves, the collection modal/navigate/
grid-refresh, stay; the dead rules half went). match-mode endpoint and
preview survive (legacy channel + the dashboard custom-section-builder,
flagged for its own migration). Contract-pin changelog documents the
deletions per the asymmetry rule.
The builder component gains collection mounts: scopeKey (the row
universe for counts/apply), data-tree prefill, treeJSON (the apply
wire with scope riding the top level), and an apply hook (the page
owns the write; All-Books never sets one). ensureReadyRule defaults to
the scope's first field instead of hardcoding author.
Tests migrated onto trees: TestSmartCollections' live-membership /
reject-mutations / include-auto / system-collection subtests now drive
TREE collections; the deleted-endpoint subtests die with them.
"One filter system to maintain" now holds at the presentation layer
too: the field/op list the builder renders comes from
services.FilterFieldsPayload(scope) — the Go registry's wire face,
scope-filtered and group-ordered — rendered into the bookshelf mount
as a data attribute (templates.JSONArg), ingested once by the Alpine
component. The registry carries operator presentation (labels,
negation, arity) in FilterOpWire alongside the grammar; per-field
value-cell semantics (valueKind) derive from the existing Kind.
The hardcoded TS mirror (OPS, op(), F(), TEXT_OPS/NUM_OPS/DATE_OPS,
FIELD_GROUPS, its known drift surface) is deleted; filter_builder.ts
now ingests the wire (ingestFilterFields — malformed payloads leave
empty menus rather than client-invented rows), and the field menu's
episode fields are HIDDEN on the All-Books page for free (the books
scope payload excludes them — the plan's "hide podcasts from the
rules on the all books page" is now structural, not enforced by a
second hand-kept list).
Unit tests ingest a minimal wire fixture (the payload shape is the
contract, not the mirror); vitest 167 green, tsc + vp clean.
collections gains filter_tree JSONB (idempotent in-place; same qf DSL
as saved filters — the scope key included). IsCollectionSmart is the
ONE detection (tree OR legacy rules), and EvaluateSmartMembership is
the ONE evaluator: tree-bearing collections run the composed engine
via ComposedLiveMedia (full media rows, so the collection surfaces
keep their episode-card stamping, feed positions for the replay
chain, and the family lens rides the mandatory head), while legacy
rule arrays keep per-item FieldMatches until rewritten — one call
surface, two storage epochs. All smart surfaces (detail JSON, counts,
detail page, reader chain) dispatch through it; the search service is
injected into the collection subsystem (variadic — main.go builds one
shared instance; the test harness mirrors it).
Write path: CreateCollection/UpdateCollection accept the tree,
validated against the engine grammar at the HANDLER (an invalid tree
is a 400, never a stored render-time 500) and exposed in the envelope
(pin widened in TestContractPin_CollectionsJSON + changelog). Update
semantics: an edit form that doesn't send a tree keeps the stored one.
RulesToFilterTree converts legacy rules + match mode at upgrade time:
episode rules map mechanically (feed→in-list, played/downloaded→
booleans, published_within→published_date+within_days), tree root
mode absorbs rule_match_mode, episode rules force the episodes scope,
and legacy starts_with/ends_with fall back to contains (documented).
Read-compat by design: nothing is migrated at boot — legacy arrays
evaluate exactly as before until the collection is rewritten.
Pinned by TestRulesToFilterTree (unit conversion) +
TestEpisodeCollectionsFilterTree (a tree collection evaluates through
the engine, the chain follows the tree's live membership, invalid
trees 400 at write, the tree wins when legacy rules coexist, and an
edit without a tree keeps it).
The qf tree gains an optional top-level `scope` ("books" default |
"episodes" — absent = books, so every existing saved tree and legacy
client stays valid). The row head is scope-aware AND this closes a
real leak: the composed engine's head had NO podcasts exclusion, so
episodes already surfaced on All-Books qf queries (plan §9 violation —
the sqlc flat head excluded them, the composed head didn't). Books
scope now says library_type_name <> 'podcasts'; episodes scope is
source = 'podcast_episode'.
The registry fields declare their scopes (the plan's "reusable
template with conditionals" = scope-driven rendering in 3.3): episode
fields declare exactly the episodes scope and the books scope REJECTS
them at parse (the "hide podcasts from the rules on the all books
page" rule enforced server-side); books fields inside episodes trees
are allowed and honestly no-op on NULLs.
New fields (the item-2.4 rule vocabulary expressed in SQL):
- feed — KindIdentifier over podcast_feeds (RLS-governed suggestion
source and match; unsubscribed feeds match nothing, by the engine)
- downloaded — is_true/is_false on download_state = 'downloaded'
- played — THE canonical finished predicate (viewer progress >= 0.999,
the hide_finished definition shared with item 2.4), reading_progress
join per viewer
- published_date — clauseDateFactory like added_at, within_days
default op (the published_within rule's SQL form)
Engine hygiene: the stray ComposedSearch SQL debug printf is gone.
Pinned by TestFilterScopeValidation (unit: scope grammar, books×
episode-field 400s, legacy no-scope trees) and
TestFilterScopesAndEpisodeFields (integration: books head excludes
podcasts including the 404-no-results contract, episodes head is
episodes only, feed-in/downloaded/played/published-date evaluate in
SQL against a subscribed-feed fixture with a played viewer).
TestContractPin_CollectionsJSON tripped on the additive keys the
episode-collections work introduced — the tripwire worked exactly as
designed (additive keys must be whitelisted here AND in
docs/contract_pins.md first).
- Collections envelope gains rule_match_mode (item 2.4): the
smart-collection match semantics on GET /api/collections/:id.
- BookInfo gains four STABLE keys (item 2.3): is_episode,
feed_title, download_state, hidden — dropped omitempty so every
card surface emits the same key set (the viewer_percentage
precedent; adapters that predate episode awareness emit zero
values).
docs/contract_pins.md changelog entry written. Nothing removed,
nothing renamed.
ACCEPTANCE (plan §0 item 2.5): "Daily Fiction" = feed ∈ {A} ∧ unplayed
∧ <7d matches exactly today's episodes, and PLAY = today's episodes
back-to-back. Two pieces were missing after item 2.4:
- The chain builder only read static membership (collection_items) — a
smart collection has none. buildSmartCollectionSequenceContext now
evaluates the rules LIVE for the requesting user (mode from the
collection, family lens via the visible set) and chains the
downloaded episode matches in NATIVE feed order — the collection's
computed order is content (R1: never written back; OrderEpisodes is
the order chokepoint per R3). A match that aged out of the briefing
(played since, published beyond the window) falls back to the feed
queue instead of breaking the page.
- The smart-detail branch rendered match cards as books: episode
matches are now flagged (IsEpisode + download state) so they carry
the play affordance and the header's Play-as-Sequence button;
feed titles stay empty on smart cards for now (batch lookup is
polish, not acceptance).
Pinned by TestPodcastSmartCollections' acceptance subtests: the
briefing starts at index 1/2 and chains into the second with the
queue param on both hops; the order control stays hidden (computed
order); a stale match falls back to the feed queue honestly.
The query said `SET view_settings = $2, updated_at = NOW()` but the
collections table has never had an updated_at column — every call
failed with 42703. The one caller (media_scanner's series-import
provenance write) discards the error, so the failure was silent:
imported-from-series view settings never persisted. Surfaced while
writing the sibling UpdateCollectionMatchMode query for item 2.4 (the
item-2.4 commit message prematurely claimed this fix — this is that
fix, for real).
The daily briefing needs AND: today's rules engine was OR-only. The
match semantics become a per-collection column
(collections.rule_match_mode, 'any' legacy default | 'all'), flipped on
its own endpoint (PUT /api/collections/:id/match-mode, owner-only —
nothing else on the row changes) and surfaced in the rules page as an
ANY/ALL toggle with matching copy.
Episode rule fields join the vocabulary (plan item 2.4): feed (the
multi-select rides the new `in` operator — comma-separated feed ids),
played (the canonical finished predicate — viewer progress ≥ 0.999,
same as hide_finished), downloaded, published_within (N days), plus
title/description contains. The rules candidate query now carries the
viewer's progress (LEFT JOIN reading_progress) so played/unplayed is
per-user, and TestRules/preview speak the SAME evaluator: the handler's
duplicated field switch is deleted and delegates to
CollectionService.FieldMatches — one vocabulary, one semantics.
Legitimate semantics note: in ANY mode `played:false` matches unplayed
BOOKS too (everything with no progress row) — correct per-rule
behavior, and exactly why the briefing mode exists.
Backward compatibility: every legacy row defaults to 'any' (the old
OR), UpdateCollection keeps the stored mode when the form doesn't send
one, and the evaluator treats absent data honestly (no date → no
match; no progress → unplayed).
Pinned by TestPodcastSmartCollections: the AND-mode briefing
(feed∈{A} ∧ unplayed ∧ <7d = exactly the new episode), OR-mode
widening, owner-only flips, the preview honoring the mode, and
downloaded/contains rules. Also fixes a latent bug the new query
surfaced: UpdateCollectionViewSettings referenced a nonexistent
updated_at column (never exercised until now — the new sibling query
was written without it).
D11 lifts: podcast episodes (media_items rows) join collections, and
"Play collection as sequence" feeds the player from the collection's
order. No new tables — collection_items already carries ordered
membership (item 2.1 audit).
PICKER: the collection modal gains a Books/Episodes tab pair. The
episodes tab hits /api/media-items/search?scope=episodes — a dedicated
picker query (SearchEpisodesForPicker) whose visibility is entirely
RLS: the media_items subscription predicate drops unsubscribed feeds'
episodes, and the podcast_feeds join is governed the same way. The
books scope keeps its podcasts exclusion (plan §9) — this scope is the
only door in. Pending episodes are pickable (plan ahead; the detail
page shows them as not downloaded). Selection rides the existing
bookPicker store — ids never collide across scopes, and AddBooks works
unchanged on episode ids.
DETAIL: the All-Libraries membership view switches to
GetCollectionItemsDetailed — LEFT JOIN + per-row visibility. An
unsubscribed episode (or family-lens hidden book) survives as a
Hidden card rendered as an "Unavailable — remove entry" hint row
instead of silently vanishing; visible episodes render feed title,
download state, and a Play link. The library-scoped view keeps its
existing dashboard-query behavior (the switcher's contract is
untouched for books).
CHAIN: ?queue=collection:<id> on any reader URL makes
buildPodcastSequenceContext produce the COLLECTION context (§10 R2
same shape, source=collection): downloaded episodes in collection
order (positions are content — R1; no feed-order override), prev/next
URLs carrying the queue param so the chain survives hops, and
backToDetail landing on the collection page. Ownership gates the chain
— foreign/deleted ids fall back to the feed queue; the order control
stays hidden (2.2's source gate). The header's "Play as Sequence"
button starts from the first downloaded episode in list order.
Pinned by TestPodcastEpisodeCollections (7 subtests): RLS-scoped
picker + family-lens auditor case (subscribed episode in a hidden
library must not reach the picker), books-scope exclusion intact,
hint rows, chain carrying the queue param, start-from-first, foreign
fallback, and reorder driving the chain (the R4 primitive works on
episode ids unchanged).
Foundation rule R2 (§10): the queue context is the only queue, and
every producer emits the same shape. PodcastSequenceContext gains
source (feed|collection), feed_id, and order; the feed producer sets
them now — the collection producer (2.3) slots into the same shape
without the player ever special-casing beyond showing/hiding controls.
The player header gains the FEED-mode order control next to the
"Ep N/M" counter: a dropdown (native/reverse/oldest/newest) plus a
one-click reverse button. Both PUT the existing per-user feed order
override — the SAME state the feed page writes, so flipping it in the
player updates the feed page's dropdown too. The control renders only
when source=feed: a collection queue's order is curated content
(R1), never a listening preference. Behavior is hx-put +
hx-swap="none" (response discarded) with a module-level
htmx:afterRequest listener in audio.ts reloading the page — the
server re-renders prev/next/index in the new order; templ decides
what it looks like, JS only triggers the round trip (CSP-clean, no
inline handlers).
Pinned by TestPodcastPlayerOrderControl: the config JSON carries
source/feed_id/order/episode_index; the control renders feed-mode;
an order override re-renders with the dropdown reflecting it (data-
json pins must match templ's " attribute escaping).
reader_settings.setting_key was VARCHAR(50), and the composite
podcast_order:<uuid> key already sits at exactly 50 chars — zero
headroom. The deferred queue-editor idea (plan §0 item 2b) and any
future per-feed/per-collection preference key would have been forced
to payload-stuff inside the order row by a column width. Composite
prefix:<uuid> keys are the norm now (§10 R5), so the ceiling widens
to VARCHAR(120): idempotent in-place ALTER (varchar widening is
metadata-only, no table rewrite), re-runs harmlessly on every boot
per the schema convention.
No new episode-membership tables: collection_items already carries
ordered playlist membership (position, A2c) against media_items,
which podcast episodes ARE — episodes join collections through the
existing table.
Pinned by TestReaderSettingsCompositeKeyHeadroom: a 57-char composite
key (the podcast_custom_order:<uuid> shape) upserts and reads back.
Under the old width this fails; sqlc codegen is byte-identical.
The app-shell sidebar read as too wide against the content area.
--sidebar-w is the single source of truth for the shell geometry, so
one line moves all three consumers together: the .app-sidebar width,
the body:has(.app-sidebar) content offset, and the topbar's left
inset on desktop. No templ or JS change needed — nothing else pinned
16rem.
style.css is the regenerated tailwind build (build:css:prod), diff
limited to the one var; committed with its source per the generated-
artifacts rule.
The admin endpoint DELETE /api/media-items/:id predated podcasts and
would delete any media row — including podcast episodes — if hit
directly. Invariant 2 (absolute form, user decision 2026-10-30) says
episode rows are removed ONLY by whole-feed removal: "the feed should
match the XML until the entire feed is gone." The plan audit found
this endpoint was the one remaining surface that could violate it:
the poller, unsubscribe, and the admin file-deletion endpoints are
files-only, and the hash-conflict merge can't see episodes (they
never hash).
The guard sits right after the GetMediaItem read and BEFORE the
storage-down check, so a refused episode never touches storage
resolution. Same segregation pattern as the generic edit path
(D9/D10, UpdateMediaItem).
Pinned by TestDeleteMediaItemPodcastGuard: admin DELETE of an episode
→ 409 with the row surviving; a book delete through the same endpoint
→ 204 control.
The bulk 'Delete selected files' button posted with hx-swap='none',
which makes htmx DISCARD the response — and the response is the
refreshed episode list partial. Rows (Play → Download, badges) stayed
stale until a manual refresh. Target the swap at #episode-list /
outerHTML like the Download All button; the afterSwap listeners then
reapply the (now-empty) selection and any active filter.
The plan's per-EPISODE bulk select for the admin view (§0 item 8
follow-up — trimmed in the first admin-view commit; user called it
back). Admins now tick checkboxes on DOWNLOADED episode rows and delete
their files in one request:
- Checkbox per row, rendered ONLY for admins + downloaded episodes
(nothing to delete otherwise), via the existing podcastDetail Alpine
component: selection ids/count survive the list's htmx swaps
(reapplied on afterSwap — same pattern as the text filter) and clear
after a successful bulk delete.
- Toolbar segment outside #episode-list (survives swaps): 'N selected'
+ 'Delete selected files' (hx-include gathers the CHECKED boxes);
disabled-looking at zero selection; hx-confirm states the
re-materialization contract. CSP-safe method bindings only
(bulkSelectionLabel/Class) — no ternaries in Alpine expressions.
- POST /feeds/:id/episodes/delete-files (admin): repeated
episode-file-select keys parsed via ParseForm directly (never bound —
repeated-key wire is another echo-bind gotcha); each id validated to
belong to THIS feed, so a forged foreign id is refused while the real
selections apply; rows reset to pending, files removed.
- Tests: bulk of two real ids + one forged id → deleted:2, files gone,
rows pending; non-admin 403; all prior single/feed-level pins hold.
§0 item 8 (Phase-4 item 1 pulled forward, user directive: maintenance
tooling ships WITH the features it maintains):
- /admin/podcasts (AdminMiddleware): every feed from EVERY user via the
RLS admin escape (first-admin identity, poller/fetcher contract) —
feed + host, library, subscriber count/names, downloaded/total, DISK
USAGE, last fetch. Sorted disk DESC ('what's eating the disk'),
totals header (feeds · episodes · bytes on disk).
- Per-feed actions: Refresh · Delete files (rows → pending, disk freed;
confirms spell out the re-materialization contract) · Remove feed
(rows + files for everyone — the same admin delete as the detail
page). All confirms ride the global confirm modal.
- New sqlc: ListAllPodcastFeedsAdmin (correlated aggregates),
ResetPodcastEpisodeFileDeleted / ResetPodcastFeedFilesDeleted —
downloaded rows reset to pending with download-derived facts cleared
(audio_metadata, chapter_count, cover, size, mime); file_path KEPT
(fetch-maintained layout — a re-download lands on the same path).
file deletion via the existing DeleteFeedFiles file-removal pattern
(episode files + .cover.jpg siblings).
- Poll upsert never touches download_state, so deleted files STAY
deleted — a feed refresh re-lists metadata but re-materializes
nothing (pinned by TestAdminDeleteEpisodeFiles).
- Per-episode admin delete-file: POST /feeds/:id/episodes/:eid/
delete-file (admin) — one episode's file gone, row pending, siblings
untouched; HX swaps the episode list. (Found while testing: the
handler first read c.Param("id") — the FEED's uuid — instead of
:eid; the 404 'episode not found' trail led straight to it.)
- Tests: admin sees other users' feeds + subscriber names; regular
users 403 on page AND actions; single + feed-level deletion reset
rows/clear facts/delete files; refresh leaves pending rows pending.
- Admin nav gains 'Podcasts' under Administration.
Wrong-gotcha note for future agents: parseUUID on c.Param('id') inside
a /feeds/:id/episodes/:eid route resolves the FEED, not the episode.
User design (round 3): one set of big buttons does the big navigation.
The outer prev/next step buttons were chapter-only even on podcasts,
while episode nav hid as small icons in the top-left cluster. Now:
- Podcasts (sequence context present): the outer buttons navigate
EPISODES via goToPrev/goToNextEpisode (progress pushed before the
hop); disabled ghost buttons at the first/last episode keep the row
shape. Embedded-chapter jumps stay on the Chapter pill's list.
- Books: unchanged chapter stepping.
- Keyboard (h/l shortcuts) and hardware media keys (MediaTrackPrev/
Next) funnel through stepChapter, which is now sequence-aware —
headset next-track = next episode on podcasts.
- The top-left episode icons are gone; the 'Ep N/M' counter stays.
- Help-modal labels say 'chapter, or episode on podcasts'.
Same icons both modes (extracted audioStepPrev/NextIcon partials);
books are untouched — no sequence context, byte-identical behavior.
User report: Remove Feed still used the browser dialog. Headless-proven
root cause (tests/ui/confirm_probe.spec.ts): the htmx:confirm handler
read evt.detail.element, which htmx 2 does not provide (its detail is
{ elt, question, issueRequest, ... }) — the handler returned at
'if (!elt) return' BEFORE preventDefault, so every hx-confirm sitewide
fell through to htmx's native confirm. The acbcc61a migration only ever
worked for direct confirmDialog() callers; hx-confirm interception was
dead on arrival against htmx 2 and no test watched the JS side.
Fix: read detail.question (htmx's own closest-attribute resolution of
hx-confirm); return without preventing default when absent. Probe
spec pins: click Remove Feed → #confirm-modal opens with the message,
cancel hides it, no dialog event fires.
Companion fixes in the same user-test round (one commit, one defect
family — templ boolean attributes + confirm-modal placement):
- Add-modal Download order never revealed: #podcast-download-all was
nested in a wrapper div, so ':checked ~ .podcast-add-order' had no
sibling relationship. Input, label, and order row are now siblings.
- Order/auto-download dropdowns always DISPLAYED the last option
('Newest first'/'Newest') whatever was saved: templ selected={ bool }
emits selected="false" — presence selects, so every option was
selected and the browser honored the last. Swept to selected?={...}
across podcast_detail (14), book_detail_modals (10, incl. two
always-checked checkboxes), collection_modal (2, same). Saves were
correct all along; display only.
- Global confirm modal moved from SidebarSignIn (signed-out only!) into
Header — signed-in pages had no #confirm-modal, so confirmDialog
degraded to window.confirm even with a working interceptor.
Facet bar scale (user question): values now carry feed counts, sort
frequency-first, cap at 6 chips per group with a native <details>
'+N more' expander.
Probes: tests/ui/confirm_probe.spec.ts + add_order_probe.spec.ts
(live-server, BH_FEED-gated) — both pass against the dev instance.
1. Add-modal 'Download order' never revealed: the CSS sibling selector
(#podcast-download-all:checked ~ .podcast-add-order) can't fire —
the checkbox was nested inside a wrapper div, so the select row was
no SIBLING of the input. Restructured: input, label, and order row
are now siblings under one flex-wrap wrapper.
2. Order + auto-download dropdowns always displayed the LAST option
('Newest first'/'Newest') regardless of the saved value: templ's
selected={ boolExpr } compiles to selected="false" — and the mere
PRESENCE of the attribute selects the option, so every option was
selected and the browser honored the last. Sweep to the boolean-
attribute form selected?={...} (omits on false) across
podcast_detail (14), book_detail_modals (10: manga type/reading
direction selects + abridged/explicit checkboxes — same defect,
always-checked), collection_modal (2: hide-finished/show-up-next
always-checked). Values were SAVING correctly all along; this was
pure display. audio_reader/collections :selected/:checked are
Alpine bindings — unaffected.
3. Remove Feed (and every hx-confirm) still showed the browser dialog
for signed-in users: the global confirm modal rendered at the end
of SidebarSignIn — signed-IN pages have no SidebarSignIn, so no
#confirm-modal existed and confirmDialog() degraded to
window.confirm. Moved the modal into Header proper (rendered on
every page, after </aside>).
4. Facet bar at scale: values now carry feed COUNTS, sort by frequency
DESC then alphabetically, cap at 6 chips per group flat, and the
rest collapse into a native <details> '+N more' expander — hundreds
of hosts/tags degrade gracefully instead of a wall of chips.
Verified live on the dev instance (order select reflects the saved
override, override row persists, reveal fires, modal renders on
signed-in pages). Podcast + logout + media-pin suites green; vitest
167 green.
The contract pin caught the §0 item 4 column surfacing on media-list
raw sqlc rows — its job. Widened additively per doctrine: rawRowKeys +
docs/contract_pins.md changelog entry. NULL for scan rows and
never-queued episodes; nothing removed or renamed.
TestSettingDefaults enumerates registry types; 'float' hit the unknown-
type branch. Added float validation (default parses, within [min,max])
mirroring the int case.
Pocket Casts steal pulled forward from Phase 4 (§0 item 6): the episode
list gains a client-side 'Filter episodes…' box and a 'Hide
played/downloaded' toggle. Pure presentation — rows carry templ-rendered
data-* attributes (downloaded, listen-pct, search text); the filter
toggles DOM visibility with the 99.5% finishedThreshold. The toolbar
lives OUTSIDE #episode-list so its state survives the list's polling
swaps, and an htmx:afterSwap listener re-applies the active filter to
freshly swapped rows. No query change, no server round trip, no new JS
rendering surface (Alpine component method, CSP-safe).
Download All previously queued in feed-native order — 'getting the
last chapter first doesn't help me'. Now the queue follows the feed's
effective listening order (§0 item 5b): QueueFeedDownloadsAs reads the
CALLER's order override and runs episodes through services.OrderEpisodes
(same helper as the sequence/reader/page surfaces — the consolidation
from item 5a landed with the reverse-order commit). Queue tickets
(SEQUENCE-stamped, item 4) make that enqueue order the claim order, so
chapters land in the order you'll actually listen.
Add modal (§0 item 5c): with 'Download all episodes now' checked, a
'Download order' select appears (pure-CSS :checked sibling reveal — no
JS island, CSP-clean). Its choice is the feed's INITIAL order override
(reader_settings upsert; 'feed' = no row) AND the queue order for the
initial batch. Per-book by design: the override IS the download order.
Test TestPodcastAddModalDownloadOrder: add with download_all+reverse →
the override row exists AND persisted tickets run position 3→1. Tickets
are asserted (not live queued rows) because the shared dev server's
worker legitimately races this test's queue — tickets survive every
state flip.
Replaces the sequential worker with a download-manager-style pool
(§0 item 4). Model: the user's connection is the limit; semantics are
fair-share ACROSS feeds, not goroutine-per-book (a global FIFO would
make Book B wait behind Book A's 100 files).
- Atomic claim: GetNextQueuedEpisode + SetPodcastEpisodeDownloading
was a read-then-write race under concurrency; ClaimQueuedPodcastEpisode
converts queued→downloading in ONE guarded statement (two claimants,
one winner — the loser gets ErrNoRows). SetPodcastEpisodeDownloading
retired.
- Queue tickets: download_queue_pos BIGINT fed by a Postgres SEQUENCE
stamped at enqueue — the enqueue loop's ordering IS the claim order
(the foundation §0 item 5 builds on; previously all rows in a batch
shared one updated_at, so within-feed order was arbitrary).
- Pool: up to podcast_download_concurrency (default 4, live range 1–16)
claim-loop goroutines per drain round; the setting re-reads every
round so admin changes take effect without restart.
- Fair share: claim prefers the feed least-recently-served
(pickFeedToServe; never-served first), earliest ticket within a feed —
books proceed in parallel while chapters land in order (start
listening while the rest download).
- Bandwidth: shared token bucket across all workers
(podcast_download_speed_mbps, default 0 = unlimited); take() reserves
bandwidth and callers sleep outside the lock.
- Settings: new float setting type end to end (registry defaults +
clamping, save validation, admin UI number-with-step input) + the two
schema seeds; worker wires the registry via SetSettings (tests run on
registered defaults).
- Tests: unit coverage for the bucket reservation math + fair-share
pick; integration TestPodcastDownloadPool pins ticket order = enqueue
order, claim exclusivity, and a full 3-episode pool drain.
Dev-DB note: the column/sequence/settings were applied live (idempotent
DDL, same statements the schema runs at next boot) since the running
container predates them.
The 'Oldest first' order sorts by PUB DATE, which breaks on Scribl's
migrated podiobooks feeds (messy/all-equal dates). New order value
'reverse' = feed position DESC with a date-DESC tiebreak — dates are
not involved, so nothing can break.
- services.OrderEpisodes: THE single order source of truth (§0 items
3+5a). The four surfaces (sequence JSON, reader chain context, SSR
page list) previously carried three near-identical handwritten
comparators — all now route through one generic helper. The SSR page
list gains a SortDate key (lexical 2006-01-02) so the pub-date
tiebreak is actually chronological; the old comparator compared
'Jan 2, 2006' display strings alphabetically (latent drift bug).
- Dropdown becomes four options: Feed order / Feed order (reversed) /
Oldest first (pub date) / Newest first (pub date).
- Serial suggestion button renamed 'Reverse feed order' (sets reverse,
not oldest) and its response out-of-band-swaps the order select
(PodcastOrderSelectOOB — same mechanism as Download All's policy
select), so the dropdown mirrors the server state immediately.
- Validation: oneof=native reverse oldest newest everywhere.
- Tests: reverse binds via the form wire; messy-dates sequence
returns 3,2,1; existing reorder-safety stays green.
Reproduced the §0 item 2 suspect: echo v5 bool-binds form values via
strconv.ParseBool, which accepts "true"/"1" but REJECTS the
browser-native <input type=checkbox> value "on" — and a bind failure
kills the WHOLE AddFeed request (400 "invalid request"), not just
the checkbox. The add modal posts urlencoded (htmx), so any
checkbox-serialization that emits "on" made Download All silently
unusable.
Fix: FlexBool (encoding.TextUnmarshaler-style UnmarshalParam) binds
download_all permissively on every wire — "true"/"on"/"yes"/"checked"
=> true, explicit false/0/off => false, ABSENCE (unchecked box on any
wire) => false; JSON bool binding is native and unaffected. Tolerates
the exact wires htmx and browsers emit.
AddFeed's subscription flip now logs "[PODCAST] flip ... <feed> —
user" so the live flip is provable in app logs (user suspected it
never fired).
Integration test TestPodcastDownloadAllFormEncoding pins the urlencoded
wire (what htmx actually sends): true + on both flip the caller's
policy to 'all', an unchecked add leaves the default.
A fresh detail-page load while downloads were active rendered the
NON-polling episode container (PollingEpisodeList was only reachable
via the /episodes partial route after Refresh), so the page froze
until the user refreshed manually. PodcastDetailPage now picks the
variant the same way the route does — HasActiveDownloads selects the
every-2s poll container.
Also: the indeterminate progress bar (feeds without Content-Length,
e.g. Scribl) used animate-pulse, which fades in place and reads as
frozen; replaced with a sliding shimmer keyframe
(.podcast-progress-indeterminate, 33% sliver, translating -100% ->
300%) in input.css. The live MB counter stays.
§0 item 1.
Logout wiped localStorage but never cleared the token cookie (HttpOnly —
only a server response can) and never called the logout API, leaving the
refresh token live for 30 days and every cookie-authenticated SSR page
accessible: the fake-signout bug. Reproduced on prod and dev; unrelated
to RLS (it sits behind authentication) and long-standing.
- /api/auth/logout now ALWAYS clears the session cookie (Set-Cookie
empty + Max-Age=-1, HttpOnly/Path/SameSite mirroring login) — even a
malformed body must sign the browser out; refresh-token revocation
unchanged.
- The header logout button (and the rules-page copy) now POST to the
API with the stored refresh token BEFORE clearing local state and
redirecting — best effort: local cleanup never blocks on API failure.
- toast 401 handler fixed: it cleared 'refreshToken' but storage uses
'refresh_token' — the real key.
- Regression test TestLogoutClearsSession: real login → logout →
response deletes the cookie; logged-out refresh token can't mint
access tokens; body-less logout still clears.
Also carries regenerated artifacts from the confirm-modal and
episode-list work (header_templ/podcast_detail_templ/style.css).
Apple scheme normalization (follow-up to the Scribl discussion): Add
Feed accepts the launcher spellings podcast pages ship next to plain
RSS — itpc://, podcasts://, feed:// — and rewrites them to the https
feed URL they wrap (scheme match case-insensitive). http stays and the
guard still rejects it. Unit-pinned.
Build stamp (stale-container drift guard): the 2026-10-29 podcast
session lost an hour to a container running a build that predated the
feature. version.Commit rides ldflags from a GIT_COMMIT build arg; the
dev compose passes git rev-parse --short HEAD via the Makefile; the
app logs 'Bookhoard <ver> (commit <hash>)' at boot and /api/health
reports it — the UI and the repo can no longer silently diverge.
ONE styled modal answers every hx-confirm in the app: a body-level
htmx:confirm interceptor (web/src/confirm.ts, canonical htmx pattern)
preventDefaults and shows the dialog, issuing the request on Confirm.
Nine hx-confirm call sites upgrade without touching them. The modal
lives in the header partial so every page has it; confirmDialog() is
exported for JS call sites and degrades to the native dialog if the
header is absent (bare layouts).
Migrated the five window.confirm sites (admin.ts ×3, book-detail.ts ×2)
to the same promise API.
Podcast feed detail (issues #4/#7): the unsubscribe copy is
role-aware — non-admins get 'Unsubscribe from this feed?' with silent
retention (file state is admin knowledge); an admin who is the LAST
subscriber gets a three-choice modal instead: Unsubscribe (files kept,
the D6 default — sliding-window feeds make purge unrecoverable) /
Remove feed + files / Cancel. Admins with co-subscribers keep the
files-note confirm.
WHY: downloads needed manual refreshes to see progress; the 204 +
HX-Trigger event indirection missed its listener, and the worker
buffered whole enclosures so no progress existed to show.
- Worker streams enclosures to disk through a counting loop (64KiB
chunks, scheme/redirect/size guards preserved) with an in-memory
per-episode progress snapshot (received/total; total from
Content-Length, seeded by the feed's declared length). Deliberately
transient — the DB row state remains the durable truth.
- Episode-list renderer seam: the router injects a closure into
PodcastHandler (handlers can't import templates — MediaDetail cycle)
so htmx download/retry/Download All/refresh/order/policy actions
answer with the REFRESHED partial directly (200, hx-target=
#episode-list) — no events to miss. Non-HX callers keep JSON; tests
without a renderer keep 204+trigger.
- Polling tightens to 2s while downloads are active; downloading rows
render a percentage bar + MB counter (quote-free inline style per
templ gotcha #8) or an indeterminate bar when the server sends no
length.
- Issue #3: Download All (button AND the add-feed checkbox) flips the
CALLER's subscription policy to 'all' — per-user, never global (D6:
files shared, policies not) — and the auto-download select
out-of-band-swaps so the dropdown shows what the user is getting.
They can turn it down afterwards; the flip is honest communication,
not coercion (user-approved 2026-10-29).
PINs: Download All with HX answers 200 + episode-list + oob select and
flips newest → all; add-feed download_all sets policy all.
WHY: the /podcasts search box filtered feed CARDS only (title, display
name, host, publisher, feed description). D14's episode search existed
only as the /api/podcasts/search endpoint and was never wired into the
page — searching 'Quarter' (present only in episode titles) cleared the
grid while 'Trader' (feed metadata) survived.
The page query now consults SearchMyPodcastEpisodes (subscription-scoped
via RLS — the caller's identity) when a query is present; a feed
surfaces when its OWN metadata OR any of its episodes matches, with an
'N matching episode(s)' badge on the card.
PIN: SSR test with a needle that exists only in an episode title — the
feed must surface with the match count.
WHY: htmx sends non-GET requests urlencoded by default; the SetFeedOrder
request struct only carried a json tag, so the binder left Order empty
and 'required' fired — the exact error the order dropdown AND the
Serial-suggestion button produced (hx-vals JSON also arrives
form-encoded). One form tag fixes both controls.
Regression test drives the exact wire format: urlencoded PUT with
order=oldest/native → 200, empty/invalid → 400.
Invariant 3: progress anchors on the episode row (guid identity +
chapter/offset), never the ordinal. Renumbering feed_position (the
fetcher's per-fetch renumber case) leaves the progress row untouched
and the sequence membership stable — downloaded-only with no gaps.
cmd/server/tests/podcast_test.go — six suites, the shared-dev-DB
doctrine (test*-named libraries, cascading cleanup, no binary
fixtures):
- TestPodcastRLSSurfaces: THE AUDITOR (invariant 13) — non-subscriber
404 on feed detail + SSR page (existence leak), episodes absent from
generic media reads, admin escape, subscriber visibility, then the
hidden-library flip hiding an already-subscribed feed (the feed
policy's library-visibility join). Seeding runs as the FIRST ADMIN —
the seed identity has neither subscription nor users row, so podcast
INSERT..RETURNING (SELECT policy under FORCE) 42501s for it;
production pods run as exactly that admin (invariant 12). All direct
row probes read through the admin scope for the same reason.
- TestPodcastAddFeedFlow: member add + subscribe + metadata-only
upsert; duplicate-URL dedupe (no second row); allow_member_feeds
gate (member 403 / admin 201 / subscribe-still-allowed).
- TestPodcastShrunkenFeedRetention: feed purges an episode from the
XML, both rows persist (upsert-without-delete pin, D14).
- TestPodcastEditSegregation: generic edit 400 + write-to-file refusal
for episodes; control book edit unaffected.
- TestPodcastDownloadWorker (ffmpeg-generated tone at runtime): file
lands at <download-root>/<feed-slug>/NNNN-…, audio_metadata probed;
re-download of a downloaded episode never re-fetches (idempotency
pin, enclosure-hit counted); dead enclosure → failed + persisted
reason + metadata intact. Fetch/SSRF tests swap the client via
SetPodcastHTTPClientForTests (the production client refuses
loopback httptest servers by design — the real client gets its own
negative test in services).
- TestPodcastScannerExclusion: SetFolders on a podcasts library root
yields zero scanner folders (invariant 4 pin) — MediaScanner
.FolderCount() accessor added.
Unit: internal/services/podcast_rss_test.go — namespace-aware parsing
fixture (three XML namespaces, fallback chains, hosts/guests),
duration spellings, RFC-822 variants, explicit variants, slug/path
layout, enclosure-extension selection, stripHTML, and the SSRF matrix
(private/loopback/link-local/multicast/mapped-v6 blocked; a loopback
httptest server PROVES the production client refuses it).
suites green: full integration run 161s / 0 failures; vitest 12 files
163 tests.
Contract widening: TestContractPin_MediaListBothBranches' raw-row
allowlist gains the 15 podcast episode columns (additive; documented in
docs/contract_pins.md — inert defaults on scan rows). Bruno: bruno/
podcasts/ mirrors the §8 endpoints (12 requests) + Bookhoard env vars.
tests/ui/COVERAGE.md rows 18-23: podcasts pages/lifecycle/RLS/chain —
Playwright sweep pending.
D7 chain-by-default, delivered with minimal player surgery:
- ReaderMetadata gains PodcastSequence: the neighboring DOWNLOADED
episodes in the caller's effective order (D5 override applied,
feed-native fallback; computed in the reader route for
source='podcast_episode' items — RLS already scoped the reads; gaps
never enter the chain, invariant 6). ?single=1 escapes chaining (the
'play single episode' affordance).
- Audio reader renders the sequence into body[data-json]
(AudioReaderConfigJSON — books produce the identical shape minus the
sequence key, so the reader path is unchanged for them) and an
'Ep N/M' indicator + prev/next episode controls in the header.
- audio.ts: the last-track 'ended' handler chains into the next
episode's reader page (final progress push precedes navigation);
back button exits episodes to their feed page. Books are untouched —
no sequence context, zero behavior change. Day-one chaining is
page-navigation based (progress persists; position restores).
Continue aggregation (2.5): GetPodcastContinueForUser (in-progress
episodes, most-recent-first, ≥95% excluded) feeds a Continue section
on /podcasts — sorted most recently listened, cards show 'Ep N · title'
— ahead of All Feeds. Cards prefer the continue line over the
description.
Rendering ownership holds throughout: every piece of server data is
templ-rendered; htmx swaps partials; the only TS is the Add-Feed modal
toggle + error surfacing (pure client-local state, bundled via
web/src/podcasts.ts with automatic bundle-hash stamping).
Sidebar: /podcasts nav entry (D13, between Series and Collections),
gated by a ShowPodcastsNav flag resolved centrally in
getTemplateUserWithTheme (one indexed COUNT — same class of
per-request lookup as the live-user guard). New 'podcast' icon.
/podcasts: header + Add Feed modal (htmx post → HX-Redirect to the
feed page; inline error surfacing on failure), subscription-scoped
search + facet bar (host/tag/language — server-side GET filters, no
JS), grid cards with download counts / new-episodes / complete /
explicit badges. Empty states for no-library (admin creates it first)
and no-subscriptions.
Feed detail: hero (artwork, host/publisher, tags, fetch state), action
row — Download All, Refresh, ORDER SELECTOR (D5 override), serial
one-click oldest-first suggestion (never silent), auto-download policy
select (D6), Unsubscribe (files retained), admin Remove Feed — gap
banner (D8 count), and the EPISODE LIST with the full lifecycle inline
(D8/D19): pending greyed + Download button (every non-downloaded
episode has one), queued/downloading pulse, failed RED + reason text +
Retry (same endpoint). The list is a self-refreshing partial: re-fetch
on the podcast:changed HX-Trigger and polls every 4s ONLY while the
server says downloads are active (it renders the polling variant).
API endpoints grew HX awareness: 204 + HX-Trigger podcast:changed (or
HX-Redirect for add/unsubscribe/remove) — the sanctioned Go-handler
response channels; JSON callers keep JSON. AddFeed result exposes the
fetch-error text for the redirect.
Dates render in the requesting user's timezone (D17) via
utils.FormatTimeInTimezone; episodes play through the existing
/readers/{id} audio reader (episodes ARE audiobook items — no new
player code day one). CSP tripwire (templates/csp_source_test.go)
passes: no inline scripts, no on*= handlers, no hx-on/js: vals.
D9/D10: users never edit podcast episode metadata. Three gates:
- Generic edit endpoint (UpdateMediaItem) rejects source=
'podcast_episode' items with 400 — the next fetch refresh would
clobber any edit anyway, and the fetcher is the sole metadata writer.
- Write-to-file (MetadataWriter.WriteMetadataToFile) refuses podcast
episodes for the same reason (tag write-back for downloaded episodes
is a named Phase-4 item needing its own iTunes-tag mapping).
- The book-detail page hides the admin Edit affordance for
podcasts-type items (direct-URL visits), and the metadata-editor
modal's field list has no podcast fields by construction (fixed
templ field list) — episodes render read-only via /podcasts
surfaces.
The §8 surface, wired through main.go → router.Config →
/api/podcasts (JWT). Scoping is the engine's (D20): lookups outside the
caller's scope simply return no rows → 404; the only Go-level checks
are explicit-argument authorization (AddFeed's library +
allow_member_feeds gate) and the admin role on feed deletion.
Endpoints: POST/GET /feeds (add: fetch+parse+upsert+subscribe, optional
download_all; list: subscriptions + episode counts + facet values
computed in Go — feed counts are tens), GET /feeds/:id (detail with
download states + gap count), PUT/DELETE /subscription (the ONLY
user-writable podcast state, D9; unsubscribe retains files per D6),
POST /refresh, POST /episodes/:eid/download (+ episode_ids[] bulk and
eid='all' = Download All, exactly the non-downloaded set),
PUT /order (D5 per-user override in the reader_settings KV —
podcast_order:<feed_id>; 'native' deletes the row so default and
explicit choice are indistinguishable), GET /sequence (effective order
— override applied, native fallback — DOWNLOADED-only with the gap
list; this payload powers web chain mode and, per §15, the future
media3 queue), GET /search (subscription-scoped feed + episode search
covering un-downloaded and historical rows, D14), DELETE /feeds/:id
(admin; best-effort file removal, cascade deletes rows).
Gap definition (countGaps): an un-downloaded episode with a downloaded
episode AFTER it in list order — the D8 'middle episodes missing'
banner condition.
One ticker re-reading its interval from the settings registry each
cycle (live tunable — the cache stays honest because writes go through
the settings API path, AGENTS.md gotcha #2). Per tick: list feeds
(first-admin identity, invariant 12), skip is_complete feeds (back
off; explicit refresh still works), poll each (conditional GET via
stored validators), and apply the subscriber auto-download policy
union after each successful fetch so new episodes queue without
waiting a tick.
Schema/registry: podcast_poll_interval_seconds (default 3600s, floor
300s, category podcasts) — seeded with ON CONFLICT DO NOTHING per the
idempotency rule.
One sequential worker (day one) over a DB-backed queue
(download_state='queued' IS the worklist — the episode list always
renders truth, D8/D19). State machine: pending → queued → downloading
→ downloaded | failed (+ download_error + download_attempts).
Idempotency: downloaded re-request is a no-op; a file already on disk
that probes cleanly re-marks the row without re-fetching
(crash-between-move-and-update reconciliation). Retry: manual path
resets attempts (human override), auto policies re-queue up to
maxAutoDownloadAttempts=5 then stay failed until a human retries.
Pipeline: enclosure fetch (same SSRF/size guards as the fetcher) →
.part temp → declared-size pre-check → MIME/size validation (empty /
over-cap / content-sniff refuses non-audio) → ffprobe reconcile +
embedded-cover extraction (scanner machinery reuse: runFFProbe,
extractEmbeddedAudioCover, AudioMetadataPayload) → atomic rename →
SUCCESS-ONLY row update (invariant 7: a failure never clears existing
episode data).
Policy union (D6): subscriber policies per feed — 'all' → everything
missing; 'newest N' → the N tail positions; 'off' contributes nothing;
auto-retry budget respected at queue time. DeleteFeedFiles for the
admin-only removal path (episode files + cover siblings + artwork).
The fetcher is the SOLE writer of feed/episode metadata (D9). Three
files, no new deps (stdlib encoding/xml + net/http):
podcast_rss.go — RSS 2.0 + iTunes + Podcasting-2.0 structs. iTunes
fields are namespace-qualified in tags (local names collide with core
RSS: title/subtitle/author/summary/explicit/image/duration);
podcast:* fields match on local name (collision-free, tolerant of both
declared namespace URIs). Description/author fallback chains per §5;
itunes:duration parses seconds/HH:MM:SS/MM:SS/prose; pubDate accepts
the RFC-822/1123 spellings feeds actually emit.
podcast_http.go — the security boundary for user-supplied URLs:
https-only on every hop (CheckRedirect), private/loopback/link-local/
multicast IP ranges blocked at DIAL time (Dialer.Control sees the
RESOLVED address — DNS rebinding into 127.0.0.1 is refused at the
socket), 10-redirect cap, hard body caps (25MiB feeds / 20MiB artwork /
1GiB enclosures), no proxy.
podcast_service.go — fetch → parse → upsert. Episodes upsert
metadata-only keyed on (source_feed_id, episode_guid); rows for items
that vanish from the XML are NEVER deleted (D14 search covers
historical episodes). itunes:new-feed-url + permanent HTTP redirects
rewrite feed_url in place (row/subscriptions/episodes stay put);
podcast:guid dedupe reuses the canonical row instead of forking.
Artwork downloads feed-level under <feed-slug>/artwork.<ext>.
AddFeed authorizes ONLY on explicit arguments (library visible +
allow_member_feeds for non-admins — the D20 scoping model); detached
work runs under the first-admin identity via adminCtx (invariant 12).
Feed layout: <download-root>/<feed-slug>/<pos:04d>-<slug>.<ext> so
ServeFile/Range streaming resolve unchanged.
§9 / D3 / D13 / D15 / D16: podcasts-type libraries are excluded from
every generic book surface. Under RLS these are pure UX/type filters
(engine already hides other users' episodes); their job is preventing
episode grids in surfaces designed for books. The discriminator is the
denormalized media_items.library_type_name (IS DISTINCT FROM 'podcasts'
— NULL rows predate the column and stay visible).
SQL (queries.sql, marked '-- podcasts exclusion (plan §9)'):
ListMediaItems/ListMediaItemsSorted (All Books) · SearchMediaItems +
SearchMediaItemsUnified (generic search) · FacetCountsUnderFilters (all
14 arms) + the six Search*Values facet-suggestion queries ·
GetDistinctSeries + Count (series pages) · GetLibraryItems (collections
picker) · dashboard system sections (Continue Reading / Recently Added
/ Recently Read / Not Started) · GetPopularBooks + GetTopBooksByMinutes
(book rankings must not rank episodes).
Go: statistics summary counts episodes' sessions into time-spent (D15
keeps that) but skips them in BooksActive/BooksFinished — an episode is
not a book. OPDS filters podcasts libraries via one helper at all 7
library-walk sites (D16).
Templ: LibrarySwitcher skips podcasts entries and TotalMediaCount
excludes them — presentation-layer filter so every switcher consumer
(dashboard, collections, custom-section builder) is correct at once
(D13); podcasts surface via /podcasts only.
WHY: podcast download roots hold fetcher-written episode files; the FS
walker must never discover them (invariant 4) — a walk would double-
ingest episodes as scanner items and fight the fetcher's sole-writer
rule (D9). Every scan/watch path funnels through MediaScanner.SetFolders
(StartScanner, ScanLibrary jobs incl. scoped single-folder scans,
watch mode), so the filter lives there once: folders resolving to a
podcasts-type library are dropped before registration, with a log line.
Belt and braces: ScanLibrary's per-folder scope refuses purpose='podcast'
folders explicitly (400) — the panel-level path that bypasses type
resolution. AddLibraryFolder now derives purpose from the library type
('podcast' for podcasts libraries, enforced single download root;
'scan' default elsewhere), so folder purpose and library type cannot
diverge through the API.
Podcast integration task 1.1 (plan: podcast-integration.md §6/§8, D20).
WHY: podcasts are a new first-class media domain — RSS feeds users
subscribe to personally, with episodes as media_items rows so progress,
stats and the audio player work unchanged. Visibility is per-user
(strict: only your subscriptions), which after the RLS migration
(2026-10-29) belongs in the POLICIES, not Go helpers.
Schema (idempotent — schema.sql re-executes every boot):
- 'podcasts' library type seeded (ON CONFLICT DO NOTHING, ARRAY[]::TEXT[]
cast required — bare ARRAY[] 42802s under ON_ERROR_STOP)
- podcast_feeds: per-library feed entity (UNIQUE(library_id, feed_url)),
full RSS/iTunes/Podcasting-2.0 metadata inventory, conditional-GET
validators, is_complete poll backoff
- podcast_subscriptions: the ONLY user-writable podcast state (display
name, auto-download policy); deliberately UNGOVERNED — it IS the read
policy's lookup table
- media_items episode columns: source/source_feed_id/episode_guid/
feed_position/season/episode/enclosure_url/declared_duration/
transcript_url/chapters_url/hosts/guests + download state machine
(pending|queued|downloading|downloaded|failed + error + attempts),
all partial-indexed so scan rows never touch them
- libraries.allow_member_feeds (first per-library knob); library_folders
.purpose ('podcast' download roots, scanner-skipped)
RLS (inside the labeled policy block):
- p_media_items_visible gains the podcast-episode subscription predicate
+ admin escape (source IS DISTINCT FROM 'podcast_episode' guard keeps
scan rows on the cheap path); unset identity still fails loud
- podcast_feeds governed identically: library-visibility join AND
subscriber-or-admin SELECT; permissive write stand-ins (mandatory at
ENABLE time; fetcher/worker declare first-admin, add-feed upsert uses
ON CONFLICT DO UPDATE RETURNING so creators never need SELECT on
foreign rows)
- Verified against dev DB: migration identity sees nothing, admin sees
all, member sees feeds/episodes only post-subscribe (probed in a
rolled-back tx — no rows persisted)
sqlc: 30 new queries (feed upsert/CRUD, subscriptions, episode
upsert-with-write-after-success-guard, download state transitions,
poller worklist, subscription-scoped search) — no visibility clauses
anywhere, per the D20 scoping model.
The 12:21 sync (cf73eec2) made the README block a byte-copy of
docker-compose.yml as it stood before the inline db-init landed;
reconcile to the current file: initdb.d mount dropped, inline
hBcescaped bootstrap, PGPORT in db-init env (DB_PORT is a variable in
every connection path). Block verified byte-identical to
docker-compose.yml; no secrets in either version (placeholders only).
z_rls_role.sh is gone. The db-init service carries the bootstrap SQL
inline (the script's docker-exec/local-socket mode was its only reason
to exist as a file, and the no-clone deployment path could not use it:
docker bind-mounts of missing host paths auto-create empty DIRECTORIES,
so a compose-only install got '/z_rls_role.sh: no such file').
The inline form surfaced two silent bugs the script had masked:
- psql defaulted its DATABASE to the username (POSTGRES_DB is not a
libpq variable) — the ownership loop ran against 'postgres' and
no-opped on existing volumes. Fixed with PGDATABASE: bookhoard.
- PGPORT was never propagated: with a custom DB_PORT, db-init dialed
5432 while Postgres listened on the custom port. Fixed with
PGPORT: ${DB_PORT:-5432} — DB_PORT is now a variable in every
connection path (db listener, host mappings, app, db-init).
Also removed the db service's initdb.d mount: the app executes the full
schema as the runtime-role owner on every boot incl. first (pg_trgm is
a trusted extension; the db owner can create it), so the bootstrap SQL
was its only content.
Verified: dev two-boot against the existing 40-table volume (owners
converged, 0 off-owner tables/functions), custom-port throwaway
(DB_PORT=15432: db-init + app both on the custom port, healthy),
default-port throwaway (fresh volume, healthy), full Go suite green.
Manual repair path: docker compose run --rm db-init.
The 'copy this wholesale' docker-compose.yml block in the README had
drifted from the real make-of-record file: it hardcoded 5432/8765
(missing the DB_PORT/SERVER_PORT configurability from de8f71b2), lost
the loopback-only db publishing (b03ed9a2), carried its own hand-tailored
db-init service with inline SQL instead of the z_rls_role.sh script
version shipped in 313268a7, and still contained the vestigial
./uploads mount on the db service (dropped in the previous commit).
Replace the whole yaml block with a byte-for-byte copy of
docker-compose.yml so the quickstart can no longer drift — the
documented optional envs, loopback bind, db-init script path, and
env_file wiring all now match what actually ships. Verified by diffing
the fenced block against the file.
The db service mounted ./uploads:/app/uploads, which Postgres has never
used and nothing in the repo relies on: no SQL COPY TO/pg_dump path, no
doc/test reference to an uploads path inside the db container (the only
/app/uploads mentions in TEST_DATA.md are the app/test containers), and
the backup scan is app-side.
Forensics: the mount was introduced by accident in 6ebe9749 ("Simplify
docker-compose healthcheck configuration"), which incidentally REPLACED
the db service's volumes with the app service's uploads mount. The
follow-up 91456d11 ("restore essential database configuration")
restored postgres_data and the schema-init mount but nobody noticed the
stray uploads mount riding along, and it has survived every compose
refactor since.
The tests service in docker-compose.dev.yml keeps its own ./uploads
mount — that is the app-like test runner (TEST_UPLOAD_PATH=/app/uploads)
and is legitimate. The dev db inherits its volumes from the base file,
so it is fixed automatically.
The Installation section assumed 'git clone the repo'. Self-hosted
operators only need three files. The install is now: mkdir + copy this
compose block + .env (three openssl secrets) + uploads dir + pull/up.
The compose is fully self-contained: the runtime-role bootstrap is
inlined into db-init's command (same SQL as
database/schema/z_rls_role.sh — keep in sync) — no ./database/schema
bind mount, because the app itself executes the schema as the runtime
role (owner) on first boot; with the role as DATABASE owner the trusted
pg_trgm extension needs no superuser.
Verified end to end in a throwaway project: fresh volume, pasted block,
.env with generated secrets, db-init converged role+ownership, app ran
the full schema and came up healthy — no repo checkout anywhere.
Compose-quoting lessons encoded (the reason for odd dollars): compose
interpolates '532016' to '$' before the container sees it, so the plpgsql
532016 tags in the inline SQL are written 532016532016, and PGUSER must be set
explicitly (POSTGRES_USER is not a libpq variable).
Approves and lands the doctrine rewrite from the verified rollout:
RLS on media_items IS the enforcement (policy joins library_visibility
on app.user_id, declared per transaction by ScopedDB); new
media-listing queries and endpoints need no visibility plumbing.
Documents the three identity bridge sites, background-job identities,
the FORCE/owner role and db-init provisioning, per-boot schema
idempotency, and the two auditors (TestRlsInvariant + tests/ui).
Item-scoped gates may answer 404 as well as 403 (contract_pins
2026-10-29). History records the convention-era leak the dashboard
section queries carried.
Deleting any device whose stats rows still referenced it failed with
500: reading_history.device_id is a NO-ACTION FK by design (the rows
are real captured reading stats — Users reads them), so a blind DELETE
violated the constraint. Surfaced by the Android app's new logout
deregister call, but the admin Devices page Deletes hit the same stone.
Fix: DetachReadingHistoryFromDevice — NULL the device association on
the user's history rows before the device DELETE (single handler
pass). device_id is nullable; the stats survive, the device frees its
identifier. sqlc regenerated.
The role split previously required a one-time manual script run on
existing volumes (initdb.d only fires on fresh ones). This folds the
bootstrap into the compose file the way split-account deployments
standardly do it: a one-shot db-init service runs z_rls_role.sh
(idempotent, network mode via PGHOST=db + PGPASSWORD from DBPASS) on
every 'compose up', and the app gates on
db-init: condition: service_completed_successfully.
- Zero manual steps for fresh AND existing deployments: update compose,
add DBSRVPASS to .env, 'compose pull && up -d'.
- Security posture improves: the superuser password is now held ONLY by
the init service — the app container connects exclusively as the
non-superuser runtime role.
- z_rls_role.sh gains dual connection modes (local socket for docker
exec, network for the init service) with the DBPASS requirement
enforced only in network mode.
- Verified on the dev stack: init runs and exits clean, app gates on it
and comes up healthy, second 'up' re-converges idempotently, invariant
test passes through the new dependency chain.
The initdb.d mount stays as-is (harmless double coverage on fresh
volumes; both paths are idempotent).
.env.example is the template prod/new deployments copy; it was missing
the runtime role password the app now requires (DATABASE_USER
bookhoard_srv). Notes distinguish it from the bootstrap DBPASS and
point existing-volume operators at the one-time z_rls_role.sh run.
Phase 5.5 (plan §7.5) first increment: the SSR/HTMX frontend is swept
in a real browser (official Playwright image, no host browsers) for
BOTH personas — testuser (admin) and rls-phone (restricted; Audiobooks
hidden from them via the visibility API).
Green checks: auth rejection + session establishment; admin dashboard
renders sections incl. the hidden-from-others library; restricted
dashboard renders real sections (Recently Added present — not a
vacuous landing-page pass) with ZERO hidden-library traces; search and
media payloads leak nothing for the restricted persona.
Harness lessons encoded: / is a landing page (dashboard lives at
/dashboard); the login endpoint rate-limits rapid attempts so sessions
are established once per persona and reused via storageState — a sweep
must not look like a brute-force run; getByText visibility assertions
are wrong for <option> text (library switcher) — body-textContent
assertions instead.
COVERAGE.md tracks the full 14-area sweep per the plan; the remaining
areas (reader, collections CRUD, admin matrix, upload roundtrip) are
enumerated and pending their sweep runs.
Phase 5 (plan §7): the entire integration suite passes with the RLS
policies live. The suite was the triage enumerator the plan predicted:
- device_auth middleware: the THIRD identity bridge — KOReader/Kobo
device-authenticated requests read media as the DEVICE OWNER
(plan A9 called this path out; the flip was 404s on sync metadata).
- conflicts/analytics/progress handlers: detached
context.Background() reads threaded to the request context (the
requesting user's identity) — 12+ sites; resolve helpers take ctx.
- scanner: ScanFolders self-declares its configured admin identity
(server maintenance, product-correct for any caller); generic job
creation stamps the requesting admin.
- harness: setup.DBPool and the sync-test pool declare
rls.SeedIdentity at SESSION scope via BeforeAcquire — the pool never
serves a second identity, so no A3 bleed; setup.DB falls back to the
seed identity (NewSeedDB) for ungoverned-context calls; filter tests
use seedExec for raw governed-table statements.
- contract pin WIDENED (docs/contract_pins.md 2026-10-29): under RLS,
item-scoped gates may answer 404 (row invisible — no existence leak)
as well as 403 (Go-level checks on listings keep strict 403).
- TestWorker_JobStatusTracking — the pre-existing red from the Phase 0
baseline — passes with the harness seed work; no product change.
The full suite enumerated the remaining triage class: tests drive the
scanner directly (ScanFolders(context.Background())) and read seeded
media via Queries with bare contexts — every governed read raised.
- MediaScanner.ScanFolders self-declares its configured admin
(SetAdminID — set by main AND the test helpers): a scan is server
maintenance, product-correct regardless of caller.
- ScopedDB gains a Fallback identity, exposed as NewSeedDB for the
TEST HARNESS ONLY: setup.DB falls back to rls.SeedIdentity when the
context carries none, so direct seeding/verification helpers work
uniformly instead of per-site context threading. rls.SeedIdentity is
the shared constant (no library_visibility rows => full visibility).
- generic job creation (jobs.go) stamps the requesting admin into the
job Context, same as the scan endpoint.
The sync queue processor stays identity-free: sync_queue is not an
RLS-governed table.
Phase 4 of the RLS plan: detached contexts (context.Background()) that
reach media_items now carry identities — with the policies live, a bare
context made governed reads raise (fresh connections: 42704 unrecognized
parameter; reused pool connections: 22P02 empty-string cast after the
LOCAL GUC reverted — the plan's A2 signature).
- scanner jobs (API-triggered, mtime-poll, initial scan): the job's
Context now carries the acting admin's identity — the library owner,
first-admin fallback (the job's UserID already encoded this trust
decision; the Context just needed to carry it). performInitialScan
uses the scanner's configured admin.
- websocket initial-sync read (GetUserMediaItemsForSync): runs as the
connection's own user.
- hash backfill: runs as first admin (server self-heal over all media).
Verified on the dev stack: forced rescan of the 1199-file Audiobooks
library completed with 0 errors under RLS (previously every SHA lookup
and create raised); hash-backfill refusals gone. The sync queue
processor needs no identity — sync_queue is not an RLS-governed table.
Known remaining detached reads (conflicts/analytics handlers using
context.Background()) are HTTP-path triage, enumerated by the full
suite next.
The dashboard leak is closed at the database engine. The black-box
invariant (restricted user + hidden library => absent from every
media-returning surface) went RED->GREEN with this change set; clause-
protected surfaces never regressed, proving enforcement came from the
policies, not the test.
Phase 3 of plan ~/.opencode/plan/server-rls-library-visibility.md:
- schema.sql: idempotent RLS block — ENABLE + FORCE on media_items,
SELECT policy mirroring COALESCE(is_visible,true) exactly, permissive
per-command write stand-ins (FOR ALL would OR into SELECT).
- internal/rls + internal/database/scoped.go: ScopedDB declares the
acting user per statement (LOCAL GUC). DESIGN CORRECTION during
execution: the decorator never refuses — the ENGINE is the fail-loud
layer (policies raise on governed tables; auth lookups and settings,
ungoverned, simply work). A decorator refusal would have been a new
convention gate — plan B thinking — and broke login (chicken-and-egg:
credentials are checked before any identity exists).
- database/schema.go: the migration tx declares a synthetic identity —
schema.sql's dedup DO block SELECTs media_items, which raises once
policies are live (plan A11). Caught by the two-boot gate: boot 1
created the policies after the block ran; boot 2 raised.
- router.go: RLS bridge at BOTH c.Set("user") sites (JWT + device
auth) — echo-local c.Set never reaches the stdlib ctx the DB layer
receives.
- search: the search service's raw DBTX path receives the ScopedDB
itself, never the bare pool.
- harness: the server under test is wired exactly like main.go; the
seed handle is also ScopedDB-wired but carries seedCtx() — a
synthetic identity with no visibility rows — because sqlc :one
inserts use RETURNING, which evaluates the SELECT policy even on
INSERT. Regex UUID-scraping in the invariant replaced with a
structured JSON decode.
- invariant test: RED -> GREEN across every surface incl. View All;
403 pins and admin-scoping pins hold.
Design correction discovered by the auth chicken-and-egg: the
decorator-level refusal was re-created a convention gate (plan B
thinking). ScopedDB now NEVER refuses — it declares the GUC when the
context carries an identity and declares nothing otherwise; the ENGINE
is the fail-loud layer (policies raise on governed tables, ungoverned
tables like auth lookups simply work). rls.WithInternal removed as a
dead concept.
- search service's raw DBTX path now receives the ScopedDB itself
(main.go + harness) — a bare pool bypassed the GUC.
- database/schema.go: migration tx declares a synthetic identity
(schema.sql's dedup DO block SELECTs media_items; with policies
live, boot 2 raised 'unrecognized configuration parameter' — A11).
- harness: TestServerSetup.DB is a raw-pool SEED handle (write
policies permissive, other seeded tables ungoverned); the server
under test keeps the scoped wiring. 33 CreateMediaItem sites use
seedCtx() — a seed identity with no visibility rows — because
RETURNING requires SELECT visibility of the new row.
- invariant test: regex UUID-scraping replaced with a structured JSON
decode of the sections payload.
INTERRUPTION NOTE: an edit to this same effect (seedCtx) was applied,
confirmed by the tool, and subsequently vanished from disk while an
'opencode serve' process (PID 84551) is alive on the machine. This
commit pins the tree against further loss; implementation is PAUSED
pending the user killing that process.
Phase 3 of the RLS plan, pinned mid-flight because a parallel writer
was actively editing cmd/server/tests/rls_visibility_test.go (A14):
the file flip-flopped between two implementations and gained a compile
error after the previous commit. This snapshot makes the whole tree
compile; functional verification (two-boot gate, invariant green,
full suite) lands in the next commit(s).
- internal/rls: typed ctx keys — user identity (bridged from echo by
the router) and an INTERNAL marker whose semantics grant nothing:
no GUC is declared, so policy'd tables still raise at the engine;
only RLS-ungoverned tables (system_settings, users counts) work for
user-less startup paths (settings registry Load, CountAdmins).
- internal/database/scoped.go: DBTX decorator — per-statement tx,
commit-on-Close (:many) / commit-on-Scan (:one), error paths always
terminate the tx, refusal before SQL when no identity and not
internal.
- database/schema/schema.sql: idempotent policy block (DROP IF EXISTS
+ CREATE — schema.sql re-runs every boot): ENABLE + FORCE, read
policy exactly mirroring COALESCE(is_visible,true), permissive
per-command write stand-ins (FOR ALL would OR into SELECT).
- router.go: RLS bridge at BOTH c.Set("user") sites (JWT + device
auth) — echo-local storage never reaches the stdlib ctx.
- main.go + test harness: database.New over ScopedDB at both wiring
sites; startup reads use rls.WithInternal.
Three harness corrections found while executing the invariant test:
- library type is 'ebooks' (plural) in library_types, not 'ebook'
- cleanup ordering: t.Cleanup is LIFO — register the pool Close BEFORE
the fixture's library deletes, or the deletes run against a closed
pool and fail silently into ignored errors (leaving RLSMARKER rows
behind; found because a second run saw the first run's 'deleted'
books still leaking through Recently Added)
- collection detail is addressed by the section's collection_id (the
UUID in view_all_url), not the section name
The app connected as 'postgres' — the bootstrap SUPERUSER — which
bypasses row-level security unconditionally (FORCE RLS does not apply
to superusers), making any future policy a silent no-op. This is the
prerequisite commit for the RLS visibility work (plan
~/.opencode/plan/server-rls-library-visibility.md REV 2, §4.0).
What changed:
- database/schema/z_rls_role.sh: creates the bookhoard_srv role
(LOGIN, non-superuser, non-bypassrls) and transfers ownership of
every public table/sequence/function + schema + database to it.
Ownership (not mere grants) because executeSchema re-runs the whole
schema.sql on EVERY boot and needs owner DDL. Idempotent, catalog-
only — no user data is read or written.
- docker-compose.yml app service + docker-compose.dev.yml tests
service: DATABASE_USER postgres -> bookhoard_srv, DATABASE_PASSWORD
DBPASS -> DBSRVPASS (new .env key, generated locally).
- POSTGRES_USER: postgres stays as the container bootstrap account;
on FRESH volumes z_rls_role.sh runs automatically via
docker-entrypoint-initdb.d AFTER schema.sql (alphabetical: z > s).
Verified on the dev stack: role has rolsuper=f, rolbypassrls=f, owns
all 40 tables and every public function; single boot as bookhoard_srv
executed schema.sql cleanly (CREATE EXTENSION IF NOT EXISTS pg_trgm
skips fine as non-superuser when the extension exists); health +
authenticated smoke pass. Behavior otherwise unchanged — no policies
yet.
Adds the permanent auditor for per-user library visibility: a restricted
user with a hidden library sees none of its items on any media-returning
surface (media list, search, dashboard sections, series, facets, OPDS
device catalog, collection detail, scoped/detail access).
RED now, on exactly the surfaces the audit predicted: the dashboard
section queries (GetNotStartedItems / GetRecentlyAddedItems /
GetCollectionItemsForDashboard et al.) never carried the
visible_library_ids clause, and the collection-detail endpoint (mobile
View All) exposes the full section query results. Clause-protected
surfaces (media list, search, series) pass, proving the test itself
isn't a false positive.
Implementation-agnostic on purpose: it asserts the user-level invariant
through the public API, so it stays valid when RLS replaces the
per-query clauses and catches any future regression regardless of how
the code is written. Also pins: hidden library_id=403 (never silently
passed), item detail denial, and that the hide is USER-scoped (admin
still sees the hidden library).
Test-lesson encoded: cleanup ordering is LIFO — the pool Close must be
registered BEFORE fixture cleanups, or library deletes run against a
closed pool and fail silently into ignored errors.
Book cards on the library (all-books) page were oversized at full
window width, and both the all-books and series grids showed only
2 columns on phones, making covers unnecessarily large on mobile.
Shift the all-books column ladder up one at every breakpoint
(2/4/5/6 -> 3/5/6/7) so the responsive reflow the grid is liked for
is preserved while every window size gets smaller cards; the series
grid gains a column at mobile only (2 -> 3), leaving md/lg untouched
per request. Dashboard carousel sizing (w-36/w-40) and the series
stacked-covers component are deliberately unchanged.
templ codegen and the tailwind build output commit with the source
change that produced them (repo convention).
Two independent dead-ends, confirmed on prod's /admin/archived
(2026-10): Purge All rendered but clicking it did nothing at all, and
per-row Delete never even showed its confirm dialog.
Purge All is Alpine-bound (@click="purgeArchivedItems()"), but this
page's <body> lacked x-data="admin" — the only admin page without it.
Alpine evaluates directives only inside an x-data tree, so the
attribute was inert: the admin component (and its method, registered
in admin.ts) never instantiated here. The body now carries
x-data="admin", matching admin.templ and admin_library.templ.
Per-item Delete was a second, subtler dead-end: admin.ts's API helpers
guard with `if (!token) return;` BEFORE the confirm dialog, and the
reporting session had no localStorage token — SSR pages run on the
cookie session, so the web UI works while every Bearer-
authenticated admin action (purge, delete, restore, merge, scan-all,
cancel-scan) silently no-ops with zero feedback. Those guards now go
through requireToken, which toasts "Session expired — sign out and
back in, then retry" when the token is absent. The two passive
page-load probes (loadWatchStatus, checkActiveScans) deliberately keep
the silent return: a toast there would fire on every visit.
Pinned by TestAdminArchivedPageHasAlpineScope: renders the page and
asserts the body scope and the action buttons coexist in one render.
After deploy: hard-refresh prod, click Purge All (confirm appears);
if the token toast fires on any admin action, log out and back in once
to stash a fresh API token.
The archive pass looped over the scan's folders and, inside each
iteration, compared the ENTIRE library's item list against that one
folder's walk. With two registered folders every item living under the
other folder was flagged "missing (1/2)" on every reconciling scan —
prod's audiobook library (the only multi-folder one) showed 743/743
flagged — and nothing reached the two-consecutive-miss threshold only
because each scan's walk clears missing_scan_count before the pass
re-flags it. With three or more folders the defect stops bouncing: a
later iteration sees an earlier iteration's count-1 marks on items its
own walk lacks and archives them mid-pass — one registered folder away
from archiving a third of a library per scan. Single-folder libraries
were unaffected, which is why comics/ebooks/manga showed zero flags.
Folder-book presence was suspected alongside it (audioBookRoots read
during the comparison), but the map is written for every >=2-track
folder-book root on every walk (processAudioBookFile, since 6216bd9b);
the flagged folder-books were victims of the same cross-folder
comparison. The roots attribution did carry the same unchecked
strings.HasPrefix boundary bug as the rest of the scanner (8th site,
variable named `root`, missed by the earlier sweep); it now uses
folderContains.
The pass now groups folders by owning library and, per library,
compares ListMediaItemsByLibraryIncludingArchived against the UNION of
all that library's folders' walks. The coverage (partial-scope) guard,
the walk-failure skip, and the mark/archive/purge item loop are
unchanged.
Regression test (archive_multifolder_test.go) builds a two-folder
audiobooks library — single-file book in one folder, folder-of-tracks
book in the other — and scans twice: pre-fix both items are flagged
missing on both scans (verified by running the test against the old
code), post-fix neither is ever flagged or archived.
None of the three save surfaces on the admin settings page gave usable
feedback: the Base URL form re-rendered into an identical-looking form
(its only indicator was the Go-source twin's banner — markup that
diverged from the templ original), the scan form was WORSE — its handler
responds JSON, so htmx's outerHTML swap replaced the whole scan card
with literal JSON text — and the tunable rows showed only a tiny
per-row label.
Client side (toasts are the sanctioned client-local island; no new
inline scripts, CSP unchanged):
- web/src/settings.ts: one delegated htmx:afterRequest/sendError
listener keyed on a data-settings-save attribute stamped on each
form. Success → green toast per surface ("Settings saved" / "Scan
settings saved" / "Setting saved"); failure → error toast carrying
the server's JSON error field, degrading to a status line for the
tunable endpoint's HTML error body (markup must never reach a toast).
Network-level failures (htmx:sendError) toast instead of silently
re-enabling the button.
- htmx 2.x quirk handled: for FORM submissions, afterRequest's
detail.elt is the event-dispatch ancestor, not the form — the
requesting element lives on detail.requestConfig.elt (verified
empirically; saveLabelFor reads the right one).
- hx-disabled-elt="this" on all save buttons: the click is
acknowledged and double-clicks cannot double-save.
Server side:
- Scan form: hx-swap="none" (the inputs already hold the saved values;
nothing needs re-rendering) — kills the JSON-over-the-card swap.
- The Base URL form's save response is a Go-source copy of the templ
form (handlers cannot import templates: templates/utils.go imports
handlers types). The twin now carries the functional attributes
(data-settings-save, hx-disabled-elt) so feedback survives the swap,
with a drift-warning comment; the double-save e2e pins this.
Dev-DB safety: every success save in the spec re-submits the value
already on the page (zero drift), and the spec restores
system_config.updated_by after itself — the save handler stamps the
requesting admin's id there, and that FK (no cascade) otherwise blocks
the integration harness's transient-user sweep (it broke
TestUpdateProfileAdminMode/TestDeleteUserConsolidated on this run's
first pass; pollution cleaned up by hand and now self-restoring).
Tests: web/src/settings.test.ts pins the form→label map and the
error-extraction contract (including the no-markup-in-toast rule);
e2e/settings-feedback.spec.ts drives all three surfaces through the
real admin UI including the error path and the twin-drift double-save.
A bare string-prefix test (strings.CutPrefix / strings.HasPrefix) let a
registered folder whose path was a string prefix of a sibling claim the
sibling's files: "/app/audiobooks" matched "/app/audiobookssd/Brian
Jacques/...", so getRelativePath sliced the remnant "sd/" into every
stored relative path for that library (prod scan log: "[AUDIO] Created
folder book 'Taggerung' (41 tracks) at sd/Brian Jacques/Redwall/Book 14").
Scanner lookups built their keys with the same function
(getMediaItemByFilePath, scanStateLookup), so writes and lookups were
consistently poisoned and no scan ever corrected them; everything
outside the scanner resolves paths by joining back to the roots
(ResolveMediaPath), so cover/track URLs 404ed for files sitting one
directory away.
The same unchecked HasPrefix pattern also drove five folder-membership
sites, so a sibling-prefix pair misattributed more than paths:
owning-library resolution (processMediaFile, maybeProcessAudioBook),
per-library-type extension allowlisting (isScannableFile), and
folder-structure metadata extraction on both create and update paths —
a file could be scanned under the wrong library's rules entirely.
folderContains(folder, path) is now the single boundary-aware
primitive: path must equal the folder or continue with a
filepath.Separator (after Clean). All seven sites go through it, so no
folder name — however similar to another — can interfere with any
other's files again, in any order of registration; trailing slashes on
registered folders are normalized. Registration order still decides for
genuinely overlapping (nested) registered roots: pre-existing behavior,
deliberately unchanged.
Regression tests (media_scanner_path_test.go, pure unit — no DB) pin the
prod trap (audiobooks/audiobookssd), order-independence, trailing
slashes, mid-word siblings, and the helper itself.
Repair of already-poisoned rows is deliberately separate and safe only
under this fix: a rescan repoints them via the SHA-256 move path
(MoveMediaItemFilePath) once lookups stop producing poisoned keys.
John O'Keefe