From afa07690f5235f0b2746992984c58b048f6b995a Mon Sep 17 00:00:00 2001
From: BoulderBadgeDad <batmanesignal@gmail.com>
Date: Sat, 13 Jun 2026 16:55:33 -0700
Subject: [PATCH 01/11] Find & Add: match a Spotify 'Title - Remix' query to
 the base-titled library track
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

wolf's report: Spotify shows 'Calma - Remix', Find & Add searches that literal
string, but the library stores the track as just 'Calma' (only the 3:58 duration
marks it the remix). The literal LIKE '%calma - remix%' misses, so it fell to the
OR-fuzzy fallback which floods on the common word 'remix' (20 unrelated '... remix'
hits). Dropping '- Remix' (searching 'Calma') finds it instantly.

Fix: search_tracks (and api_search_tracks) now retry on the BASE title — the part
before Spotify's ' - ' version separator — BEFORE the OR-fuzzy flood. So
'Calma - Remix' resolves to 'Calma' (or 'Calma (Remix)') and the noise fallback is
never reached when the base matches. New core.text.title_match.base_title_before_dash
(splits the first spaced ' - '; leaves bare hyphens like 'Up-Tight' alone).

Tests: pure helper (3) + real-DB integration reproducing the Calma case, the
parenthesized-remix variant, plain-title-unaffected, and no-flood (4). 64
search/match tests green.
---
 core/text/title_match.py                 | 18 +++++
 database/music_database.py               | 28 +++++++-
 tests/test_base_title_search_fallback.py | 88 ++++++++++++++++++++++++
 3 files changed, 133 insertions(+), 1 deletion(-)
 create mode 100644 tests/test_base_title_search_fallback.py

diff --git a/core/text/title_match.py b/core/text/title_match.py
index 85e1c10b..36818fcd 100644
--- a/core/text/title_match.py
+++ b/core/text/title_match.py
@@ -201,9 +201,27 @@ def numeric_tokens_differ(title_a: str, title_b: str) -> bool:
     return _digit_tokens(title_a) != _digit_tokens(title_b)
 
 
+def base_title_before_dash(title: str) -> str:
+    """The base title before Spotify's ' - <qualifier>' version separator.
+
+    Spotify renders versions as 'Calma - Remix' / 'Song - Radio Edit' /
+    'Track - Remastered 2019'. Libraries (and the files people actually have)
+    very often store just the base — 'Calma' — so a literal search for
+    'Calma - Remix' finds nothing and the OR-fuzzy fallback then floods on the
+    common qualifier word ('remix' matches every remix). This returns the base
+    ('Calma') for a base-title search fallback. Splits on the FIRST ' - ' (the
+    spaced hyphen is Spotify's separator; a bare hyphen inside a word is left
+    alone). Returns the title unchanged when there's no separator."""
+    if not title:
+        return title
+    idx = title.find(' - ')
+    return title[:idx].strip() if idx > 0 else title
+
+
 __all__ = [
     "titles_plausibly_same",
     "strip_redundant_context_qualifiers",
     "strip_subtitle_qualifiers",
     "numeric_tokens_differ",
+    "base_title_before_dash",
 ]
diff --git a/database/music_database.py b/database/music_database.py
index 0d06a6db..1f6078fb 100644
--- a/database/music_database.py
+++ b/database/music_database.py
@@ -6887,11 +6887,26 @@ class MusicDatabase:
 
             # STRATEGY 1: Try basic SQL LIKE search first (fastest)
             basic_results = self._search_tracks_basic(cursor, title, artist, limit, server_source, rank_artist)
-            
+
             if basic_results:
                 logger.debug(f"Basic search found {len(basic_results)} results")
                 return basic_results
 
+            # STRATEGY 1b: Spotify renders versions as "Title - Qualifier"
+            # ("Calma - Remix") but libraries usually store just the base
+            # ("Calma"), so the literal search misses. Retry on the base title
+            # BEFORE the OR-fuzzy fallback (which would flood on the common
+            # qualifier word — every "... remix" matches "remix"). #: Calma - Remix
+            if title:
+                from core.text.title_match import base_title_before_dash
+                base_title = base_title_before_dash(title)
+                if base_title and base_title != title:
+                    base_results = self._search_tracks_basic(
+                        cursor, base_title, artist, limit, server_source, rank_artist)
+                    if base_results:
+                        logger.debug("Base-title search matched '%s' via '%s'", title, base_title)
+                        return base_results
+
             # STRATEGY 2: Broader fuzzy search - splits into individual words with OR matching
             fuzzy_results = self._search_tracks_fuzzy_fallback(cursor, title, artist, limit, server_source)
             if fuzzy_results:
@@ -6920,6 +6935,17 @@ class MusicDatabase:
             if basic_rows:
                 return [dict(r) for r in basic_rows]
 
+            # Base-title fallback for Spotify "Title - Qualifier" forms (see
+            # search_tracks STRATEGY 1b) before the OR-fuzzy flood.
+            if title:
+                from core.text.title_match import base_title_before_dash
+                base_title = base_title_before_dash(title)
+                if base_title and base_title != title:
+                    base_rows = self._search_tracks_basic_rows(
+                        cursor, base_title, artist, limit, server_source)
+                    if base_rows:
+                        return [dict(r) for r in base_rows]
+
             fuzzy_rows = self._search_tracks_fuzzy_rows(cursor, title, artist, limit, server_source)
             return [dict(r) for r in fuzzy_rows]
         except Exception as e:
diff --git a/tests/test_base_title_search_fallback.py b/tests/test_base_title_search_fallback.py
new file mode 100644
index 00000000..f9c2b936
--- /dev/null
+++ b/tests/test_base_title_search_fallback.py
@@ -0,0 +1,88 @@
+"""Find & Add: Spotify "Title - Qualifier" must find the base-titled library track.
+
+wolf's report: Spotify shows "Calma - Remix", Find & Add searches that literal
+string, the library stores the track as just "Calma" (only the duration marks it
+as the remix) → the literal search misses and the OR-fuzzy fallback floods 20
+unrelated "... remix" hits. Dropping "- Remix" (searching "Calma") finds it.
+
+Fix: search_tracks retries on the base title (before Spotify's " - " separator)
+before the OR-fuzzy flood.
+"""
+
+from __future__ import annotations
+
+import pytest
+
+from core.text.title_match import base_title_before_dash
+from database.music_database import MusicDatabase
+
+
+# --- pure helper -----------------------------------------------------------
+
+def test_base_title_before_dash_strips_spotify_version_suffix():
+    assert base_title_before_dash('Calma - Remix') == 'Calma'
+    assert base_title_before_dash('Closer - Radio Edit') == 'Closer'
+    assert base_title_before_dash('Crocodile Rock - Remastered 2014') == 'Crocodile Rock'
+
+
+def test_base_title_before_dash_leaves_plain_titles_alone():
+    assert base_title_before_dash('Tom Sawyer') == 'Tom Sawyer'
+    assert base_title_before_dash('21st Century Schizoid Man') == '21st Century Schizoid Man'
+    assert base_title_before_dash('Up-Tight') == 'Up-Tight'   # bare hyphen, not a separator
+    assert base_title_before_dash('') == ''
+
+
+def test_base_title_before_dash_splits_first_separator_only():
+    assert base_title_before_dash('A - B - C') == 'A'
+
+
+# --- integration: the real search path -------------------------------------
+
+@pytest.fixture
+def db(tmp_path):
+    return MusicDatabase(str(tmp_path / "music.db"))
+
+
+def _insert(db, tid, title, artist_id, artist_name):
+    with db._get_connection() as conn:
+        conn.execute("INSERT OR IGNORE INTO artists (id, name) VALUES (?, ?)", (artist_id, artist_name))
+        conn.execute("INSERT OR IGNORE INTO albums (id, title, artist_id) VALUES (?, ?, ?)",
+                     (artist_id, "Alb", artist_id))
+        conn.execute(
+            "INSERT INTO tracks (id, album_id, artist_id, title, track_number, duration, file_path) "
+            "VALUES (?, ?, ?, ?, 1, 238, ?)",
+            (tid, artist_id, artist_id, title, f"/m/{tid}.flac"),
+        )
+        conn.commit()
+
+
+def test_spotify_dash_remix_finds_base_titled_track(db):
+    # Library stores the remix as just "Calma" (the wolf case).
+    _insert(db, 1, "Calma", 1, "Pedro Capó")
+    results = db.search_tracks(title="Calma - Remix", rank_artist="Pedro Capó")
+    titles = [t.title for t in results]
+    assert "Calma" in titles, "base-title fallback should find 'Calma' for 'Calma - Remix'"
+
+
+def test_spotify_dash_remix_finds_parenthesized_remix(db):
+    # …and still matches when the library DID label it "(Remix)".
+    _insert(db, 1, "Calma (Remix)", 1, "Pedro Capó")
+    results = db.search_tracks(title="Calma - Remix", rank_artist="Pedro Capó")
+    assert any("Calma" in t.title for t in results)
+
+
+def test_plain_title_unaffected_uses_basic_search(db):
+    _insert(db, 1, "Tom Sawyer", 1, "Rush")
+    results = db.search_tracks(title="Tom Sawyer")
+    assert [t.title for t in results] == ["Tom Sawyer"]
+
+
+def test_dash_query_does_not_flood_when_base_matches(db):
+    # The base-title retry must short-circuit BEFORE the OR-fuzzy flood, so an
+    # unrelated "... Remix" track doesn't drown the real one.
+    _insert(db, 1, "Calma", 1, "Pedro Capó")
+    _insert(db, 2, "Some Other Song (KAIZ Remix)", 2, "Someone Else")
+    results = db.search_tracks(title="Calma - Remix", rank_artist="Pedro Capó")
+    titles = [t.title for t in results]
+    assert "Calma" in titles
+    assert "Some Other Song (KAIZ Remix)" not in titles

From 2905fe0853601534d0da84abd949e96c46f87f5f Mon Sep 17 00:00:00 2001
From: BoulderBadgeDad <batmanesignal@gmail.com>
Date: Mon, 15 Jun 2026 19:56:53 -0700
Subject: [PATCH 02/11] #880: retry 429 mid-walk when paginating Tidal Favorite
 Tracks (don't truncate)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

The Favorites collection walker (_iter_collection_resource_ids) broke on ANY
non-200 — including a transient 429. So a rate-limit mid-pagination silently
truncated the collection: the log shows `status=429` then `Retrieved 98/100`,
and the mirror saved 98 of a 524-track favorites list. The auto-sync cycle only
"worked" because it dodged the 429. The regular-playlist paginator already
retries 429; the collection walker didn't.

Fix: retry the same cursor page with backoff (5/10/15/20s, 4 attempts) on 429,
mirroring the playlist paginator; 401/403 still bail (+ reconnect flag), other
non-200 still break. Regression tests: 429 mid-walk completes the full chain;
exhausted retries return partial without hanging; 429 doesn't set reconnect.
---
 core/tidal_client.py                  | 26 ++++++++++++++
 tests/test_tidal_collection_tracks.py | 49 +++++++++++++++++++++++++++
 2 files changed, 75 insertions(+)

diff --git a/core/tidal_client.py b/core/tidal_client.py
index a18ab04f..4b542d3d 100644
--- a/core/tidal_client.py
+++ b/core/tidal_client.py
@@ -1655,6 +1655,8 @@ class TidalClient:
 
         ids: List[str] = []
         next_path: Optional[str] = None
+        consecutive_429 = 0
+        MAX_PAGE_RETRIES = 4
 
         while True:
             if next_path:
@@ -1687,6 +1689,28 @@ class TidalClient:
                 logger.warning(f"Tidal collection page request failed: {e}")
                 break
 
+            # Rate limited mid-walk → retry the SAME cursor page with backoff
+            # rather than silently truncating the collection. Without this a 429
+            # on page ~5 capped a 513-track favorites list at ~98 (issue #880);
+            # the regular-playlist paginator already retries 429 the same way.
+            if resp.status_code == 429:
+                consecutive_429 += 1
+                if consecutive_429 <= MAX_PAGE_RETRIES:
+                    backoff = 5.0 * consecutive_429  # 5s, 10s, 15s, 20s
+                    logger.warning(
+                        f"Tidal collection {expected_type} rate limited (429) — "
+                        f"retry {consecutive_429}/{MAX_PAGE_RETRIES} in {backoff}s "
+                        f"({len(ids)} fetched so far)"
+                    )
+                    time.sleep(backoff)
+                    continue  # next_path/cursor unchanged → re-request same page
+                logger.error(
+                    f"Tidal collection {expected_type} still rate limited after "
+                    f"{MAX_PAGE_RETRIES} retries — returning {len(ids)} IDs "
+                    f"(PARTIAL: the collection may be larger)"
+                )
+                break
+
             if resp.status_code != 200:
                 # 401/403 = scope/permission issue. Token predates the
                 # `collection.read` scope expansion or the user revoked
@@ -1704,6 +1728,8 @@ class TidalClient:
                     )
                 break
 
+            consecutive_429 = 0  # a good page → reset the retry budget
+
             try:
                 data = resp.json()
             except ValueError as e:
diff --git a/tests/test_tidal_collection_tracks.py b/tests/test_tidal_collection_tracks.py
index b821a93c..88695e99 100644
--- a/tests/test_tidal_collection_tracks.py
+++ b/tests/test_tidal_collection_tracks.py
@@ -252,6 +252,55 @@ class TestIterCollectionTrackIds:
 
         assert ids == []
 
+    def test_429_mid_walk_retries_and_completes(self):
+        """Regression for #880: a 429 mid-pagination must RETRY the same
+        cursor page (with backoff), not truncate the collection. Before the
+        fix a transient 429 on page ~5 capped a 513-track favorites list at
+        ~98 (the log showed `status=429` then `Retrieved 98/100`)."""
+        client = _make_authed_client()
+        # page1 (200) → 429 (transient) → page2 (200, end of chain)
+        responses = iter([
+            _FakeResp(200, _PAGE_ONE),
+            _FakeResp(429, text=""),       # rate limited — retry, don't truncate
+            _FakeResp(200, _PAGE_TWO),
+        ])
+        with patch.object(client, '_ensure_valid_token', return_value=True), \
+             patch('core.tidal_client.requests.get', side_effect=lambda *a, **kw: next(responses)), \
+             patch('core.tidal_client.time.sleep'):
+            ids = client._iter_collection_track_ids()
+
+        assert ids == ['1001', '1002', '1003', '1004', '1005']  # full chain, nothing dropped
+
+    def test_429_does_not_set_reconnect_flag(self):
+        """A rate-limit is transient, NOT a scope problem — must not tell the
+        user to reconnect."""
+        client = _make_authed_client()
+        responses = iter([_FakeResp(200, _PAGE_ONE), _FakeResp(429, text=""), _FakeResp(200, _PAGE_TWO)])
+        with patch.object(client, '_ensure_valid_token', return_value=True), \
+             patch('core.tidal_client.requests.get', side_effect=lambda *a, **kw: next(responses)), \
+             patch('core.tidal_client.time.sleep'):
+            client._iter_collection_track_ids()
+
+        assert client.collection_needs_reconnect() is False
+
+    def test_429_exhausts_retries_returns_partial(self):
+        """If the 429s never clear, give up after the retry budget and return
+        what we have (PARTIAL) rather than looping forever."""
+        client = _make_authed_client()
+
+        def gen():
+            yield _FakeResp(200, _PAGE_ONE)
+            while True:
+                yield _FakeResp(429, text="")
+
+        responses = gen()
+        with patch.object(client, '_ensure_valid_token', return_value=True), \
+             patch('core.tidal_client.requests.get', side_effect=lambda *a, **kw: next(responses)), \
+             patch('core.tidal_client.time.sleep'):
+            ids = client._iter_collection_track_ids()
+
+        assert ids == ['1001', '1002', '1003']  # page 1 survives; no hang
+
 
 # ---------------------------------------------------------------------------
 # get_collection_tracks_count

From 02d6af29edc35a8826e88906970028b6af222cb8 Mon Sep 17 00:00:00 2001
From: BoulderBadgeDad <batmanesignal@gmail.com>
Date: Mon, 15 Jun 2026 20:09:53 -0700
Subject: [PATCH 03/11] #879: a failed settings load must never overwrite the
 saved config

Reported by @Lysticity: opening Settings reset the whole config to defaults. The
chain: GET /api/settings 500s (their env: ConfigManager missing redacted_config)
-> loadSettingsData() called response.json() WITHOUT checking response.ok, so the
error body {"error": ...} was treated as settings -> every field populated as
`settings.x?.y || ''` blanked to defaults -> autosave then wrote those defaults
over the real config.

Fix (settings.js): bail BEFORE touching any field when the response isn't ok / is
an error body, set window._settingsLoadFailed, and guard BOTH save paths
(debouncedAutoSaveSettings + saveSettings) on it. The flag clears on the next
successful load. So any load failure (500, lock, network) now leaves the saved
config untouched instead of wiping it.

The redacted_config method exists in all 2.7.x source + on dev (their 500 looks
like a stale/mismatched build), but the UI must not destroy config on ANY failed
load. Regression test pins redacted_config stays a callable method on the class
(its removal is exactly what 500s the endpoint).
---
 tests/test_settings_redaction.py | 17 +++++++++++++++++
 webui/static/settings.js         | 30 +++++++++++++++++++++++++++++-
 2 files changed, 46 insertions(+), 1 deletion(-)

diff --git a/tests/test_settings_redaction.py b/tests/test_settings_redaction.py
index 3aea7a2a..a27e4d25 100644
--- a/tests/test_settings_redaction.py
+++ b/tests/test_settings_redaction.py
@@ -24,6 +24,23 @@ def _cm(config_data):
     return cm
 
 
+# ── #879: the GET /api/settings handler calls config_manager.redacted_config().
+# If that method is ever renamed/removed the endpoint 500s, the web UI treats the
+# error body as settings, blanks the form to defaults, and autosaves over the
+# user's real config. Pin the method's presence on the class so that can't ship.
+
+def test_redacted_config_is_a_method_on_the_class():
+    assert callable(getattr(ConfigManager, 'redacted_config', None)), (
+        "GET /api/settings depends on ConfigManager.redacted_config(); removing or "
+        "renaming it 500s the settings endpoint and the UI then wipes the config (#879)")
+
+
+def test_redacted_config_callable_on_instance_returns_dict():
+    cm = _cm({'spotify': {'client_secret': 'REAL'}})
+    out = cm.redacted_config()
+    assert isinstance(out, dict) and out.get('spotify', {}).get('client_secret') == S
+
+
 # ── redacted_config: secrets out, everything else intact ────────────────────
 
 def test_configured_secrets_are_masked():
diff --git a/webui/static/settings.js b/webui/static/settings.js
index 1921839b..d00c79d9 100644
--- a/webui/static/settings.js
+++ b/webui/static/settings.js
@@ -45,6 +45,9 @@ function debouncedAutoSaveSettings() {
     // fields on load — those aren't user edits and must not trigger a full
     // save (which re-initializes every backend service client).
     if (window._suppressSettingsAutoSave) return;
+    // #879: never auto-save while the last settings load failed — the form is
+    // showing defaults, not the real config, so saving would wipe it.
+    if (window._settingsLoadFailed) return;
     // #827: the Logs tab has no savable settings — its live-viewer controls
     // (source picker, filters, auto-scroll) were tripping the auto-save and
     // flooding app.log with "Settings saved" lines, drowning out the logs the
@@ -976,6 +979,18 @@ async function loadSettingsData() {
         const response = await fetch(API.settings);
         const settings = await response.json();
 
+        // #879: a failed GET /api/settings returns an error body (e.g. {"error":
+        // "..."} on a 500), NOT real settings. Populating from it blanks every
+        // field to its default ('settings.spotify?.x || ""'), and the next
+        // (auto)save then overwrites the user's real config. Abort BEFORE
+        // touching any field and flag it so saves stay blocked until a good load.
+        if (!response.ok || !settings || typeof settings !== 'object' || settings.error) {
+            window._settingsLoadFailed = true;
+            throw new Error('settings load failed (HTTP ' + response.status + '): ' +
+                ((settings && settings.error) || 'unexpected response'));
+        }
+        window._settingsLoadFailed = false;  // good load → saving is safe again
+
         // Populate Spotify settings
         document.getElementById('spotify-client-id').value = settings.spotify?.client_id || '';
         document.getElementById('spotify-client-secret').value = settings.spotify?.client_secret || '';
@@ -1461,7 +1476,10 @@ async function loadSettingsData() {
 
     } catch (error) {
         console.error('Error loading settings:', error);
-        showToast('Failed to load settings', 'error');
+        // #879: any load failure → block saves so a blank/partial form can't be
+        // written over the real config. Cleared on the next successful load.
+        window._settingsLoadFailed = true;
+        showToast('Failed to load settings — reload the page before saving (your saved config is untouched)', 'error');
     }
 }
 
@@ -2861,6 +2879,16 @@ function _getTagConfig(path) {
 }
 
 async function saveSettings(quiet = false) {
+    // #879: refuse to save if the settings never loaded successfully — the form
+    // is showing defaults, not the user's real config, so saving would wipe it.
+    // Cleared automatically on the next successful load (reload the page).
+    if (window._settingsLoadFailed) {
+        if (!quiet && typeof showToast === 'function') {
+            showToast("Settings didn't load — reload the page before saving (your config is untouched)", 'error');
+        }
+        return;
+    }
+
     // Validate file organization templates before saving
     const validationErrors = validateFileOrganizationTemplates();
     if (validationErrors.length > 0) {

From f2f0f5d849f0c5f6588ade1f3c1b07915bece3cc Mon Sep 17 00:00:00 2001
From: BoulderBadgeDad <batmanesignal@gmail.com>
Date: Mon, 15 Jun 2026 21:16:18 -0700
Subject: [PATCH 04/11] Sidebar UI: frosted-glass header blur, centered nav
 badges, admin cleanup
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

- .sidebar-header: real frosted-glass blur of content scrolling behind it —
  made the background translucent (was an opaque base layer), added
  backdrop-filter blur, and raised the header above the nav (z-index) so nav
  items actually sit in its backdrop.
- .dl-nav-badge: vertically centered on the right (top:50% + translateY) instead
  of pinned to the top-right corner.
- Removed border-top-right-radius from .sidebar and .sidebar-header (square top).
- Hide the "My Accounts" + "My Settings" header buttons for admin profiles —
  both are inert for admin (every service is "Managed in Settings", and My
  Settings is an empty pointer note); kept for non-admins who get real UI.
---
 webui/static/init.js   |  9 +++++++++
 webui/static/style.css | 31 +++++++++++++++++++++----------
 2 files changed, 30 insertions(+), 10 deletions(-)

diff --git a/webui/static/init.js b/webui/static/init.js
index d339cb18..a8a11b98 100644
--- a/webui/static/init.js
+++ b/webui/static/init.js
@@ -1116,6 +1116,15 @@ function updateProfileIndicator() {
     const statusSection = document.querySelector('.status-section--clickable');
     if (statusSection) statusSection.classList.toggle('status-section--locked', !currentProfile.is_admin);
 
+    // My Accounts (per-profile streaming OAuth) and My Settings (per-profile
+    // server library) are inert for admin — admin uses the global app account
+    // for every service and the full Settings page. Hide both for admin; keep
+    // them for non-admins, who actually get a connect/library UI.
+    const myAccountsBtn = document.getElementById('my-accounts-btn');
+    const personalSettingsBtn = document.getElementById('personal-settings-btn');
+    if (myAccountsBtn) myAccountsBtn.style.display = currentProfile.is_admin ? 'none' : '';
+    if (personalSettingsBtn) personalSettingsBtn.style.display = currentProfile.is_admin ? 'none' : '';
+
     indicator.onclick = async () => {
         const res = await fetch('/api/profiles');
         const data = await res.json();
diff --git a/webui/static/style.css b/webui/static/style.css
index d3a2d09c..608ee724 100644
--- a/webui/static/style.css
+++ b/webui/static/style.css
@@ -80,7 +80,6 @@ body {
     /* Soft translucent borders */
     border-right: 1px solid rgba(255, 255, 255, 0.08);
     border-top: 1px solid rgba(255, 255, 255, 0.12);
-    border-top-right-radius: 24px;
     border-bottom-right-radius: 24px;
 
     /* Soft floating shadow with inner glow */
@@ -291,17 +290,16 @@ body.reduce-effects .sidebar::after {
 
 .sidebar-header {
     min-height: 115px;
-    /* Opaque base layered under the accent gradient so nav items scrolling
-       past the sticky header don't bleed through the translucent stops. */
+    /* Translucent so the backdrop-filter blur below is visible — the dark tint
+       keeps the header readable while nav items scrolling behind it read as a
+       soft frosted blur instead of a sharp bleed-through. */
     background:
         linear-gradient(180deg,
-            rgba(var(--accent-rgb), 0.14) 0%,
-            rgba(var(--accent-rgb), 0.08) 30%,
-            rgba(var(--accent-rgb), 0.03) 70%,
-            rgba(18, 18, 18, 1) 100%),
-        rgb(18, 18, 18);
+            rgba(var(--accent-rgb), 0.16) 0%,
+            rgba(var(--accent-rgb), 0.10) 30%,
+            rgba(24, 24, 24, 0.62) 70%,
+            rgba(18, 18, 18, 0.72) 100%);
     border-bottom: 1px solid rgba(255, 255, 255, 0.08);
-    border-top-right-radius: 20px;
     padding: 20px 24px;
     display: flex;
     flex-direction: column;
@@ -309,9 +307,21 @@ body.reduce-effects .sidebar::after {
     gap: 8px;
     position: sticky;
     top: 0;
+    /* Above the nav (the sidebar gives its children z-index:1) so nav items
+       scrolling up sit BEHIND the header — i.e. in its backdrop, where the
+       blur below can actually act on them. Without this they paint in front
+       and backdrop-filter has nothing to blur. */
+    z-index: 2;
     overflow: hidden;
     flex-shrink: 0;
 
+    /* Intense frosted-glass blur: nav items scrolling up behind the translucent
+       accent-gradient top now read as a soft blur instead of bleeding through
+       sharply. (Only visible where the background is translucent — the opaque
+       base toward the bottom still stops any bleed.) */
+    backdrop-filter: blur(28px) saturate(1.3);
+    -webkit-backdrop-filter: blur(28px) saturate(1.3);
+
     /* Subtle inner glow */
     box-shadow: inset 0 1px 0 rgba(255, 255, 255, 0.1);
 }
@@ -60149,8 +60159,9 @@ body.reduce-effects #page-particles-canvas {
 
 .dl-nav-badge {
     position: absolute;
-    top: 4px;
+    top: 50%;
     right: 8px;
+    transform: translateY(-50%);
     background: rgb(var(--accent-rgb));
     color: #fff;
     font-size: 0.6rem;

From e7814e0acf048851a2cd8516cdff1138e6561671 Mon Sep 17 00:00:00 2001
From: BoulderBadgeDad <batmanesignal@gmail.com>
Date: Mon, 15 Jun 2026 21:30:55 -0700
Subject: [PATCH 05/11] #877: Download Discography filters mirror Artist Detail
 (fix dead EPs + add Live/Comp/Featured)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

The Download Discography modal exposed only Albums/EPs/Singles, its EPs toggle did
nothing, and Live/Compilations/Featured were missing — so you couldn't fine-filter
a bulk download the way Artist Detail lets you browse.

Root cause: the modal's endpoint (/api/artist/<id>/discography) used the base
get_artist_discography, which lumps EPs into singles, and the modal only read
{albums, singles} — so the EPs bucket was always empty (dead toggle). It also had
no content-type (Live/Compilation/Featured) classification at all.

- Backend: the endpoint now uses get_artist_detail_discography — the SAME split
  Artist Detail uses — and returns a separate `eps` list.
- Frontend: read `eps`; tag each card with data-is-live/compilation/featured via a
  new shared _classifyReleaseContent() (also adopted by the Artist Detail cards so
  the two can't drift); add Live/Compilations/Featured filter buttons; combined
  category+content filtering. The download payload is built from VISIBLE checked
  cards, so every toggle now actually changes what downloads.
- Regression test: get_artist_detail_discography splits an EP into the eps bucket.
---
 tests/metadata/test_metadata_discography.py | 23 +++++++
 web_server.py                               |  8 ++-
 webui/static/library.js                     | 68 +++++++++++++++------
 3 files changed, 81 insertions(+), 18 deletions(-)

diff --git a/tests/metadata/test_metadata_discography.py b/tests/metadata/test_metadata_discography.py
index 65ac0624..a5109dbb 100644
--- a/tests/metadata/test_metadata_discography.py
+++ b/tests/metadata/test_metadata_discography.py
@@ -731,3 +731,26 @@ def test_get_artist_discography_prefers_source_specific_artist_ids(monkeypatch):
         )
     ]
     assert deezer.album_calls == []
+
+
+def test_get_artist_detail_discography_splits_eps_from_singles(monkeypatch):
+    """#877: get_artist_detail_discography must put EPs in their OWN bucket (not
+    lumped into singles). The Download Discography modal now reads this split, so
+    its EPs filter has cards to act on and stays in sync with Artist Detail."""
+    spotify = _FakeSourceClient(
+        album_results=[
+            _album("a1", "An Album", "2024-01-01", album_type="album"),
+            _album("e1", "An EP", "2024-02-01", album_type="ep"),
+            _album("s1", "A Single", "2024-03-01", album_type="single"),
+        ]
+    )
+    clients = {"deezer": _FakeSourceClient(), "spotify": spotify, "itunes": _FakeSourceClient()}
+    monkeypatch.setattr(metadata_registry, "get_primary_source", lambda spotify_client_factory=None: "deezer")
+    monkeypatch.setattr(metadata_registry, "get_source_priority", lambda primary: [primary, "spotify", "itunes"])
+    monkeypatch.setattr(metadata_registry, "get_client_for_source", lambda source, **kwargs: clients.get(source))
+
+    result = metadata_discography.get_artist_detail_discography("artist-1", "Artist One", MetadataLookupOptions())
+
+    assert [r["id"] for r in result["albums"]] == ["a1"]
+    assert [r["id"] for r in result["eps"]] == ["e1"]
+    assert [r["id"] for r in result["singles"]] == ["s1"]
diff --git a/web_server.py b/web_server.py
index fd53c71a..e53a2711 100644
--- a/web_server.py
+++ b/web_server.py
@@ -9114,7 +9114,11 @@ def get_artist_discography(artist_id):
                 effective_override_source = 'spotify'
 
         from core.metadata.lookup import MetadataLookupOptions
-        from core.metadata_service import get_artist_discography as _get_artist_discography
+        # #877: use the artist-DETAIL discography so the Download Discography modal
+        # gets the SAME release-type split (albums / eps / singles) the Artist
+        # Detail view shows — EPs were being lumped into singles before, leaving
+        # the modal's EPs toggle dead.
+        from core.metadata.discography import get_artist_detail_discography as _get_artist_discography
 
         # Server-side per-source ID resolution. Look up the library row
         # by ANY of the IDs the frontend might send: library DB id,
@@ -9192,6 +9196,7 @@ def get_artist_discography(artist_id):
         )
 
         album_list = discography['albums']
+        eps_list = discography.get('eps', [])
         singles_list = discography['singles']
         active_source = discography['source']
         source_priority = discography['source_priority']
@@ -9303,6 +9308,7 @@ def get_artist_discography(artist_id):
 
         return jsonify({
             "albums": album_list,
+            "eps": eps_list,
             "singles": singles_list,
             "source": active_source or (source_priority[0] if source_priority else "unknown"),
             "artist_info": artist_info,
diff --git a/webui/static/library.js b/webui/static/library.js
index cd88e264..48f347f8 100644
--- a/webui/static/library.js
+++ b/webui/static/library.js
@@ -1891,16 +1891,12 @@ function createReleaseCard(release) {
     // Store mutable reference so stream updates propagate to click handler
     card._releaseData = release;
 
-    // Tag card for content-type filtering
-    const livePattern = /\b(live)\b|\(live[^)]*\)|\[live[^]]*\]/i;
-    const compilationPattern = /\b(greatest hits|best of|collection|anthology|essential)\b/i;
-    const featuredPattern = /\(?\bfeat\.?\s|\bft\.?\s|\bfeaturing\b/i;
-    const isLive = livePattern.test(release.title || '') || (release.album_type === 'compilation' && livePattern.test(release.title || ''));
-    const isCompilation = (release.album_type === 'compilation') || compilationPattern.test(release.title || '');
-    const isFeatured = featuredPattern.test(release.title || '');
-    card.setAttribute("data-is-live", isLive ? "true" : "false");
-    card.setAttribute("data-is-compilation", isCompilation ? "true" : "false");
-    card.setAttribute("data-is-featured", isFeatured ? "true" : "false");
+    // Tag card for content-type filtering (shared classifier — #877, so Artist
+    // Detail and the Download Discography modal never drift apart).
+    const cc = _classifyReleaseContent(release);
+    card.setAttribute("data-is-live", cc.isLive ? "true" : "false");
+    card.setAttribute("data-is-compilation", cc.isCompilation ? "true" : "false");
+    card.setAttribute("data-is-featured", cc.isFeatured ? "true" : "false");
 
     // Background image — use data-bg-src for IntersectionObserver lazy loading
     // (observeLazyBackgrounds is called by the caller after appending the grid).
@@ -2518,8 +2514,8 @@ async function openDiscographyModal() {
             const data = await res.json();
 
             if (!data.error) {
-                discography = { albums: data.albums || [], singles: data.singles || [] };
-                if (discography.albums.length > 0 || discography.singles.length > 0) {
+                discography = { albums: data.albums || [], eps: data.eps || [], singles: data.singles || [] };
+                if (discography.albums.length > 0 || discography.eps.length > 0 || discography.singles.length > 0) {
                     artistsPageState.artistDiscography = discography;
                     artistsPageState.sourceOverride = data.source || artistsPageState.sourceOverride || null;
                     // Use metadata source ID for the modal (needed for download API calls)
@@ -2569,6 +2565,9 @@ async function openDiscographyModal() {
                     <button class="discog-filter active" data-type="album" onclick="toggleDiscogFilter(this)">Albums</button>
                     <button class="discog-filter active" data-type="ep" onclick="toggleDiscogFilter(this)">EPs</button>
                     <button class="discog-filter active" data-type="single" onclick="toggleDiscogFilter(this)">Singles</button>
+                    <button class="discog-filter active" data-content="live" onclick="toggleDiscogFilter(this)">Live</button>
+                    <button class="discog-filter active" data-content="compilations" onclick="toggleDiscogFilter(this)">Compilations</button>
+                    <button class="discog-filter active" data-content="featured" onclick="toggleDiscogFilter(this)">Featured</button>
                 </div>
                 <div class="discog-select-actions">
                     <button class="discog-select-btn" onclick="discogSelectAll(true)">Select All</button>
@@ -2605,21 +2604,36 @@ async function openDiscographyModal() {
 
 function _esc(s) { const d = document.createElement('div'); d.textContent = s; return d.innerHTML; }
 
+// #877: single source of truth for content-type classification, shared by the
+// Artist Detail cards and the Download Discography modal so they can't drift.
+function _classifyReleaseContent(release) {
+    const t = (release && (release.title || release.name)) || '';
+    const livePattern = /\b(live)\b|\(live[^)]*\)|\[live[^\]]*\]/i;
+    const compilationPattern = /\b(greatest hits|best of|collection|anthology|essential)\b/i;
+    const featuredPattern = /\(?\bfeat\.?\s|\bft\.?\s|\bfeaturing\b/i;
+    return {
+        isLive: livePattern.test(t),
+        isCompilation: (release && release.album_type === 'compilation') || compilationPattern.test(t),
+        isFeatured: featuredPattern.test(t),
+    };
+}
+
 function _renderDiscogCard(release, index, completionData) {
     const comp = completionData?.albums?.find(c => c.id === release.id) || completionData?.singles?.find(c => c.id === release.id);
     const status = comp?.status || 'unknown';
     const isOwned = status === 'completed';
     const isPartial = status === 'partial' || status === 'nearly_complete';
     const year = release.release_date ? release.release_date.substring(0, 4) : '';
-    const tracks = release.total_tracks || 0;
+    const tracks = release.total_tracks || release.track_count || 0;
     const img = release.image_url || '';
+    const cc = _classifyReleaseContent(release);
     const checked = !isOwned;
     const statusClass = isOwned ? 'owned' : isPartial ? 'partial' : '';
     const statusIcon = isOwned ? '✓' : isPartial ? '◐' : '';
 
     const albumName = release.name || release.title || '';
     return `
-        <label class="discog-card ${statusClass}" data-type="${release._type}" style="animation-delay:${index * 0.03}s">
+        <label class="discog-card ${statusClass}" data-type="${release._type}" data-is-live="${cc.isLive}" data-is-compilation="${cc.isCompilation}" data-is-featured="${cc.isFeatured}" style="animation-delay:${index * 0.03}s">
             <input type="checkbox" class="discog-card-cb" data-album-id="${release.id}" data-album-name="${_esc(albumName)}" data-tracks="${tracks}" ${checked ? 'checked' : ''} onchange="_updateDiscogFooterCount()">
             <div class="discog-card-art">
                 ${img ? `<img src="${img}" alt="" loading="lazy">` : '<div class="discog-card-art-placeholder">🎵</div>'}
@@ -2636,9 +2650,29 @@ function _renderDiscogCard(release, index, completionData) {
 
 function toggleDiscogFilter(btn) {
     btn.classList.toggle('active');
-    const type = btn.dataset.type;
-    document.querySelectorAll(`.discog-card[data-type="${type}"]`).forEach(card => {
-        card.style.display = btn.classList.contains('active') ? '' : 'none';
+    _applyDiscogFilters();
+}
+
+// #877: combined category (Albums/EPs/Singles) + content (Live/Compilations/
+// Featured) filtering, mirroring the Artist Detail filter logic. A card is
+// hidden if its category is off OR any active content exclusion applies — and
+// because the download payload is built from VISIBLE checked cards, every
+// toggle now actually changes what gets downloaded.
+function _applyDiscogFilters() {
+    const typeActive = {};
+    document.querySelectorAll('.discog-filter[data-type]').forEach(b => {
+        typeActive[b.dataset.type] = b.classList.contains('active');
+    });
+    const contentActive = {};
+    document.querySelectorAll('.discog-filter[data-content]').forEach(b => {
+        contentActive[b.dataset.content] = b.classList.contains('active');
+    });
+    document.querySelectorAll('.discog-card').forEach(card => {
+        let hidden = typeActive[card.getAttribute('data-type')] === false;
+        if (!hidden && contentActive.live === false && card.getAttribute('data-is-live') === 'true') hidden = true;
+        if (!hidden && contentActive.compilations === false && card.getAttribute('data-is-compilation') === 'true') hidden = true;
+        if (!hidden && contentActive.featured === false && card.getAttribute('data-is-featured') === 'true') hidden = true;
+        card.style.display = hidden ? 'none' : '';
     });
     _updateDiscogFooterCount();
 }

From 93af95d8652d77b1f2bb7148670c3935425d4ffc Mon Sep 17 00:00:00 2001
From: BoulderBadgeDad <batmanesignal@gmail.com>
Date: Mon, 15 Jun 2026 21:37:42 -0700
Subject: [PATCH 06/11] #876: show the real Quarantine tab count on open (not a
 stale 0)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

The Quarantine tab badge was only populated by loadQuarantineList(), which runs
when the tab is clicked — so opening Library History showed a stale 0 until then.
Refresh the count on modal open via the existing /api/quarantine/list endpoint.
---
 webui/static/wishlist-tools.js | 14 ++++++++++++++
 1 file changed, 14 insertions(+)

diff --git a/webui/static/wishlist-tools.js b/webui/static/wishlist-tools.js
index b59ab783..8f95c421 100644
--- a/webui/static/wishlist-tools.js
+++ b/webui/static/wishlist-tools.js
@@ -3253,9 +3253,23 @@ function openLibraryHistoryModal() {
         overlay.classList.remove('hidden');
         _libraryHistoryState.page = 1;
         loadLibraryHistory();
+        _refreshQuarantineTabCount();   // #876: count correct on open, not only after clicking the tab
     }
 }
 
+// #876: keep the Quarantine tab badge accurate the moment the modal opens. The
+// full count was previously only set by loadQuarantineList() (i.e. after the tab
+// was clicked), so it showed a stale 0 until then.
+async function _refreshQuarantineTabCount() {
+    const el = document.getElementById('history-quarantine-count');
+    if (!el) return;
+    try {
+        const resp = await fetch('/api/quarantine/list');
+        const data = await resp.json();
+        el.textContent = (data.entries || []).length;
+    } catch (e) { /* leave the existing value on error */ }
+}
+
 // ──────────────────────────────────────────────────────────────────────
 // Quarantine tab — rendered inside the Library History modal as a third
 // tab next to Downloads + Server Imports. Reuses the existing list +

From 46be97b19584ca75e29b35327dbf8b117f64aa41 Mon Sep 17 00:00:00 2001
From: BoulderBadgeDad <batmanesignal@gmail.com>
Date: Mon, 15 Jun 2026 22:12:06 -0700
Subject: [PATCH 07/11] #876: group quarantine alternatives by target track-id
 + auto-clear siblings on approve
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Multiple failed source attempts at one song each land in quarantine as
separate entries. Group them by the *intended* target (sidecar context
track_info isrc -> id -> uri, falling back to normalized artist|title for
legacy thin sidecars) — an exact relationship across siblings, since the
bad files' own tags differ but the target track is constant.

- core: quarantine_group_key() + find_quarantine_siblings() seams; list
  entries now carry group_key.
- approve endpoint: remove_siblings flag auto-deletes the other attempts
  once one is accepted (captured BEFORE approve restores the file out of
  quarantine, or the id lookup would resolve nothing). Scoped to the
  quarantine manager; download-modal chooser + version-mismatch fallback
  pass no flag and are unaffected.
- UI: multi-member groups render as a collapsible parent row (album art +
  'N alternatives'); singletons unchanged. Toast reports removed count.
- 11 tests incl. ordering regression for capture-before-approve.
---
 core/imports/quarantine.py                  |  82 +++++++++++++++
 tests/imports/test_quarantine_management.py | 104 +++++++++++++++++++-
 web_server.py                               |  36 ++++++-
 webui/static/style.css                      |  57 +++++++++++
 webui/static/wishlist-tools.js              |  67 ++++++++++++-
 5 files changed, 337 insertions(+), 9 deletions(-)

diff --git a/core/imports/quarantine.py b/core/imports/quarantine.py
index 0e9af594..7dd43be6 100644
--- a/core/imports/quarantine.py
+++ b/core/imports/quarantine.py
@@ -153,6 +153,83 @@ def get_quarantined_source_keys(quarantine_dir: str) -> set:
     return keys
 
 
+def quarantine_group_key(
+    expected_artist: Any, expected_track: Any, context: Any = None
+) -> Optional[str]:
+    """Grouping key for "the same intended download target".
+
+    #876: when several sources are downloaded for one wishlist/queue
+    track they each fail verification and land in quarantine as separate
+    entries. They are *alternatives for the same song*, so they should
+    group together — and once the user accepts one, the rest are
+    redundant failed attempts at a song they now own.
+
+    The key identifies the *intended* target — what SoulSync was trying to
+    fetch — NOT the downloaded file's own tags. That matters: the file's
+    metadata is frequently *wrong* (that's why it failed acoustid /
+    integrity), whereas the target is fixed and identical across every
+    alternative for one song (they're all Soulseek uploads of the *same*
+    source track), so grouping by it is an exact relationship, not a fuzzy
+    metadata guess.
+
+    Prefers a stable target-track id from the sidecar `context.track_info`
+    when present — isrc, then source id, then uri — since those are exact
+    and constant across siblings. Falls back to the normalized
+    artist|track name only for legacy/thin sidecars that carry no context.
+    Keys are kind-prefixed so an id-based key never collides with a
+    name-based one.
+
+    Returns ``None`` when nothing identifies the target (no usable id and
+    both name fields empty). Callers treat a ``None`` key as "its own
+    singleton group" — ungroupable entries must never collapse together.
+    """
+    ti = {}
+    if isinstance(context, dict):
+        maybe_ti = context.get("track_info")
+        if isinstance(maybe_ti, dict):
+            ti = maybe_ti
+    isrc = str(ti.get("isrc") or "").strip().lower()
+    if isrc:
+        return f"isrc:{isrc}"
+    tid = str(ti.get("id") or "").strip()
+    if tid:
+        return f"id:{tid}"
+    uri = str(ti.get("uri") or "").strip()
+    if uri:
+        return f"uri:{uri}"
+    artist = " ".join(str(expected_artist or "").split()).lower()
+    track = " ".join(str(expected_track or "").split()).lower()
+    if not artist and not track:
+        return None
+    return f"nm:{artist}|{track}"
+
+
+def find_quarantine_siblings(quarantine_dir: str, entry_id: str) -> List[str]:
+    """Other entry ids that share ``entry_id``'s intended-target group key.
+
+    Returns the ids of every *other* quarantine entry whose
+    `expected_artist`/`expected_track` normalize to the same key as
+    ``entry_id`` (see :func:`quarantine_group_key`). Excludes ``entry_id``
+    itself. Returns ``[]`` when the entry is missing, has an ungroupable
+    (``None``) key, or has no siblings. Never raises.
+    """
+    if not entry_id:
+        return []
+    entries = list_quarantine_entries(quarantine_dir)
+    target_key = None
+    for e in entries:
+        if e.get("id") == entry_id:
+            target_key = e.get("group_key")
+            break
+    if target_key is None:
+        return []
+    return [
+        e["id"]
+        for e in entries
+        if e.get("id") != entry_id and e.get("group_key") == target_key
+    ]
+
+
 def list_quarantine_entries(quarantine_dir: str) -> List[Dict[str, Any]]:
     """Enumerate quarantined files paired with their sidecars.
 
@@ -213,6 +290,11 @@ def list_quarantine_entries(quarantine_dir: str) -> List[Dict[str, Any]]:
                 "reason": sidecar.get("quarantine_reason", "Unknown reason"),
                 "expected_track": sidecar.get("expected_track", ""),
                 "expected_artist": sidecar.get("expected_artist", ""),
+                "group_key": quarantine_group_key(
+                    sidecar.get("expected_artist", ""),
+                    sidecar.get("expected_track", ""),
+                    ctx,
+                ),
                 "timestamp": sidecar.get("timestamp", ""),
                 "size_bytes": size_bytes,
                 "has_full_context": isinstance(sidecar.get("context"), dict),
diff --git a/tests/imports/test_quarantine_management.py b/tests/imports/test_quarantine_management.py
index 3615ca6a..7234c497 100644
--- a/tests/imports/test_quarantine_management.py
+++ b/tests/imports/test_quarantine_management.py
@@ -5,9 +5,11 @@ from core.imports.quarantine import (
     approve_quarantine_entry,
     delete_quarantine_entry,
     entry_id_from_quarantined_filename,
+    find_quarantine_siblings,
     get_quarantine_entry_stream_info,
     get_quarantined_source_keys,
     list_quarantine_entries,
+    quarantine_group_key,
     recover_to_staging,
     serialize_quarantine_context,
 )
@@ -71,18 +73,20 @@ def test_serialize_round_trips_through_json():
 # list_quarantine_entries
 # ──────────────────────────────────────────────────────────────────────
 
-def _write_entry(quarantine_dir, entry_id, original_name, *, with_context=False, trigger="integrity", reason="boom", file_bytes=b"X" * 100):
+def _write_entry(quarantine_dir, entry_id, original_name, *, with_context=False, trigger="integrity", reason="boom", file_bytes=b"X" * 100, expected_track="Track", expected_artist="Artist", context=None):
     qfile = quarantine_dir / f"{entry_id}_{original_name}.quarantined"
     qfile.write_bytes(file_bytes)
     sidecar = {
         "original_filename": original_name,
         "quarantine_reason": reason,
-        "expected_track": "Track",
-        "expected_artist": "Artist",
+        "expected_track": expected_track,
+        "expected_artist": expected_artist,
         "timestamp": "2026-05-14T12:00:00",
         "trigger": trigger,
     }
-    if with_context:
+    if context is not None:
+        sidecar["context"] = context
+    elif with_context:
         sidecar["context"] = {"track_info": {"name": "Track"}, "context_key": entry_id}
     sidecar_path = quarantine_dir / f"{entry_id}_{os.path.splitext(original_name)[0]}.json"
     sidecar_path.write_text(json.dumps(sidecar))
@@ -454,3 +458,95 @@ def test_move_with_retry_returns_false_on_missing_source(tmp_path):
     # attempts=1 keeps the test fast (no retry sleeps)
     assert _move_with_retry(str(tmp_path / "nope.flac"), str(tmp_path / "dst.flac"),
                             attempts=1, delay=0) is False
+
+
+# ──────────────────────────────────────────────────────────────────────
+# #876: grouping alternatives for one song — quarantine_group_key /
+# find_quarantine_siblings, and the group_key field on list entries.
+# ──────────────────────────────────────────────────────────────────────
+
+def test_group_key_prefers_isrc_over_everything():
+    ctx = {"track_info": {"isrc": "USRC12345678", "id": "spid", "uri": "spotify:track:x"}}
+    assert quarantine_group_key("Artist", "Track", ctx) == "isrc:usrc12345678"
+
+
+def test_group_key_falls_back_to_source_id_then_uri():
+    assert quarantine_group_key("A", "T", {"track_info": {"id": "abc123"}}) == "id:abc123"
+    assert quarantine_group_key("A", "T", {"track_info": {"uri": "spotify:track:z"}}) == "uri:spotify:track:z"
+
+
+def test_group_key_falls_back_to_normalized_name_without_context():
+    # Trivial case/whitespace differences still collapse to one key.
+    k1 = quarantine_group_key("Kendrick  Lamar", "DNA.")
+    k2 = quarantine_group_key("kendrick lamar", "dna.")
+    assert k1 == k2 == "nm:kendrick lamar|dna."
+
+
+def test_group_key_none_when_nothing_identifies_target():
+    assert quarantine_group_key("", "", {}) is None
+    assert quarantine_group_key("", "", None) is None
+
+
+def test_list_entries_carry_group_key(tmp_path):
+    _write_entry(tmp_path, "20260514_120000", "a.flac",
+                 context={"track_info": {"isrc": "USABC1234567"}})
+    entries = list_quarantine_entries(str(tmp_path))
+    assert entries[0]["group_key"] == "isrc:usabc1234567"
+
+
+def test_find_siblings_returns_same_target_attempts(tmp_path):
+    # Two failed source attempts at the SAME target track (same isrc) + an
+    # unrelated entry. Siblings of #2 = {#1}, never the unrelated one.
+    same = {"track_info": {"isrc": "USAAA0000001"}}
+    other = {"track_info": {"isrc": "USZZZ9999999"}}
+    q1, _ = _write_entry(tmp_path, "20260514_120000", "src1.flac", context=same)
+    q2, _ = _write_entry(tmp_path, "20260514_120001", "src2.flac", context=same)
+    _write_entry(tmp_path, "20260514_120002", "diff.flac", context=other)
+
+    id1 = entry_id_from_quarantined_filename(q1.name)
+    id2 = entry_id_from_quarantined_filename(q2.name)
+    assert find_quarantine_siblings(str(tmp_path), id2) == [id1]
+
+
+def test_find_siblings_groups_by_intended_target_not_file_tags(tmp_path):
+    # Same intended target (isrc) even though the bad files differ — that's
+    # the whole point: the file metadata is wrong, the target is constant.
+    same = {"track_info": {"isrc": "USAAA0000001", "name": "Whatever"}}
+    q1, _ = _write_entry(tmp_path, "20260514_120000", "garbage_wrong_song.flac", context=same)
+    q2, _ = _write_entry(tmp_path, "20260514_120001", "another_bad_rip.flac", context=same)
+    id1 = entry_id_from_quarantined_filename(q1.name)
+    id2 = entry_id_from_quarantined_filename(q2.name)
+    assert find_quarantine_siblings(str(tmp_path), id1) == [id2]
+
+
+def test_find_siblings_empty_for_ungroupable_entry(tmp_path):
+    # No id and blank expected fields -> None key -> never grouped.
+    q1, _ = _write_entry(tmp_path, "20260514_120000", "orphan.flac",
+                         expected_track="", expected_artist="")
+    id1 = entry_id_from_quarantined_filename(q1.name)
+    assert find_quarantine_siblings(str(tmp_path), id1) == []
+
+
+def test_find_siblings_empty_for_missing_entry(tmp_path):
+    _write_entry(tmp_path, "20260514_120000", "a.flac")
+    assert find_quarantine_siblings(str(tmp_path), "does_not_exist") == []
+
+
+def test_siblings_must_be_captured_before_accepted_entry_leaves_quarantine(tmp_path):
+    # Regression for the approve-endpoint ordering: approving RESTORES (moves)
+    # the accepted entry out of quarantine, after which an id-based sibling
+    # lookup for that id can't resolve its group_key and returns []. The
+    # endpoint therefore captures siblings BEFORE approving. This pins that
+    # invariant: lookup-before == sibling found, lookup-after == empty.
+    same = {"track_info": {"isrc": "USAAA0000001"}}
+    q1, _ = _write_entry(tmp_path, "20260514_120000", "a.flac", context=same)
+    q2, _ = _write_entry(tmp_path, "20260514_120001", "b.flac", context=same)
+    id1 = entry_id_from_quarantined_filename(q1.name)
+    id2 = entry_id_from_quarantined_filename(q2.name)
+
+    captured = find_quarantine_siblings(str(tmp_path), id1)  # while id1 present
+    assert captured == [id2]
+
+    delete_quarantine_entry(str(tmp_path), id1)  # simulate approve restoring it
+
+    assert find_quarantine_siblings(str(tmp_path), id1) == []  # too late now
diff --git a/web_server.py b/web_server.py
index e53a2711..7cfd9387 100644
--- a/web_server.py
+++ b/web_server.py
@@ -7556,6 +7556,18 @@ def approve_quarantine_item(entry_id):
             docker_resolve_path(config_manager.get('soulseek.download_path', './downloads')),
             'Transfer',
         )
+        _req = request.get_json(silent=True) or {}
+        # #876: capture the sibling alternatives BEFORE approving — the approve
+        # restores (moves) this entry's file out of quarantine, after which its
+        # own group_key can no longer be looked up by id. Read-only here; the
+        # actual deletion happens only after the re-import is safely kicked off.
+        _sibling_ids = []
+        if _req.get('remove_siblings'):
+            from core.imports.quarantine import find_quarantine_siblings
+            try:
+                _sibling_ids = find_quarantine_siblings(_get_quarantine_dir(), entry_id)
+            except Exception as sib_exc:
+                logger.warning(f"[Quarantine] Sibling lookup for {entry_id} failed: {sib_exc}")
         result = approve_quarantine_entry(_get_quarantine_dir(), entry_id, restore_dir)
         if result is None:
             return jsonify({
@@ -7574,7 +7586,6 @@ def approve_quarantine_item(entry_id):
         # context lost task_id/batch_id (the wrapper pops them before quarantine),
         # so we re-supply them here. Manager-tab approvals (no task_id) keep the
         # original inner-pipeline path.
-        _req = request.get_json(silent=True) or {}
         _task_id = (_req.get('task_id') or '').strip() or None
         _batch_id = None
         if _task_id:
@@ -7594,7 +7605,28 @@ def approve_quarantine_item(entry_id):
             _reprocess = lambda: _post_process_matched_download(context_key, context, restored_path)
         threading.Thread(target=_reprocess, daemon=True).start()
         logger.info(f"[Quarantine] Approved {entry_id} (original_trigger={trigger}, bypass=all, task={_task_id}) → re-running pipeline")
-        return jsonify({"success": True, "trigger_bypassed": "all", "original_trigger": trigger})
+        # #876: once one alternative for a song is accepted, the other
+        # quarantined attempts at the SAME intended target are redundant
+        # failed downloads of a track the user now owns. Delete the siblings
+        # captured above (scoped to the quarantine manager via `remove_siblings`
+        # — the download-modal chooser passes no flag and is unaffected).
+        removed_siblings = []
+        if _sibling_ids:
+            from core.imports.quarantine import delete_quarantine_entry
+            try:
+                for sib_id in _sibling_ids:
+                    if delete_quarantine_entry(_get_quarantine_dir(), sib_id):
+                        removed_siblings.append(sib_id)
+                if removed_siblings:
+                    logger.info(f"[Quarantine] Auto-removed {len(removed_siblings)} sibling alternative(s) of {entry_id}: {removed_siblings}")
+            except Exception as sib_exc:
+                logger.warning(f"[Quarantine] Sibling cleanup for {entry_id} failed: {sib_exc}")
+        return jsonify({
+            "success": True,
+            "trigger_bypassed": "all",
+            "original_trigger": trigger,
+            "removed_siblings": removed_siblings,
+        })
     except Exception as e:
         logger.error(f"[Quarantine] Error approving {entry_id}: {e}")
         return jsonify({"success": False, "error": str(e)}), 500
diff --git a/webui/static/style.css b/webui/static/style.css
index 608ee724..6c344644 100644
--- a/webui/static/style.css
+++ b/webui/static/style.css
@@ -10530,6 +10530,63 @@ body.helper-mode-active #dashboard-activity-feed:hover {
     gap: 4px;
 }
 
+/* #876: quarantine "alternatives for one song" group — a collapsible parent
+   row wrapping the standard entry rows for each failed source attempt. */
+.lh-quarantine-group {
+    border-left: 3px solid rgba(var(--accent-rgb), 0.35);
+    border-radius: 8px;
+    margin-bottom: 4px;
+    background: rgba(var(--accent-rgb), 0.03);
+}
+
+.lh-quarantine-group-header {
+    display: flex;
+    align-items: center;
+    gap: 12px;
+    padding: 10px 12px;
+    cursor: pointer;
+    border-radius: 8px;
+    transition: background 0.15s ease;
+}
+
+.lh-quarantine-group-header:hover {
+    background: rgba(var(--accent-rgb), 0.06);
+}
+
+.lh-quarantine-group-text {
+    flex: 1;
+    min-width: 0;
+}
+
+.lh-quarantine-group-count {
+    flex-shrink: 0;
+    font-size: 12px;
+    color: rgba(var(--accent-rgb), 0.95);
+    border: 1px solid rgba(var(--accent-rgb), 0.4);
+    border-radius: 999px;
+    padding: 2px 10px;
+    white-space: nowrap;
+}
+
+.lh-quarantine-group-header .lh-expand-btn {
+    transition: transform 0.18s ease;
+}
+
+.lh-quarantine-group:not(.lh-qgroup-collapsed) .lh-quarantine-group-header .lh-expand-btn {
+    transform: rotate(180deg);
+}
+
+.lh-quarantine-group-members {
+    padding: 0 8px 8px 16px;
+    display: flex;
+    flex-direction: column;
+    gap: 4px;
+}
+
+.lh-quarantine-group.lh-qgroup-collapsed .lh-quarantine-group-members {
+    display: none;
+}
+
 .library-history-entry-row1 {
     display: flex;
     align-items: flex-start;
diff --git a/webui/static/wishlist-tools.js b/webui/static/wishlist-tools.js
index 8f95c421..c0ad23c1 100644
--- a/webui/static/wishlist-tools.js
+++ b/webui/static/wishlist-tools.js
@@ -3300,13 +3300,66 @@ async function loadQuarantineList() {
             list.innerHTML = '<div class="library-history-empty">🛡️<br><br>No quarantined files. Nice and clean.</div>';
             return;
         }
-        list.innerHTML = entries.map(renderQuarantineEntry).join('');
+        list.innerHTML = _groupQuarantineEntries(entries).map(renderQuarantineGroupOrEntry).join('');
     } catch (err) {
         console.error('Error loading quarantine entries:', err);
         list.innerHTML = '<div class="library-history-empty">Error loading quarantine</div>';
     }
 }
 
+// #876: bucket entries that are alternatives for the SAME intended target
+// (same backend group_key — derived from expected_artist/expected_track, the
+// track SoulSync was trying to fetch, not the bad file's own tags). Entries
+// with a null group_key (legacy/orphan, ungroupable) each stand alone.
+// Preserves the newest-first order the backend already sorted by.
+function _groupQuarantineEntries(entries) {
+    const groups = [];
+    const byKey = new Map();
+    for (const entry of entries) {
+        const key = entry.group_key;
+        if (!key) { groups.push({ key: null, members: [entry] }); continue; }
+        let g = byKey.get(key);
+        if (!g) { g = { key, members: [] }; byKey.set(key, g); groups.push(g); }
+        g.members.push(entry);
+    }
+    return groups;
+}
+
+function renderQuarantineGroupOrEntry(group) {
+    if (group.members.length === 1) return renderQuarantineEntry(group.members[0]);
+    return renderQuarantineGroup(group);
+}
+
+// A collapsible parent row for multiple alternatives of one song. Header shows
+// the shared track/artist + an alternatives count; members reuse the standard
+// entry markup so per-row Approve/Recover/Delete keep working unchanged.
+function renderQuarantineGroup(group) {
+    const first = group.members[0] || {};
+    const title = first.expected_track || first.original_filename || 'Unknown';
+    const artist = first.expected_artist || '';
+    const n = group.members.length;
+    const sub = artist ? escapeHtml(artist) : '';
+    // Reuse the album art the sidecar context carried, when any member has it.
+    const thumb = (group.members.find(m => m.thumb_url) || {}).thumb_url || '';
+    const thumbHtml = thumb
+        ? `<img class="library-history-thumb" src="${escapeHtml(thumb)}" alt="" onerror="this.style.display='none'">`
+        : '<div class="library-history-thumb-placeholder">🛡️</div>';
+    return `<div class="lh-quarantine-group lh-qgroup-collapsed">
+        <div class="lh-quarantine-group-header" onclick="this.parentElement.classList.toggle('lh-qgroup-collapsed')">
+            ${thumbHtml}
+            <div class="lh-quarantine-group-text">
+                <div class="library-history-entry-title">${escapeHtml(title)}</div>
+                <div class="library-history-entry-meta">${sub}</div>
+            </div>
+            <span class="lh-quarantine-group-count">${n} alternatives</span>
+            <span class="lh-expand-btn">&#x25BE;</span>
+        </div>
+        <div class="lh-quarantine-group-members">
+            ${group.members.map(renderQuarantineEntry).join('')}
+        </div>
+    </div>`;
+}
+
 function renderQuarantineEntry(entry) {
     const triggerLabels = { integrity: 'Duration / Integrity', acoustid: 'AcoustID Mismatch', bit_depth: 'Bit Depth Filter', unknown: 'Unknown' };
     const triggerColors = { integrity: '#facc15', acoustid: '#ef5350', bit_depth: '#fb923c', unknown: '#888' };
@@ -3388,12 +3441,20 @@ async function approveQuarantineEntry(entryId) {
     });
     if (!ok) return;
     try {
-        const r = await fetch(`/api/quarantine/${encodeURIComponent(entryId)}/approve`, { method: 'POST' });
+        const r = await fetch(`/api/quarantine/${encodeURIComponent(entryId)}/approve`, {
+            method: 'POST',
+            headers: { 'Content-Type': 'application/json' },
+            // #876: clear the other quarantined alternatives for this same song
+            // once one is accepted — they're redundant failed attempts now.
+            body: JSON.stringify({ remove_siblings: true }),
+        });
         const data = await r.json();
         if (!data.success) {
             showToast(`Approve failed: ${data.error}`, 'error');
         } else {
-            showToast(`Approved — skipped ${data.trigger_bypassed} check, re-running pipeline.`, 'success');
+            const removed = (data.removed_siblings || []).length;
+            const extra = removed ? ` — removed ${removed} other option${removed === 1 ? '' : 's'}.` : '';
+            showToast(`Approved — skipped ${data.trigger_bypassed} check, re-running pipeline.${extra}`, 'success');
         }
     } catch (err) {
         showToast(`Approve failed: ${err.message}`, 'error');

From 48e86a1a58c79cc4940e917de5749e790718f682 Mon Sep 17 00:00:00 2001
From: BoulderBadgeDad <batmanesignal@gmail.com>
Date: Mon, 15 Jun 2026 22:50:39 -0700
Subject: [PATCH 08/11] =?UTF-8?q?#874:=20wishlist=20ignore-list=20?=
 =?UTF-8?q?=E2=80=94=20stop=20auto-retrying=20removed/cancelled=20tracks?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

A user who removes a wishlist track, or cancels an in-flight wishlist
download, would have it re-added on the next auto cycle (watchlist scan,
failed-track capture, or the cancel handler's own re-add), so the same
release downloaded -> failed/cancelled -> re-queued forever.

Adds a TTL'd skip-gate (30 days), softer than the blocklist: it expires
so the track is reconsidered later, and never blocks a manual
force-download — only the automatic re-queue.

- core/wishlist/ignore.py: pure TTL/normalization/display logic + a
  best-effort orchestrator (no DB handle, caller passes now).
- database/music_database.py: migration-safe wishlist_ignore table +
  add/check/remove/list(+purge)/clear methods, and the gate in
  add_to_wishlist beside the blocklist guard. Fail-open throughout — an
  ignore error can never block a legitimate add; a manual add bypasses
  the gate AND clears the ignore.
- routes.py: user remove (single/album/batch) records an ignore. Hooked
  at the route layer, NOT the DB remove, so success-cleanup never
  ignores (regression-tested).
- web_server.py: cancel now ignores + removes from the wishlist instead
  of re-adding for endless retry; three /api/wishlist/ignore-list*
  endpoints.
- downloads.js: 'Ignored' modal (view / un-ignore / clear all).
- 13 tests: pure logic, DB seam, gate (block/bypass/fail-open),
  route wiring, and the success-cleanup-does-not-ignore regression.
---
 core/wishlist/ignore.py                | 162 +++++++++++++++++++++
 core/wishlist/routes.py                |  44 +++++-
 database/music_database.py             | 181 ++++++++++++++++++++++-
 tests/wishlist/test_wishlist_ignore.py | 190 +++++++++++++++++++++++++
 web_server.py                          |  83 +++++++++--
 webui/static/downloads.js              | 114 +++++++++++++++
 6 files changed, 755 insertions(+), 19 deletions(-)
 create mode 100644 core/wishlist/ignore.py
 create mode 100644 tests/wishlist/test_wishlist_ignore.py

diff --git a/core/wishlist/ignore.py b/core/wishlist/ignore.py
new file mode 100644
index 00000000..beced2c4
--- /dev/null
+++ b/core/wishlist/ignore.py
@@ -0,0 +1,162 @@
+"""Wishlist ignore-list — a TTL'd skip-gate for the wishlist (#874).
+
+When a user removes a track from the wishlist or cancels an in-flight
+wishlist download, SoulSync would otherwise re-add it on the next
+automatic cycle (watchlist scan, failed-track capture, or the cancel
+handler's own re-add), so the same release downloads → fails/cancels →
+re-queues forever. The ignore list records the user's "stop
+auto-grabbing this" intent, and the wishlist *add* path checks it,
+skipping automatic re-adds until the entry ages out.
+
+It is deliberately softer than the blocklist:
+  - it **expires** after ``IGNORE_TTL_DAYS`` so the track is re-attempted
+    again later rather than banned forever, and
+  - it **never blocks a manual force-download** — only the automatic
+    re-queue. (Manual downloads don't go through ``add_to_wishlist`` at
+    all, and an explicit manual *add* both bypasses the gate and clears
+    any existing ignore for the track.)
+
+This module is pure decision logic — no database handle and no clock of
+its own; the caller passes ``now``. The SQL lives in ``MusicDatabase``
+as a thin wrapper around these helpers, which keeps the TTL / id-matching
+rules unit-testable without a database.
+"""
+
+from __future__ import annotations
+
+from datetime import datetime, timedelta
+from typing import Any, Dict, Iterable, List, Optional, Set, Tuple
+
+IGNORE_TTL_DAYS = 30
+
+# Recognised reasons (free-text tolerated; these are the canonical two).
+REASON_REMOVED = "removed"
+REASON_CANCELLED = "cancelled"
+
+
+def normalize_ignore_id(track_id: Any) -> str:
+    """Canonical key for a wishlist track id.
+
+    The wishlist stores some ids as a composite ``<track_id>::<album_id>``
+    (when ``wishlist.allow_duplicate_tracks`` is on). The add-path gate
+    keys on the bare track id (``spotify_track_data['id']``), so we strip
+    the ``::album`` suffix here so an ignore recorded from a composite-id
+    wishlist row still matches the bare-id add attempt, and vice-versa.
+    Returns ``''`` for falsy/blank input.
+    """
+    s = str(track_id or "").strip()
+    if not s:
+        return ""
+    return s.split("::", 1)[0]
+
+
+def _parse_ts(value: Any) -> Optional[datetime]:
+    """Best-effort parse of a sqlite TIMESTAMP / ISO string / datetime."""
+    if isinstance(value, datetime):
+        return value
+    if not value:
+        return None
+    s = str(value).strip()
+    for fmt in (
+        "%Y-%m-%d %H:%M:%S",
+        "%Y-%m-%dT%H:%M:%S",
+        "%Y-%m-%d %H:%M:%S.%f",
+        "%Y-%m-%dT%H:%M:%S.%f",
+    ):
+        try:
+            return datetime.strptime(s, fmt)
+        except ValueError:
+            continue
+    try:
+        return datetime.fromisoformat(s)
+    except ValueError:
+        return None
+
+
+def is_expired(created_at: Any, now: datetime, ttl_days: int = IGNORE_TTL_DAYS) -> bool:
+    """True when an ignore entry created at ``created_at`` has aged past TTL.
+
+    Fail-SAFE in the gate's favour: an unparseable/blank timestamp is
+    treated as **expired** (returns True) so a corrupt row can never wedge
+    a track out of the wishlist permanently — the worst case is the
+    ignore silently lapses and the track becomes eligible again.
+    """
+    created = _parse_ts(created_at)
+    if created is None:
+        return True
+    return now >= created + timedelta(days=ttl_days)
+
+
+def active_ignored_ids(
+    rows: Iterable[Dict[str, Any]], now: datetime, ttl_days: int = IGNORE_TTL_DAYS
+) -> Set[str]:
+    """Set of normalized track ids whose ignore entry is still within TTL."""
+    out: Set[str] = set()
+    for row in rows or []:
+        tid = normalize_ignore_id(row.get("track_id"))
+        if tid and not is_expired(row.get("created_at"), now, ttl_days):
+            out.add(tid)
+    return out
+
+
+def is_ignored(
+    rows: Iterable[Dict[str, Any]], track_id: Any, now: datetime, ttl_days: int = IGNORE_TTL_DAYS
+) -> bool:
+    """Whether ``track_id`` matches an in-TTL entry among ``rows``."""
+    key = normalize_ignore_id(track_id)
+    if not key:
+        return False
+    return key in active_ignored_ids(rows, now, ttl_days)
+
+
+def extract_display(data: Any) -> Tuple[str, str]:
+    """Pull a (track_name, artist_name) pair from a Spotify-shaped dict.
+
+    Tolerates ``artists`` as a list of dicts or bare strings, and missing
+    fields. Used to give ignore-list rows a human label for the UI.
+    Returns ``('', '')`` when nothing usable is present.
+    """
+    if not isinstance(data, dict):
+        return "", ""
+    name = str(data.get("name") or "").strip()
+    artist = ""
+    artists = data.get("artists") or []
+    if isinstance(artists, list) and artists:
+        first = artists[0]
+        if isinstance(first, dict):
+            artist = str(first.get("name") or "").strip()
+        else:
+            artist = str(first or "").strip()
+    return name, artist
+
+
+def ignore_wishlist_track(
+    database: Any,
+    profile_id: int,
+    track_id: Any,
+    reason: str,
+    spotify_data: Any = None,
+) -> bool:
+    """Record an ignore entry for a wishlist track. Best-effort; never raises.
+
+    Copies the track's display name/artist from ``spotify_data`` when
+    provided (callers should capture it BEFORE removing the wishlist row,
+    since the row may be gone afterwards); otherwise leaves them blank.
+    Returns True when an entry was written.
+    """
+    key = normalize_ignore_id(track_id)
+    if not key or database is None:
+        return False
+    name, artist = extract_display(spotify_data or {})
+    try:
+        return bool(
+            database.add_to_wishlist_ignore(
+                key,
+                track_name=name,
+                artist_name=artist,
+                reason=reason or REASON_REMOVED,
+                profile_id=profile_id,
+            )
+        )
+    except Exception:
+        return False
diff --git a/core/wishlist/routes.py b/core/wishlist/routes.py
index c05a626c..a467676c 100644
--- a/core/wishlist/routes.py
+++ b/core/wishlist/routes.py
@@ -310,12 +310,30 @@ def remove_track_from_wishlist(
         if not spotify_track_id:
             return {"success": False, "error": "No spotify_track_id provided"}, 400
 
-        success = get_wishlist_service().remove_track_from_wishlist(
+        service = get_wishlist_service()
+        _db = getattr(service, "database", None)
+        # #874: capture the track's display info BEFORE removal (the row is
+        # gone afterwards) so the ignore-list entry carries a human label.
+        _ignore_data = None
+        try:
+            if _db is not None:
+                _ignore_data = _db.get_wishlist_spotify_data(
+                    spotify_track_id, profile_id=runtime.profile_id)
+        except Exception:
+            _ignore_data = None
+
+        success = service.remove_track_from_wishlist(
             spotify_track_id,
             profile_id=runtime.profile_id,
         )
 
         if success:
+            # #874: a user-initiated remove means "stop auto-requeuing this".
+            # Record a TTL'd ignore so the watchlist/auto-processor doesn't
+            # re-add it. Best-effort — never fails the remove.
+            from core.wishlist.ignore import ignore_wishlist_track, REASON_REMOVED
+            ignore_wishlist_track(_db, runtime.profile_id,
+                                  spotify_track_id, REASON_REMOVED, spotify_data=_ignore_data)
             runtime.logger.info("Successfully removed track from wishlist: %s", spotify_track_id)
             return {"success": True, "message": "Track removed from wishlist"}, 200
 
@@ -358,13 +376,20 @@ def remove_album_from_wishlist(
             if matched:
                 spotify_track_id = track.get("track_id") or track.get("spotify_track_id") or track.get("id")
                 if spotify_track_id:
-                    tracks_to_remove.append(spotify_track_id)
+                    # Keep the loaded spotify_data alongside the id so the #874
+                    # ignore entry can be labelled without a second DB read.
+                    tracks_to_remove.append((spotify_track_id, spotify_data))
 
+        from core.wishlist.ignore import ignore_wishlist_track, REASON_REMOVED
+        _db = getattr(wishlist_service, "database", None)
         removed_count = 0
         album_remove_pid = runtime.profile_id
-        for spotify_track_id in tracks_to_remove:
+        for spotify_track_id, track_spotify_data in tracks_to_remove:
             if wishlist_service.remove_track_from_wishlist(spotify_track_id, profile_id=album_remove_pid):
                 removed_count += 1
+                # #874: user removed the whole album → ignore each track.
+                ignore_wishlist_track(_db, album_remove_pid,
+                                      spotify_track_id, REASON_REMOVED, spotify_data=track_spotify_data)
 
         if removed_count > 0:
             runtime.logger.info("Successfully removed %s tracks from album %s", removed_count, album_id)
@@ -390,11 +415,22 @@ def remove_batch_from_wishlist(
         if not spotify_track_ids or not isinstance(spotify_track_ids, list):
             return {"success": False, "error": "Missing or invalid spotify_track_ids"}, 400
 
+        from core.wishlist.ignore import ignore_wishlist_track, REASON_REMOVED
+        service = get_wishlist_service()
+        _db = getattr(service, "database", None)
         removed = 0
         pid = runtime.profile_id
         for track_id in spotify_track_ids:
-            if get_wishlist_service().remove_track_from_wishlist(track_id, profile_id=pid):
+            # Capture label before the row is deleted (#874).
+            _data = None
+            try:
+                if _db is not None:
+                    _data = _db.get_wishlist_spotify_data(track_id, profile_id=pid)
+            except Exception:
+                _data = None
+            if service.remove_track_from_wishlist(track_id, profile_id=pid):
                 removed += 1
+                ignore_wishlist_track(_db, pid, track_id, REASON_REMOVED, spotify_data=_data)
 
         runtime.logger.info("Batch removed %s track(s) from wishlist", removed)
         return {
diff --git a/database/music_database.py b/database/music_database.py
index 1f6078fb..30d2c2b3 100644
--- a/database/music_database.py
+++ b/database/music_database.py
@@ -344,7 +344,29 @@ class MusicDatabase:
                     source_info TEXT  -- JSON of source context (playlist name, album info, etc.)
                 )
             """)
-            
+
+            # Wishlist ignore-list (#874): a TTL'd skip-gate. When a user
+            # removes a track from the wishlist or cancels an in-flight
+            # wishlist download, the track is recorded here so the automatic
+            # re-add paths (watchlist scan, failed-track capture, cancel
+            # re-add) skip it until the entry ages out (see core.wishlist.
+            # ignore.IGNORE_TTL_DAYS). Softer than `blocklist`: it expires
+            # and never blocks a manual force-download. Keyed on the bare
+            # track id; unique per (profile, track) so re-ignoring refreshes.
+            cursor.execute("""
+                CREATE TABLE IF NOT EXISTS wishlist_ignore (
+                    id INTEGER PRIMARY KEY AUTOINCREMENT,
+                    profile_id INTEGER NOT NULL DEFAULT 1,
+                    track_id TEXT NOT NULL,
+                    track_name TEXT DEFAULT '',
+                    artist_name TEXT DEFAULT '',
+                    reason TEXT DEFAULT 'removed',  -- 'removed' | 'cancelled'
+                    created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP,
+                    UNIQUE(profile_id, track_id)
+                )
+            """)
+            cursor.execute("CREATE INDEX IF NOT EXISTS idx_wishlist_ignore_profile ON wishlist_ignore (profile_id, track_id)")
+
             # Watchlist table for storing artists to monitor for new releases
             cursor.execute("""
                 CREATE TABLE IF NOT EXISTS watchlist_artists (
@@ -8851,6 +8873,21 @@ class MusicDatabase:
                                 _blocked[0], _blocked[1])
                     return False
 
+                # Ignore-list guard (#874): a user who removed or cancelled this
+                # track asked us to stop AUTO-requeuing it — every automatic
+                # re-add funnels through here, so one check covers them all.
+                # A *manual* add is explicit user intent → bypass the gate AND
+                # clear any stale ignore so it sticks. Fail-open: any error here
+                # must never block a legitimate wishlist add.
+                try:
+                    if source_type == 'manual':
+                        self.remove_from_wishlist_ignore(track_id, profile_id=profile_id)
+                    elif self.is_track_ignored(track_id, profile_id=profile_id):
+                        logger.info("Skipping wishlist add — track is on the ignore-list (#874): %s", track_id)
+                        return False
+                except Exception as _ignore_exc:
+                    logger.debug("Wishlist ignore-list check skipped (fail-open): %s", _ignore_exc)
+
                 from core.library import manual_library_match as _mlm
                 if _mlm.get_match_for_track(self, profile_id, spotify_track_data):
                     logger.info(
@@ -8995,7 +9032,147 @@ class MusicDatabase:
         except Exception as e:
             logger.error(f"Error removing track from wishlist: {e}")
             return False
-    
+
+    # ── Wishlist ignore-list (#874) ──────────────────────────────────────
+    # A TTL'd skip-gate consulted by add_to_wishlist so user-removed /
+    # user-cancelled tracks are not auto-re-queued. All methods fail-open
+    # (an error here must never block a legitimate wishlist add).
+
+    def add_to_wishlist_ignore(self, track_id: str, track_name: str = "",
+                               artist_name: str = "", reason: str = "removed",
+                               profile_id: int = 1) -> bool:
+        """Record (or refresh) an ignore entry for a wishlist track id.
+
+        Keyed on the bare track id; UNIQUE(profile_id, track_id) means a
+        repeat ignore replaces the row and so refreshes its TTL clock.
+        """
+        from core.wishlist.ignore import normalize_ignore_id
+        key = normalize_ignore_id(track_id)
+        if not key:
+            return False
+        try:
+            with self._get_connection() as conn:
+                cursor = conn.cursor()
+                cursor.execute("""
+                    INSERT OR REPLACE INTO wishlist_ignore
+                    (profile_id, track_id, track_name, artist_name, reason, created_at)
+                    VALUES (?, ?, ?, ?, ?, CURRENT_TIMESTAMP)
+                """, (profile_id, key, track_name or "", artist_name or "", reason or "removed"))
+                conn.commit()
+                logger.info("Added track to wishlist ignore-list (%s): '%s' [%s]",
+                            reason or "removed", track_name or key, key)
+                return True
+        except Exception as e:
+            logger.error("Error adding to wishlist ignore-list: %s", e)
+            return False
+
+    def is_track_ignored(self, track_id: str, profile_id: int = 1,
+                         ttl_days: Optional[int] = None) -> bool:
+        """Whether ``track_id`` has a non-expired ignore entry. Fail-open False."""
+        from core.wishlist.ignore import normalize_ignore_id, is_expired, IGNORE_TTL_DAYS
+        ttl = IGNORE_TTL_DAYS if ttl_days is None else ttl_days
+        key = normalize_ignore_id(track_id)
+        if not key:
+            return False
+        try:
+            with self._get_connection() as conn:
+                cursor = conn.cursor()
+                cursor.execute(
+                    "SELECT created_at FROM wishlist_ignore WHERE profile_id = ? AND track_id = ?",
+                    (profile_id, key))
+                row = cursor.fetchone()
+                if not row:
+                    return False
+                return not is_expired(row["created_at"], datetime.now(), ttl)
+        except Exception as e:
+            logger.debug("is_track_ignored failed open: %s", e)
+            return False
+
+    def remove_from_wishlist_ignore(self, track_id: str, profile_id: int = 1) -> bool:
+        """Un-ignore a track (manual override / UI action). Returns True if a row went."""
+        from core.wishlist.ignore import normalize_ignore_id
+        key = normalize_ignore_id(track_id)
+        if not key:
+            return False
+        try:
+            with self._get_connection() as conn:
+                cursor = conn.cursor()
+                cursor.execute(
+                    "DELETE FROM wishlist_ignore WHERE profile_id = ? AND track_id = ?",
+                    (profile_id, key))
+                conn.commit()
+                return cursor.rowcount > 0
+        except Exception as e:
+            logger.error("Error removing from wishlist ignore-list: %s", e)
+            return False
+
+    def get_wishlist_ignore(self, profile_id: int = 1,
+                            ttl_days: Optional[int] = None) -> List[Dict[str, Any]]:
+        """Active (non-expired) ignore entries, newest first; purges lapsed rows."""
+        from core.wishlist.ignore import is_expired, IGNORE_TTL_DAYS
+        ttl = IGNORE_TTL_DAYS if ttl_days is None else ttl_days
+        try:
+            with self._get_connection() as conn:
+                cursor = conn.cursor()
+                cursor.execute(
+                    "SELECT track_id, track_name, artist_name, reason, created_at "
+                    "FROM wishlist_ignore WHERE profile_id = ? ORDER BY created_at DESC",
+                    (profile_id,))
+                rows = cursor.fetchall()
+                now = datetime.now()
+                active, expired_ids = [], []
+                for r in rows:
+                    if is_expired(r["created_at"], now, ttl):
+                        expired_ids.append(r["track_id"])
+                    else:
+                        active.append({
+                            "track_id": r["track_id"],
+                            "track_name": r["track_name"] or "",
+                            "artist_name": r["artist_name"] or "",
+                            "reason": r["reason"] or "removed",
+                            "created_at": r["created_at"],
+                        })
+                # Opportunistic housekeeping so the table can't grow unbounded.
+                if expired_ids:
+                    cursor.executemany(
+                        "DELETE FROM wishlist_ignore WHERE profile_id = ? AND track_id = ?",
+                        [(profile_id, tid) for tid in expired_ids])
+                    conn.commit()
+                return active
+        except Exception as e:
+            logger.error("Error reading wishlist ignore-list: %s", e)
+            return []
+
+    def clear_wishlist_ignore(self, profile_id: int = 1) -> int:
+        """Drop every ignore entry for a profile. Returns rows removed."""
+        try:
+            with self._get_connection() as conn:
+                cursor = conn.cursor()
+                cursor.execute("DELETE FROM wishlist_ignore WHERE profile_id = ?", (profile_id,))
+                conn.commit()
+                return cursor.rowcount
+        except Exception as e:
+            logger.error("Error clearing wishlist ignore-list: %s", e)
+            return 0
+
+    def get_wishlist_spotify_data(self, track_id: str, profile_id: int = 1) -> Dict[str, Any]:
+        """Parsed ``spotify_data`` for a wishlist row, or {}. Used to label an
+        ignore entry with the track's name/artist before the row is removed."""
+        try:
+            with self._get_connection() as conn:
+                cursor = conn.cursor()
+                cursor.execute(
+                    "SELECT spotify_data FROM wishlist_tracks WHERE spotify_track_id = ? AND profile_id = ?",
+                    (track_id, profile_id))
+                row = cursor.fetchone()
+                if not row or not row["spotify_data"]:
+                    return {}
+                data = json.loads(row["spotify_data"])
+                return data if isinstance(data, dict) else {}
+        except Exception as e:
+            logger.debug("get_wishlist_spotify_data failed: %s", e)
+            return {}
+
     def get_wishlist_tracks(self, limit: Optional[int] = None, profile_id: int = 1,
                             offset: int = 0, category: Optional[str] = None) -> List[Dict[str, Any]]:
         """Get tracks in the wishlist for the given profile, ordered by date added
diff --git a/tests/wishlist/test_wishlist_ignore.py b/tests/wishlist/test_wishlist_ignore.py
new file mode 100644
index 00000000..783511f2
--- /dev/null
+++ b/tests/wishlist/test_wishlist_ignore.py
@@ -0,0 +1,190 @@
+"""#874 — wishlist ignore-list: TTL skip-gate for user-removed/cancelled tracks.
+
+Two layers:
+  * pure logic (core.wishlist.ignore) — TTL, id-normalization, display extract
+  * DB seam (MusicDatabase on a temp db) — add/check/remove/list/clear, the
+    add_to_wishlist gate, the manual-add bypass, and the regression that the
+    success-cleanup removal path does NOT ignore.
+"""
+
+from datetime import datetime, timedelta
+
+import pytest
+
+from core.wishlist.ignore import (
+    IGNORE_TTL_DAYS,
+    REASON_CANCELLED,
+    REASON_REMOVED,
+    active_ignored_ids,
+    extract_display,
+    is_expired,
+    is_ignored,
+    normalize_ignore_id,
+)
+from database.music_database import MusicDatabase
+
+
+# ── pure logic ──────────────────────────────────────────────────────────
+
+def test_normalize_strips_composite_album_suffix():
+    assert normalize_ignore_id("track123::album456") == "track123"
+    assert normalize_ignore_id("  track123  ") == "track123"
+    assert normalize_ignore_id("") == ""
+    assert normalize_ignore_id(None) == ""
+
+
+def test_is_expired_true_past_ttl_false_within():
+    now = datetime(2026, 6, 15, 12, 0, 0)
+    fresh = (now - timedelta(days=5)).strftime("%Y-%m-%d %H:%M:%S")
+    stale = (now - timedelta(days=40)).strftime("%Y-%m-%d %H:%M:%S")
+    assert is_expired(fresh, now) is False
+    assert is_expired(stale, now) is True
+
+
+def test_is_expired_unparseable_is_treated_expired_fail_open():
+    # Corrupt timestamp must lapse (never wedge a track out of the wishlist).
+    assert is_expired("not-a-date", datetime(2026, 6, 15)) is True
+    assert is_expired("", datetime(2026, 6, 15)) is True
+    assert is_expired(None, datetime(2026, 6, 15)) is True
+
+
+def test_is_ignored_matches_composite_and_bare_ids():
+    now = datetime(2026, 6, 15, 12, 0, 0)
+    created = (now - timedelta(days=1)).strftime("%Y-%m-%d %H:%M:%S")
+    rows = [{"track_id": "abc", "created_at": created}]
+    # Stored bare, queried with composite (and vice-versa) — both match.
+    assert is_ignored(rows, "abc::album9", now) is True
+    rows2 = [{"track_id": "abc", "created_at": created}]
+    assert is_ignored(rows2, "abc", now) is True
+    assert is_ignored(rows2, "different", now) is False
+
+
+def test_active_ignored_ids_drops_expired():
+    now = datetime(2026, 6, 15, 12, 0, 0)
+    fresh = (now - timedelta(days=2)).strftime("%Y-%m-%d %H:%M:%S")
+    stale = (now - timedelta(days=99)).strftime("%Y-%m-%d %H:%M:%S")
+    rows = [
+        {"track_id": "keep", "created_at": fresh},
+        {"track_id": "drop", "created_at": stale},
+    ]
+    assert active_ignored_ids(rows, now) == {"keep"}
+
+
+def test_extract_display_handles_dict_and_string_artists():
+    assert extract_display({"name": "Song", "artists": [{"name": "A"}]}) == ("Song", "A")
+    assert extract_display({"name": "Song", "artists": ["B"]}) == ("Song", "B")
+    assert extract_display({}) == ("", "")
+    assert extract_display(None) == ("", "")
+
+
+# ── DB seam (temp database — never the live db) ─────────────────────────
+
+@pytest.fixture
+def db(tmp_path):
+    return MusicDatabase(str(tmp_path / "m.db"))
+
+
+def _track(track_id="t1", name="Some Song", artist="Some Artist", album_id="alb1"):
+    return {
+        "id": track_id,
+        "name": name,
+        "artists": [{"name": artist}],
+        "album": {"id": album_id, "name": "Some Album", "images": []},
+    }
+
+
+def test_add_check_remove_roundtrip(db):
+    assert db.is_track_ignored("t1") is False
+    assert db.add_to_wishlist_ignore("t1", "Song", "Artist", REASON_REMOVED) is True
+    assert db.is_track_ignored("t1") is True
+    # Composite id of the same base track is also considered ignored.
+    assert db.is_track_ignored("t1::albX") is True
+    assert db.remove_from_wishlist_ignore("t1") is True
+    assert db.is_track_ignored("t1") is False
+
+
+def test_get_and_clear_ignore_list(db):
+    db.add_to_wishlist_ignore("a", "SongA", "ArtA", REASON_REMOVED)
+    db.add_to_wishlist_ignore("b", "SongB", "ArtB", REASON_CANCELLED)
+    entries = db.get_wishlist_ignore()
+    assert {e["track_id"] for e in entries} == {"a", "b"}
+    assert any(e["reason"] == "cancelled" for e in entries)
+    assert db.clear_wishlist_ignore() == 2
+    assert db.get_wishlist_ignore() == []
+
+
+def test_expired_entry_is_not_active_and_gets_purged(db):
+    db.add_to_wishlist_ignore("old", "Old", "Art", REASON_REMOVED)
+    # Backdate it well past the TTL.
+    with db._get_connection() as conn:
+        conn.execute(
+            "UPDATE wishlist_ignore SET created_at = ? WHERE track_id = ?",
+            ((datetime.now() - timedelta(days=IGNORE_TTL_DAYS + 5)).strftime("%Y-%m-%d %H:%M:%S"), "old"),
+        )
+        conn.commit()
+    assert db.is_track_ignored("old") is False        # lapsed
+    assert db.get_wishlist_ignore() == []             # and purged on read
+    with db._get_connection() as conn:
+        remaining = conn.execute("SELECT COUNT(*) c FROM wishlist_ignore").fetchone()["c"]
+    assert remaining == 0
+
+
+# ── the gate: add_to_wishlist honours the ignore-list ───────────────────
+
+def test_gate_blocks_auto_readd_but_manual_bypasses_and_clears(db):
+    track = _track("t1")
+    # Auto add works first time.
+    assert db.add_to_wishlist(track, source_type="playlist") is True
+    # User removes + ignores it.
+    db.remove_from_wishlist("t1")
+    db.add_to_wishlist_ignore("t1", "Some Song", "Some Artist", REASON_REMOVED)
+    # Auto re-add (watchlist / failed-capture / cancel) is now blocked.
+    assert db.add_to_wishlist(track, source_type="playlist") is False
+    assert db.is_track_ignored("t1") is True
+    # A MANUAL add bypasses the gate AND clears the ignore so it sticks.
+    assert db.add_to_wishlist(track, source_type="manual") is True
+    assert db.is_track_ignored("t1") is False
+
+
+def test_gate_failopen_when_ignore_table_errors(db, monkeypatch):
+    # If the ignore check raises, the add must still succeed (never block).
+    monkeypatch.setattr(db, "is_track_ignored", lambda *a, **k: (_ for _ in ()).throw(RuntimeError("boom")))
+    assert db.add_to_wishlist(_track("t9"), source_type="playlist") is True
+
+
+def test_route_remove_track_records_ignore(db, monkeypatch):
+    # Pin the route → ignore wiring (not just the DB layer): a user-initiated
+    # remove via the route function must drop the row AND ignore the track.
+    import types
+    from core.wishlist import routes as routes_module
+
+    db.add_to_wishlist(_track("rt1", name="Route Song", artist="Route Artist"),
+                       source_type="playlist")
+
+    class _Svc:
+        database = db
+
+        def remove_track_from_wishlist(self, tid, profile_id=1):
+            return db.remove_from_wishlist(tid, profile_id=profile_id)
+
+    monkeypatch.setattr(routes_module, "get_wishlist_service", lambda: _Svc())
+    runtime = types.SimpleNamespace(profile_id=1, logger=routes_module.module_logger)
+
+    payload, status = routes_module.remove_track_from_wishlist(runtime, "rt1")
+    assert status == 200
+    assert db.is_track_ignored("rt1") is True
+    # The ignore carries the captured label.
+    entry = db.get_wishlist_ignore()[0]
+    assert entry["track_name"] == "Route Song"
+    assert entry["reason"] == REASON_REMOVED
+
+
+def test_regression_success_cleanup_does_not_ignore(db):
+    # The post-download success path calls remove_from_wishlist directly — it
+    # must NOT add anything to the ignore-list (only user remove/cancel do).
+    track = _track("t5")
+    assert db.add_to_wishlist(track, source_type="playlist") is True
+    db.remove_from_wishlist("t5")                 # simulate success cleanup
+    assert db.is_track_ignored("t5") is False     # NOT ignored
+    # And so a later legitimate auto-add still works.
+    assert db.add_to_wishlist(track, source_type="playlist") is True
diff --git a/web_server.py b/web_server.py
index 7cfd9387..a873ac7b 100644
--- a/web_server.py
+++ b/web_server.py
@@ -16819,6 +16819,45 @@ def remove_batch_from_wishlist():
         logger.error(f"Error batch removing from wishlist: {e}")
         return jsonify({"success": False, "error": str(e)}), 500
 
+@app.route('/api/wishlist/ignore-list', methods=['GET'])
+def get_wishlist_ignore_list():
+    """#874: active (non-expired) wishlist ignore entries for this profile."""
+    try:
+        runtime = _build_wishlist_route_runtime()
+        entries = get_wishlist_service().database.get_wishlist_ignore(profile_id=runtime.profile_id)
+        from core.wishlist.ignore import IGNORE_TTL_DAYS
+        return jsonify({"success": True, "entries": entries, "ttl_days": IGNORE_TTL_DAYS})
+    except Exception as e:
+        logger.error(f"Error reading wishlist ignore-list: {e}")
+        return jsonify({"success": False, "error": str(e), "entries": []}), 500
+
+@app.route('/api/wishlist/ignore-list/remove', methods=['POST'])
+def remove_from_wishlist_ignore_list():
+    """#874: un-ignore a track so it can be auto-acquired again."""
+    try:
+        data = request.get_json() or {}
+        track_id = data.get('track_id') or data.get('spotify_track_id')
+        if not track_id:
+            return jsonify({"success": False, "error": "No track_id provided"}), 400
+        runtime = _build_wishlist_route_runtime()
+        ok = get_wishlist_service().database.remove_from_wishlist_ignore(
+            track_id, profile_id=runtime.profile_id)
+        return jsonify({"success": True, "removed": ok})
+    except Exception as e:
+        logger.error(f"Error removing from wishlist ignore-list: {e}")
+        return jsonify({"success": False, "error": str(e)}), 500
+
+@app.route('/api/wishlist/ignore-list/clear', methods=['POST'])
+def clear_wishlist_ignore_list():
+    """#874: clear the entire wishlist ignore-list for this profile."""
+    try:
+        runtime = _build_wishlist_route_runtime()
+        count = get_wishlist_service().database.clear_wishlist_ignore(profile_id=runtime.profile_id)
+        return jsonify({"success": True, "cleared": count})
+    except Exception as e:
+        logger.error(f"Error clearing wishlist ignore-list: {e}")
+        return jsonify({"success": False, "error": str(e)}), 500
+
 @app.route('/api/add-album-to-wishlist', methods=['POST'])
 def add_album_track_to_wishlist():
     """Endpoint to add a single track from an album to the wishlist."""
@@ -18956,26 +18995,44 @@ def _check_batch_completion_v2(batch_id):
 
 
 def _add_cancelled_task_to_wishlist(task):
-    """
-    Helper function to add cancelled task to wishlist.
-    Separated for clarity and error isolation.
+    """Handle a user-cancelled download's wishlist state.
+
+    #874: a manual cancel means "stop auto-retrying this release". The
+    previous behaviour re-added the cancelled track to the wishlist, which
+    the auto-processor then re-downloaded → re-cancelled → re-added,
+    forever. Instead we record a TTL'd ignore entry (so the watchlist /
+    auto-processor skip it) and drop it from the active wishlist. The user
+    can still force-download it manually (manual paths bypass the gate),
+    and the ignore lapses after core.wishlist.ignore.IGNORE_TTL_DAYS so it
+    is reconsidered later. Error-isolated; never raises into the caller.
     """
     if not task:
         return
-        
+
     try:
         from core.wishlist_service import get_wishlist_service
+        from core.wishlist.ignore import extract_display, REASON_CANCELLED
         wishlist_service = get_wishlist_service()
-        payload = _build_cancelled_task_wishlist_payload(task, profile_id=get_current_profile_id())
-        success = wishlist_service.add_spotify_track_to_wishlist(**payload)
-        
-        if success:
-            logger.info(f"[Atomic Cancel] Added '{task.get('track_info', {}).get('name')}' to wishlist")
-        else:
-            logger.error(f"[Atomic Cancel] Failed to add '{task.get('track_info', {}).get('name')}' to wishlist")
-            
+        profile_id = get_current_profile_id()
+        track_info = task.get('track_info', {}) or {}
+        track_id = track_info.get('id')
+        name = track_info.get('name')
+
+        if not track_id:
+            logger.warning("[Atomic Cancel] No track id on cancelled task — cannot ignore '%s'", name)
+            return
+
+        disp_name, disp_artist = extract_display(track_info)
+        wishlist_service.database.add_to_wishlist_ignore(
+            track_id, track_name=disp_name, artist_name=disp_artist,
+            reason=REASON_CANCELLED, profile_id=profile_id)
+        # Drop it from the active wishlist so the auto-processor stops
+        # re-attempting it; the gate then blocks any automatic re-add.
+        wishlist_service.remove_track_from_wishlist(track_id, profile_id=profile_id)
+        logger.info("[Atomic Cancel] Ignored (TTL) + removed from wishlist: '%s'", name or track_id)
+
     except Exception as e:
-        logger.error(f"[Atomic Cancel] Critical error adding to wishlist: {e}")
+        logger.error(f"[Atomic Cancel] Critical error handling cancelled task: {e}")
 
 @app.route('/api/playlists/<batch_id>/cancel_batch', methods=['POST'])
 def cancel_batch(batch_id):
diff --git a/webui/static/downloads.js b/webui/static/downloads.js
index 26303e3b..fe322a78 100644
--- a/webui/static/downloads.js
+++ b/webui/static/downloads.js
@@ -1186,6 +1186,9 @@ async function openWishlistOverviewModal() {
                         <button class="playlist-modal-btn playlist-modal-btn-warning" onclick="cleanupWishlistOverview()">
                             🧹 Cleanup Wishlist
                         </button>
+                        <button class="playlist-modal-btn playlist-modal-btn-secondary" onclick="openWishlistIgnoreModal()" title="Tracks you removed or cancelled — auto-skipped until they expire">
+                            🚫 Ignored
+                        </button>
                     </div>
                     <div class="playlist-modal-footer-right">
                         <button class="playlist-modal-btn playlist-modal-btn-secondary" onclick="closeWishlistOverviewModal()">Close</button>
@@ -1312,6 +1315,117 @@ function closeWishlistOverviewModal() {
     console.log('✅ Modal closed');
 }
 
+// ── #874: Wishlist ignore-list ("Ignored") modal ────────────────────────
+// Tracks the user removed from the wishlist or cancelled mid-download are
+// auto-skipped (not re-queued) until they expire. This modal lets the user
+// see what's currently ignored and lift the skip (un-ignore / clear all).
+
+async function openWishlistIgnoreModal() {
+    let modal = document.getElementById('wishlist-ignore-modal');
+    if (modal) modal.remove();
+    modal = document.createElement('div');
+    modal.id = 'wishlist-ignore-modal';
+    modal.className = 'modal-overlay';
+    modal.style.cssText = 'display:flex;position:fixed;inset:0;z-index:10050;align-items:center;justify-content:center;background:rgba(0,0,0,0.6);';
+    modal.innerHTML = `
+        <div class="playlist-modal-content" style="max-width:560px;width:90%;max-height:80vh;display:flex;flex-direction:column;">
+            <div class="playlist-modal-header">
+                <h2 style="margin:0;">🚫 Ignored Tracks</h2>
+                <button class="playlist-modal-btn playlist-modal-btn-secondary" onclick="closeWishlistIgnoreModal()">Close</button>
+            </div>
+            <p style="opacity:0.7;font-size:13px;margin:8px 16px 0;">Removed or cancelled tracks are skipped by auto-download until they expire. Un-ignore to allow auto-download again (you can always download manually).</p>
+            <div id="wishlist-ignore-list" class="playlist-tracks-scroll" style="flex:1;overflow-y:auto;padding:12px 16px;">
+                <div class="loading-indicator">Loading...</div>
+            </div>
+            <div class="playlist-modal-footer">
+                <div class="playlist-modal-footer-left">
+                    <button id="wishlist-ignore-clear-btn" class="playlist-modal-btn playlist-modal-btn-danger" onclick="clearWishlistIgnoreList()" style="display:none;">Clear All</button>
+                </div>
+            </div>
+        </div>`;
+    document.body.appendChild(modal);
+    modal.addEventListener('click', (e) => { if (e.target === modal) closeWishlistIgnoreModal(); });
+    await loadWishlistIgnoreList();
+}
+
+function closeWishlistIgnoreModal() {
+    const modal = document.getElementById('wishlist-ignore-modal');
+    if (modal) modal.remove();
+}
+
+async function loadWishlistIgnoreList() {
+    const list = document.getElementById('wishlist-ignore-list');
+    const clearBtn = document.getElementById('wishlist-ignore-clear-btn');
+    if (!list) return;
+    try {
+        const resp = await fetch('/api/wishlist/ignore-list');
+        const data = await resp.json();
+        const entries = (data && data.entries) || [];
+        if (clearBtn) clearBtn.style.display = entries.length ? '' : 'none';
+        if (!entries.length) {
+            list.innerHTML = '<div class="playlist-empty-state" style="text-align:center;opacity:0.6;padding:30px 0;">🎉<br><br>Nothing ignored.</div>';
+            return;
+        }
+        const ttl = (data && data.ttl_days) || 30;
+        list.innerHTML = entries.map(e => {
+            const title = escapeHtml(e.track_name || e.track_id || 'Unknown');
+            const artist = escapeHtml(e.artist_name || '');
+            const reason = e.reason === 'cancelled' ? 'Cancelled' : 'Removed';
+            const tid = escapeHtml(String(e.track_id || ''));
+            return `<div class="wishlist-ignore-row" style="display:flex;align-items:center;gap:10px;padding:8px 4px;border-bottom:1px solid rgba(255,255,255,0.06);">
+                <div style="flex:1;min-width:0;">
+                    <div style="font-weight:600;white-space:nowrap;overflow:hidden;text-overflow:ellipsis;">${title}</div>
+                    <div style="opacity:0.6;font-size:12px;white-space:nowrap;overflow:hidden;text-overflow:ellipsis;">${artist}${artist ? ' · ' : ''}${reason} · skips ${ttl}d</div>
+                </div>
+                <button class="playlist-modal-btn playlist-modal-btn-secondary" style="flex-shrink:0;" data-track-id="${tid}" onclick="unignoreWishlistTrack(this.dataset.trackId)">Un-ignore</button>
+            </div>`;
+        }).join('');
+    } catch (err) {
+        list.innerHTML = '<div class="playlist-empty-state" style="text-align:center;opacity:0.6;padding:30px 0;">Error loading ignored tracks</div>';
+    }
+}
+
+async function unignoreWishlistTrack(trackId) {
+    if (!trackId) return;
+    try {
+        const resp = await fetch('/api/wishlist/ignore-list/remove', {
+            method: 'POST',
+            headers: { 'Content-Type': 'application/json' },
+            body: JSON.stringify({ track_id: trackId }),
+        });
+        const data = await resp.json();
+        if (data && data.success) {
+            showToast('Track un-ignored — it can be auto-downloaded again.', 'success');
+            await loadWishlistIgnoreList();
+        } else {
+            showToast(`Un-ignore failed: ${(data && data.error) || 'unknown'}`, 'error');
+        }
+    } catch (err) {
+        showToast(`Un-ignore failed: ${err.message}`, 'error');
+    }
+}
+
+async function clearWishlistIgnoreList() {
+    if (!await showConfirmDialog({
+        title: 'Clear Ignored List',
+        message: 'Allow all currently-ignored tracks to be auto-downloaded again?',
+        confirmText: 'Clear All',
+        cancelText: 'Cancel',
+    })) return;
+    try {
+        const resp = await fetch('/api/wishlist/ignore-list/clear', { method: 'POST' });
+        const data = await resp.json();
+        if (data && data.success) {
+            showToast(`Cleared ${data.cleared || 0} ignored track(s).`, 'success');
+            await loadWishlistIgnoreList();
+        } else {
+            showToast(`Clear failed: ${(data && data.error) || 'unknown'}`, 'error');
+        }
+    } catch (err) {
+        showToast(`Clear failed: ${err.message}`, 'error');
+    }
+}
+
 async function cleanupWishlistOverview() {
     console.log('🧹 cleanupWishlistOverview() called');
 

From d15b3a185d2f7be1369c9861300b76697ce6b712 Mon Sep 17 00:00:00 2001
From: BoulderBadgeDad <batmanesignal@gmail.com>
Date: Mon, 15 Jun 2026 23:35:48 -0700
Subject: [PATCH 09/11] Track "01" bug: recover real track position instead of
 fabricating 1
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Single tracks (esp. Deezer-sourced) imported as "01 - Title" regardless
of their real album position — e.g. Fly Away (track 2 of Greatest Hits)
landed as 01, littering album folders with duplicate "01" files.

Root cause: a Deezer single track is matched via /search/track, which
omits track_position, so the context never carried the real number; then
service.py + context.py fabricated a confident track_number=1 from that
gap. Because the resolver puts that first, the fake 1 beat the source.
It is source-agnostic (slskd-with-Deezer-metadata hits it too) — albums
work because /album/<id>/tracks DOES include positions.

Fix (at the shared import funnel, strictly additive):
- track_number.py: new read_embedded_track_number() (mutagen, local, no
  network) + an optional embedded_track_number arg on resolve_track_number.
  The downloaded file already carries the source-written position (deemix
  wrote it); consult it LAST — only when metadata AND the "NN - Title"
  filename both come up empty — so it can only fill the gap that would
  otherwise hit the default-1 floor. Never overrides a value the pre-fix
  resolver produced (no regression for correctly-named/mistagged files).
- pipeline.py: read the file tag at the resolve step and pass it in.
- De-poison: service.py:217 + context.py default to 0 (the existing
  "unknown" sentinel, like total_tracks), NOT 1 — so the fake 1 no longer
  blocks recovery. Frontend already treats falsy track_number as unknown
  (omits it), so this also drops the bogus "1." in the UI.

13 new resolver tests incl. the no-regression precedence guards; full
imports + wishlist suites green (583), no behavior change for albums.
---
 core/imports/context.py                     |   6 +-
 core/imports/pipeline.py                    |  12 ++-
 core/imports/track_number.py                |  74 +++++++++++--
 core/wishlist/service.py                    |   6 +-
 tests/imports/test_track_number_resolver.py | 112 +++++++++++++++++++-
 5 files changed, 196 insertions(+), 14 deletions(-)

diff --git a/core/imports/context.py b/core/imports/context.py
index f718f014..024c3f36 100644
--- a/core/imports/context.py
+++ b/core/imports/context.py
@@ -319,7 +319,11 @@ def build_import_album_info(
         (album_info or {}).get("track_number")
         or track_info.get("track_number")
         or original_search.get("track_number")
-        or 1
+        # "Track 01" bug: default to 0 (the codebase's "unknown" sentinel,
+        # same as total_tracks below), NOT 1. A fabricated 1 looks
+        # authoritative and blocks the pipeline's downstream recovery
+        # (embedded file tag / resolve chain); 0 lets it fall through.
+        or 0
     )
     disc_number = (
         (album_info or {}).get("disc_number")
diff --git a/core/imports/pipeline.py b/core/imports/pipeline.py
index b6a077a9..ad6a0801 100644
--- a/core/imports/pipeline.py
+++ b/core/imports/pipeline.py
@@ -629,9 +629,17 @@ def post_process_matched_download(context_key, context, file_path, runtime, meta
         # See ``core/imports/track_number.py`` for the resolution
         # chain — pure function, unit-tested in isolation, single
         # place to fix the rule.
-        from core.imports.track_number import resolve_track_number
+        from core.imports.track_number import resolve_track_number, read_embedded_track_number
         track_info_for_resolve = context.get('track_info') if isinstance(context, dict) else None
-        track_number = resolve_track_number(album_info, track_info_for_resolve, file_path)
+        # "Track 01" bug: a single Deezer track is matched via an endpoint
+        # that omits track_position, so the context never carried the real
+        # number. The downloaded file itself does (deemix/source wrote it),
+        # so read it as a source between metadata and the filename guess.
+        embedded_track_number = read_embedded_track_number(file_path)
+        track_number = resolve_track_number(
+            album_info, track_info_for_resolve, file_path,
+            embedded_track_number=embedded_track_number,
+        )
         logger.debug(
             "Final track_number processing: source=%s album_info=%s resolved=%s",
             album_info.get('source', 'unknown'),
diff --git a/core/imports/track_number.py b/core/imports/track_number.py
index 43c98029..63b53b03 100644
--- a/core/imports/track_number.py
+++ b/core/imports/track_number.py
@@ -65,17 +65,64 @@ def _coerce_spotify_data(track_info: Any) -> dict:
     return {}
 
 
+def read_embedded_track_number(file_path: str) -> Optional[int]:
+    """Read the track position from a downloaded audio file's own tags.
+
+    Streaming sources (Deezer/deemix, Qobuz, Tidal) and most Soulseek
+    uploads write the correct album position into the file itself. That
+    tag is authoritative for the *source's* idea of the track's place on
+    its album — more reliable than a filename guess — so the resolver
+    consults it before falling back to the filename / default-1 floor.
+
+    Issue #874-adjacent / "Track 01" bug: a single Deezer track is matched
+    via Deezer's ``/search/track`` endpoint, which omits ``track_position``
+    (core/deezer_client.py), so the metadata context never carried the
+    real number — but the downloaded file *does* (deemix wrote it). This
+    recovers it with no network call.
+
+    Returns a positive int, or None when the file has no usable
+    tracknumber tag / can't be read. Never raises. Handles the common
+    ``"2/15"`` (number/total) form by taking the leading number.
+    """
+    if not file_path:
+        return None
+    try:
+        from mutagen import File as MutagenFile
+        audio = MutagenFile(file_path, easy=True)
+        if audio is None:
+            return None
+        raw = audio.get('tracknumber')
+        if isinstance(raw, list):
+            raw = raw[0] if raw else None
+        if raw is None:
+            return None
+        # "2/15" -> "2"; bare "2" -> "2".
+        text = str(raw).split('/', 1)[0].strip()
+        return _coerce_positive(text)
+    except Exception:
+        return None
+
+
 def resolve_track_number(
     album_info: Any,
     track_info: Any,
     file_path: str,
+    embedded_track_number: Any = None,
 ) -> Optional[int]:
     """Walk the resolution chain and return the first valid positive
     int found, or None when every source is missing / unusable.
 
-    Caller is responsible for the final default-1 floor — leaving
-    that out of this function so tests can pin "everything missing
+    Order: album_info -> track_info -> nested spotify_data -> filename ->
+    ``embedded_track_number`` (the source-written file tag, when the caller
+    supplies it). Caller is responsible for the final default-1 floor —
+    leaving that out of this function so tests can pin "everything missing
     returns None" separate from the floor behaviour.
+
+    ``embedded_track_number`` is passed in (not read here) so this stays a
+    pure function — the file I/O lives in :func:`read_embedded_track_number`.
+    It is consulted **last**, only when every other source came up empty, so
+    it can never override a value the pre-fix resolver already produced — it
+    only fills the gap that would otherwise hit the default-1 floor.
     """
     album_info = album_info if isinstance(album_info, dict) else {}
     track_info = track_info if isinstance(track_info, dict) else {}
@@ -96,10 +143,19 @@ def resolve_track_number(
     # default-1 floor is the single source of that fallback —
     # otherwise this resolver would silently fill 1 and the
     # downstream floor logic would have no effect.
-    if not file_path:
-        return None
-    try:
-        from_filename = extract_explicit_track_number(file_path)
-    except Exception:
-        from_filename = None
-    return _coerce_positive(from_filename)
+    if file_path:
+        try:
+            from_filename = extract_explicit_track_number(file_path)
+        except Exception:
+            from_filename = None
+        ff = _coerce_positive(from_filename)
+        if ff is not None:
+            return ff
+
+    # Embedded source-written file tag is consulted LAST — only when every
+    # other source (metadata + the ripped-album "NN - Title" filename) came
+    # up empty. This is deliberate: it can ONLY fill the gap that would
+    # otherwise hit the caller's default-1 floor, so it never overrides a
+    # value the pre-fix resolver would have used. A correctly-named file
+    # with a stale/wrong embedded tag is therefore never regressed.
+    return _coerce_positive(embedded_track_number)
diff --git a/core/wishlist/service.py b/core/wishlist/service.py
index 6dc8c859..df3be19a 100644
--- a/core/wishlist/service.py
+++ b/core/wishlist/service.py
@@ -214,7 +214,11 @@ class WishlistService:
                     "preview_url": track_data.get("preview_url") if isinstance(track_data, dict) else None,
                     "external_urls": track_data.get("external_urls", {}) if isinstance(track_data, dict) else {},
                     "popularity": track_data.get("popularity", 0) if isinstance(track_data, dict) else 0,
-                    "track_number": track_data.get("track_number", 1) if isinstance(track_data, dict) else 1,
+                    # "Track 01" bug: 0 = "unknown position", NOT a fabricated 1.
+                    # A fake 1 looks authoritative and blocks the import
+                    # pipeline's track-number recovery; 0 lets it recover the
+                    # real position (file tag / source lookup) before the floor.
+                    "track_number": track_data.get("track_number", 0) if isinstance(track_data, dict) else 0,
                     "disc_number": track_data.get("disc_number", 1) if isinstance(track_data, dict) else 1,
                 }
 
diff --git a/tests/imports/test_track_number_resolver.py b/tests/imports/test_track_number_resolver.py
index 088a4bcb..381d5f86 100644
--- a/tests/imports/test_track_number_resolver.py
+++ b/tests/imports/test_track_number_resolver.py
@@ -14,7 +14,7 @@ from __future__ import annotations
 
 from unittest.mock import patch
 
-from core.imports.track_number import resolve_track_number
+from core.imports.track_number import read_embedded_track_number, resolve_track_number
 
 
 # ---------------------------------------------------------------------------
@@ -226,3 +226,113 @@ def test_non_dict_track_info_treated_as_empty():
     """Defensive — non-dict track_info won't crash the resolver."""
     result = resolve_track_number({}, 'not a dict', '/dir/file.flac')
     assert result is None
+
+
+# ---------------------------------------------------------------------------
+# "Track 01" bug (Deezer single tracks): embedded file-tag source.
+# ---------------------------------------------------------------------------
+
+
+def test_embedded_tag_used_when_metadata_missing():
+    """The Deezer single-track case: context carried no position (search
+    endpoint omits track_position), but the downloaded file did. With the
+    metadata-context de-poisoned to 0/None, the embedded tag is consulted
+    BEFORE the filename guess."""
+    result = resolve_track_number(
+        album_info={'track_number': 0},        # de-poisoned "unknown"
+        track_info={'track_number': 0},
+        file_path='/dl/Lenny Kravitz - Fly Away.flac',  # no number prefix
+        embedded_track_number=2,
+    )
+    assert result == 2
+
+
+def test_real_metadata_still_beats_embedded_tag():
+    """No regression for album downloads: an authoritative album_info
+    position wins over the file tag (they normally agree, but the
+    context is the source of truth when present)."""
+    result = resolve_track_number(
+        album_info={'track_number': 5},
+        track_info={},
+        file_path='/dl/file.flac',
+        embedded_track_number=2,
+    )
+    assert result == 5
+
+
+def test_filename_beats_embedded_tag_no_regression():
+    """SAFETY: embedded tag is consulted LAST, so a correctly-named ripped
+    file ('05 - Song') with a stale/wrong embedded tag is NOT regressed —
+    the filename still wins, exactly as before the fix."""
+    result = resolve_track_number(
+        album_info={},
+        track_info={},
+        file_path='/dl/09 - Mislabelled.flac',
+        embedded_track_number=2,   # wrong/stale tag must NOT win
+    )
+    assert result == 9
+
+
+def test_embedded_only_fills_the_floor_gap():
+    """When metadata AND filename are both empty (the Deezer case), the
+    embedded tag fills what would otherwise be the blind default-1 floor."""
+    result = resolve_track_number(
+        album_info={}, track_info={},
+        file_path='/dl/Lenny Kravitz - Fly Away.flac',  # no number prefix
+        embedded_track_number=2,
+    )
+    assert result == 2
+
+
+def test_resolver_without_embedded_arg_is_backwards_compatible():
+    """The new parameter is optional — existing 3-arg callers unaffected."""
+    assert resolve_track_number({}, {'track_number': 4}, '/x.flac') == 4
+
+
+# read_embedded_track_number — mutagen-backed file tag reader.
+
+def _fake_mutagen(tag_value):
+    """Patch target factory: returns a callable standing in for
+    ``mutagen.File`` that yields an object whose .get('tracknumber')
+    returns tag_value (mimicking easy=True list-valued tags)."""
+    class _Audio(dict):
+        pass
+    def _factory(path, easy=False):
+        a = _Audio()
+        if tag_value is not None:
+            a['tracknumber'] = tag_value
+        return a
+    return _factory
+
+
+def test_read_embedded_handles_number_slash_total():
+    with patch('mutagen.File', _fake_mutagen(['2/15'])):
+        assert read_embedded_track_number('/x.flac') == 2
+
+
+def test_read_embedded_handles_bare_number():
+    with patch('mutagen.File', _fake_mutagen(['7'])):
+        assert read_embedded_track_number('/x.flac') == 7
+
+
+def test_read_embedded_none_when_tag_absent():
+    with patch('mutagen.File', _fake_mutagen(None)):
+        assert read_embedded_track_number('/x.flac') is None
+
+
+def test_read_embedded_none_when_file_unreadable():
+    def _factory(path, easy=False):
+        return None
+    with patch('mutagen.File', _factory):
+        assert read_embedded_track_number('/x.flac') is None
+
+
+def test_read_embedded_never_raises():
+    def _boom(path, easy=False):
+        raise RuntimeError('corrupt')
+    with patch('mutagen.File', _boom):
+        assert read_embedded_track_number('/x.flac') is None
+
+
+def test_read_embedded_empty_path_returns_none():
+    assert read_embedded_track_number('') is None

From 4bdde1248eb252f7e5d4fb6ec20eb115c2fd4ac7 Mon Sep 17 00:00:00 2001
From: BoulderBadgeDad <batmanesignal@gmail.com>
Date: Mon, 15 Jun 2026 23:52:39 -0700
Subject: [PATCH 10/11] #874 fixup: import get_wishlist_service in the
 ignore-list endpoints

ruff F821 caught a real NameError: the three /api/wishlist/ignore-list*
endpoints called get_wishlist_service() without the local import every
other call site in web_server.py uses, so they'd crash the moment the
Ignored modal queried them. Add the import; ruff check now clean.
---
 web_server.py | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/web_server.py b/web_server.py
index a873ac7b..449e9405 100644
--- a/web_server.py
+++ b/web_server.py
@@ -16823,6 +16823,7 @@ def remove_batch_from_wishlist():
 def get_wishlist_ignore_list():
     """#874: active (non-expired) wishlist ignore entries for this profile."""
     try:
+        from core.wishlist_service import get_wishlist_service
         runtime = _build_wishlist_route_runtime()
         entries = get_wishlist_service().database.get_wishlist_ignore(profile_id=runtime.profile_id)
         from core.wishlist.ignore import IGNORE_TTL_DAYS
@@ -16839,6 +16840,7 @@ def remove_from_wishlist_ignore_list():
         track_id = data.get('track_id') or data.get('spotify_track_id')
         if not track_id:
             return jsonify({"success": False, "error": "No track_id provided"}), 400
+        from core.wishlist_service import get_wishlist_service
         runtime = _build_wishlist_route_runtime()
         ok = get_wishlist_service().database.remove_from_wishlist_ignore(
             track_id, profile_id=runtime.profile_id)
@@ -16851,6 +16853,7 @@ def remove_from_wishlist_ignore_list():
 def clear_wishlist_ignore_list():
     """#874: clear the entire wishlist ignore-list for this profile."""
     try:
+        from core.wishlist_service import get_wishlist_service
         runtime = _build_wishlist_route_runtime()
         count = get_wishlist_service().database.clear_wishlist_ignore(profile_id=runtime.profile_id)
         return jsonify({"success": True, "cleared": count})

From 6c0d79a84a843195065af88909a8d10bf3dc7333 Mon Sep 17 00:00:00 2001
From: BoulderBadgeDad <batmanesignal@gmail.com>
Date: Tue, 16 Jun 2026 00:03:57 -0700
Subject: [PATCH 11/11] Release 2.7.3: bump version + What's New / version
 modal + docker-publish default tag

- web_server.py: _SOULSYNC_BASE_VERSION 2.7.2 -> 2.7.3
- helper.js: WHATS_NEW + VERSION_MODAL_SECTIONS rewritten for 2.7.3
  (current release + rolled-down 2.7.2/2.7.1/2.7.0 summary)
- docker-publish.yml: workflow_dispatch default tag -> 2.7.3

2.7.3 = Quality Upgrade Finder + #867 Tidal discovery + #880/#879/#877/
#876/#874/#870/#868 fixes + the 'Track 01' track-number recovery.
---
 .github/workflows/docker-publish.yml |  4 +-
 web_server.py                        |  2 +-
 webui/static/helper.js               | 97 ++++++++++++++--------------
 3 files changed, 50 insertions(+), 53 deletions(-)

diff --git a/.github/workflows/docker-publish.yml b/.github/workflows/docker-publish.yml
index 9ffbb51a..a216db03 100644
--- a/.github/workflows/docker-publish.yml
+++ b/.github/workflows/docker-publish.yml
@@ -9,9 +9,9 @@ on:
   workflow_dispatch:
     inputs:
       version_tag:
-        description: 'Version tag (e.g. 2.7.2)'
+        description: 'Version tag (e.g. 2.7.3)'
         required: true
-        default: '2.7.2'
+        default: '2.7.3'
 
 jobs:
   build-and-push:
diff --git a/web_server.py b/web_server.py
index 449e9405..6d9cdb3d 100644
--- a/web_server.py
+++ b/web_server.py
@@ -40,7 +40,7 @@ logger = setup_logging(_log_level, _log_path)
 
 # App version — single source of truth for backup metadata, system-info, update check, etc.
 # Semver: MAJOR.MINOR.PATCH. Bump at each dev→main release.
-_SOULSYNC_BASE_VERSION = "2.7.2"
+_SOULSYNC_BASE_VERSION = "2.7.3"
 
 def _build_version_string():
     """Append short commit hash to version when available (e.g. 2.35+abc1234)."""
diff --git a/webui/static/helper.js b/webui/static/helper.js
index b1168b2d..7774a8ac 100644
--- a/webui/static/helper.js
+++ b/webui/static/helper.js
@@ -3404,21 +3404,19 @@ function closeHelperSearch() {
 const WHATS_NEW = {
     // Convention: keep only the CURRENT release here, plus a single brief
     // "Earlier versions" summary entry. Don't accumulate old per-version blocks.
-    '2.7.2': [
-        { date: 'June 2026 — 2.7.2 release' },
-        { title: 'Organize playlists into folders', desc: 'optionally mirror each playlist into its own folder on disk — symlink (no extra space) or copy — so external players (plex / jellyfin / music assistant) see your soulsync playlists as real folders. rebuilds itself after every sync, prunes removed tracks, and there\'s a manual rebuild button in settings.', page: 'settings' },
-        { title: 'Export server playlists as M3U', desc: 'one-click "Export M3U" button in the Server Playlists compare/editor toolbar — writes a standard .m3u of the playlist\'s tracks and downloads it to your browser. handy for music assistant and friends.', page: 'sync' },
-        { title: 'Follow-only watchlist', desc: 'each watchlist artist now has an "auto-download" toggle (on by default). turn it off to just follow an artist — scans still discover and surface new releases, they just don\'t get auto-added to the wishlist.', page: 'artists' },
-        { title: 'Download from a SoundCloud link (#865)', desc: 'paste a soundcloud track url — including unlisted / private share links — into manual search and it resolves + downloads directly.', page: 'downloads' },
-        { title: 'YouTube playlists keep the artist (#863)', desc: 'youtube / youtube-music playlists that used to import as "Unknown Artist" now recover the real artist from the track\'s music metadata, the "Artist - Title" pattern, or the uploading channel — and fall back to the matched artist when youtube gives nothing.', page: 'sync' },
-        { title: 'Rename mirrored playlists', desc: 'a mirrored playlist now has a rename (✏️) button — set a custom name that changes how it shows in soulsync and how it syncs, survives upstream refreshes, and still tracks the same server playlist.', page: 'sync' },
-        { title: 'New maintenance jobs', desc: 'ReplayGain Filler (#437) computes + writes missing replaygain tags across your library, and Empty Folder Cleaner sweeps out leftover empty directories. both under Tools → library maintenance.', page: 'tools' },
-        { title: 'Export your watchlist + library', desc: 'one Export button with a scope selector dumps your watchlist roster and/or whole library roster to JSON / CSV / text.', page: 'artists' },
-        { title: 'Custom completed-download path for Torrent/Usenet (#857)', desc: 'point soulsync at the in-container folder your torrent/usenet client drops finished files into (e.g. a category subfolder), so completed grabs are found instead of going missing.', page: 'settings' },
-        { title: 'HiFi instances: restore + new working instance', desc: 'a "Restore Defaults" button re-adds any built-in instances you removed (keeps your own), the ✔/✖ controls have bigger tap targets, and a freshly-confirmed working instance is auto-pushed to everyone (thanks Sokhi).', page: 'settings' },
-        { title: 'Artist DB Record inspector', desc: 'an artist\'s detail page can now show the raw "DB Record" — everything the database knows about that artist — for debugging metadata.', page: 'library' },
-        { title: 'Fixes', desc: 'a hung database update self-heals now instead of wedging on "Starting..." forever (#859); Library Reorganize finally works on media-server libraries by falling back to tag mode (#862); spotify (no-auth) shows as connected and the dashboard test reports it correctly instead of claiming deezer; navidrome reconnects itself instead of latching disconnected; the orphan detector hard-bails on a mass-orphan flood; plus more #852 lock-screen hardening and login-password management in Manage Profiles.', page: 'settings' },
-        { title: 'Earlier versions', desc: '2.7.1 added download verification & a review queue (acoustid fingerprint-checks every download), closed the websocket login-bypass (#852), and the acoustid Relocate fix. 2.7.0 brought multi-user for real — per-profile streaming accounts, opt-in login, reverse-proxy support. before that the 2.6.x cycle brought the blocklist, the download-retry overhaul, Download Origins, Spotify-no-auth metadata, and Library Re-tag.' },
+    '2.7.3': [
+        { date: 'June 2026 — 2.7.3 release' },
+        { title: 'Quality Upgrade Finder', desc: 'a new findings-based job that scans your library for tracks you own in worse quality than is available and lets you upgrade them. matches by ISRC first, then album→track, then artist+title, with a direct track-ID tier, a dedup-skip, and a duration guard so it never swaps in the wrong song. replaces the old auto-acting Quality Scanner.', page: 'tools' },
+        { title: 'Tidal playlist discovery shows everything (#867)', desc: 'tidal playlist discovery used to cap at ~21 tracks — it now walks the whole playlist. and the discovery modal opens instantly in its "discovering" state instead of freezing the UI for ~10s on a pre-fetch.', page: 'sync' },
+        { title: 'Wishlist ignore-list (#874)', desc: 'remove a track from the wishlist or cancel an in-flight wishlist download and soulsync stops auto-re-adding it (it used to re-grab the same thing forever). softer than a blocklist — it expires after 30 days and never blocks a manual download. new "Ignored" view to see/undo it.', page: 'downloads' },
+        { title: 'Quarantine: group duplicates + auto-clear (#876)', desc: 'multiple failed attempts at the same song now collapse into one collapsible group; approve the good one and the other options auto-clear. plus the quarantine tab shows the correct count the moment you open it.', page: 'downloads' },
+        { title: 'Download Discography filters match Artist Detail (#877)', desc: 'the bulk-download modal now has the same Albums / EPs / Singles + Live / Compilations / Featured filters Artist Detail does — and the EPs toggle actually works now (it was always empty before).', page: 'artists' },
+        { title: 'Settings can no longer wipe your config (#879)', desc: 'if the settings page failed to load it used to fall back to blank defaults and then autosave the blanks over your real config. it now bails on a failed load and tells you to reload — your saved config is left untouched.', page: 'settings' },
+        { title: 'Tidal Favorites mirror grabs everything (#880)', desc: 'a transient tidal rate-limit (429) mid-sync used to truncate your Favorite Tracks mirror to ~98 of 500+. it retries the page with backoff now instead of stopping short.', page: 'sync' },
+        { title: 'Track numbers fixed (the "Track 01" bug)', desc: 'single tracks — especially deezer-sourced — imported as "01 - Title" regardless of their real album position, littering album folders with duplicate 01s. soulsync now recovers the real position from the downloaded file\'s own tag instead of guessing 1.', page: 'downloads' },
+        { title: 'More fixes', desc: 'deezer ARL stops appearing to "reset itself" — the saved token is tested directly, not a redacted mask (#870); an artist sharing a name with another no longer gets the wrong discography (disambiguated by the catalog you actually own, #868); a "Title - Remix" search now matches the base-titled track in your library; and a colon in a title (T:T) matches an underscore variant (T_T).', page: 'library' },
+        { title: 'Sidebar polish', desc: 'frosted-glass header blur, vertically-centered nav count badges, and the My-Accounts / Personal-Settings buttons are hidden for admins (who use the global app account anyway).', page: 'settings' },
+        { title: 'Earlier versions', desc: '2.7.2 added playlist-folder mirroring, server-playlist M3U export, follow-only watchlist, soundcloud-link + better youtube imports, and ReplayGain / Empty-Folder maintenance jobs. 2.7.1 added download verification (acoustid fingerprint-checks every download) + a review queue and closed the websocket login-bypass (#852). 2.7.0 made multi-user real — per-profile streaming accounts, opt-in login, reverse-proxy support. before that the 2.6.x cycle brought the blocklist, the download-retry overhaul, Download Origins, and Library Re-tag.' },
     ],
 };
 
@@ -3449,58 +3447,57 @@ const WHATS_NEW = {
 //                  usage_note?: 'optional hint shown at the bottom' }
 const VERSION_MODAL_SECTIONS = [
     {
-        title: "Organize playlists into folders",
-        description: "soulsync can now mirror each of your playlists into its own folder on disk, so external players (plex / jellyfin / music assistant) see them as real folders instead of just living inside soulsync.",
+        title: "Quality Upgrade Finder",
+        description: "a new findings-based job that scans your library for tracks you own in worse quality than is available, and lets you upgrade them. replaces the old auto-acting Quality Scanner.",
         features: [
-            "symlink (no extra disk space) or copy — your choice",
-            "rebuilds itself after every sync and prunes tracks you removed",
-            "separate output root + a manual rebuild button in settings",
+            "matches by ISRC first, then album→track, then artist+title — using the IDs enrichment already embedded",
+            "a direct track-ID tier for exact source matches",
+            "dedup-skip so it won't re-find the same upgrade, and a duration guard so it never swaps in the wrong track",
         ],
-        usage_note: "Settings → Playlists → organize into folders",
+        usage_note: "Tools → Quality Upgrade Finder",
     },
     {
-        title: "Better YouTube & SoundCloud imports",
-        description: "the two weak spots in playlist/link importing got fixed.",
+        title: "Wishlist ignore-list (#874)",
+        description: "remove a track from the wishlist, or cancel an in-flight wishlist download, and soulsync stops auto-re-adding it — it used to re-download the same release forever.",
         features: [
-            "#863 — youtube / youtube-music playlists that imported as \"Unknown Artist\" now recover the real artist from music metadata, the \"Artist - Title\" pattern, or the uploading channel (and fall back to the matched artist)",
-            "#865 — paste a soundcloud track link, including unlisted / private share urls, into manual search to download it directly",
+            "softer than a blocklist: it expires after 30 days and never blocks a manual download",
+            "a new \"Ignored\" view lets you see what's skipped and un-ignore anything",
+        ],
+        usage_note: "Wishlist → Ignored",
+    },
+    {
+        title: "Quarantine: group duplicates + auto-clear (#876)",
+        description: "the quarantine workflow got a cleanup pass.",
+        features: [
+            "multiple failed attempts at the same song collapse into one collapsible group",
+            "approve the good one and the other options auto-clear",
+            "the quarantine tab shows the correct count the moment you open it (no more stale 0)",
         ],
     },
     {
-        title: "Export server playlists as M3U",
-        description: "a one-click \"Export M3U\" button now sits in the Server Playlists compare/editor toolbar — writes a standard .m3u of the playlist and downloads it to your browser. great for music assistant.",
-        features: [],
-        usage_note: "Sync → Server Playlists → Export M3U",
-    },
-    {
-        title: "Follow-only watchlist + more",
-        description: "a grab-bag of requested features.",
+        title: "Tidal discovery & favorites (#867, #880)",
+        description: "two tidal sync fixes.",
         features: [
-            "per-artist \"auto-download\" toggle: turn it off to just follow an artist — scans still surface new releases, they just don't auto-add to the wishlist",
-            "rename mirrored playlists (✏️): a custom name that changes the display + sync name, survives refreshes, still tracks the server playlist",
-            "export your watchlist and/or whole library roster to JSON / CSV / text",
-            "ReplayGain Filler (#437) and Empty Folder Cleaner library-maintenance jobs",
-            "custom in-container completed-download path for Torrent / Usenet sources (#857)",
-            "HiFi instances: Restore Defaults button, bigger tap targets, and a new confirmed-working instance pushed to everyone",
-            "Artist detail \"DB Record\" inspector for debugging metadata",
+            "#867 — playlist discovery now walks the whole playlist instead of capping at ~21 tracks, and the modal opens instantly instead of freezing the UI for ~10s",
+            "#880 — a transient rate-limit (429) mid-sync no longer truncates your Favorite Tracks mirror to ~98 of 500+; it retries with backoff",
         ],
     },
     {
         title: "Fixes this release",
-        description: "a stack of issue fixes on top of 2.7.1.",
+        description: "a stack of issue fixes on top of 2.7.2.",
         features: [
-            "#859 — a hung database update self-heals instead of wedging on \"Starting...\" forever",
-            "#862 — Library Reorganize now works on media-server libraries (falls back to tag mode when there are no source IDs)",
-            "spotify (no-auth) shows as connected and the dashboard test reports it correctly, instead of claiming a deezer fallback",
-            "navidrome reconnects itself instead of latching \"disconnected\"",
-            "the orphan detector hard-bails on a mass-orphan flood instead of plowing ahead",
-            "more #852 lock-screen hardening + login-password management in Manage Profiles",
-            "Aria2 added to the torrent client list",
+            "the \"Track 01\" bug — single tracks (especially deezer) imported as \"01\" regardless of real album position; now recovered from the file's own tag",
+            "#879 — a failed Settings load can no longer overwrite your saved config with blank defaults",
+            "#877 — Download Discography filters now match Artist Detail (working EPs toggle + Live / Compilations / Featured)",
+            "#870 — deezer ARL stops appearing to \"reset itself\" (the saved token is tested, not a redacted mask)",
+            "#868 — an artist sharing a name with another no longer gets the wrong discography",
+            "Find & Add matches a \"Title - Remix\" search to the base-titled library track; a colon (T:T) matches an underscore variant (T_T)",
+            "sidebar polish: frosted-glass header, centered nav badges, admin-only cleanup",
         ],
     },
     {
-        title: "Earlier in 2.7.1 / 2.7.0",
-        description: "2.7.1 added download verification & an unverified review queue (acoustid fingerprint-checks every download against what you asked for), closed the websocket login-bypass (#852), and added the acoustid Relocate fix action. 2.7.0 made multi-user real: per-profile streaming accounts (My Accounts), auto-sync running as its owner, opt-in username/password login with recovery, and reverse-proxy support. before that, the 2.6.x cycle brought the blocklist, the download-retry overhaul, Download Origins, Spotify-no-auth metadata, and Library Re-tag.",
+        title: "Earlier in 2.7.2 / 2.7.1 / 2.7.0",
+        description: "2.7.2 added playlist-folder mirroring, server-playlist M3U export, follow-only watchlist, soundcloud-link + better youtube imports, and ReplayGain / Empty-Folder maintenance jobs. 2.7.1 added download verification (acoustid fingerprint-checks every download against what you asked for) + an unverified review queue, and closed the websocket login-bypass (#852). 2.7.0 made multi-user real: per-profile streaming accounts (My Accounts), opt-in username/password login with recovery, and reverse-proxy support. before that, the 2.6.x cycle brought the blocklist, the download-retry overhaul, Download Origins, and Library Re-tag.",
         features: [],
     },
 ];