Merge bugfix/downloads: repair 5 Downloads backend bugs (B1-B5)

Sticky errored asset re-download, ref_count leak on errored-job delete, admin
storage cap reads the DB value, and 400-not-500 on malformed edit/format specs.
Reviewed clean (a B2 concurrency race was caught + fixed). Local dev only — prod
publish waits for the end of the whole code-hygiene review.

backend/app/downloads/edit.py

@@ -79,13 +79,17 @@ def normalize_edit_spec(spec: dict | None) -> dict:
     # Single TRIM (one output file; the frontend fans a "separate" split out into N of these).
     trim = spec.get("trim") or {}
     if trim:
-        start = max(0.0, float(trim.get("start_s") or 0.0))
+        # Guard the numeric coercion like the crop/segments branches do — a malformed value must
-        end_raw = trim.get("end_s")
+        # drop the trim (→ empty spec → 400 EditError), not raise an unhandled 500.
+        try:
+            start = max(0.0, float(trim.get("start_s") or 0.0))
+            end_raw = trim.get("end_s")
+            end = float(end_raw) if end_raw is not None else None
+        except (TypeError, ValueError):
+            start, end = 0.0, None
         t: dict = {"start_s": round(start, 3)}
-        if end_raw is not None:
+        if end is not None and end > start:
-            end = float(end_raw)
+            t["end_s"] = round(end, 3)
-            if end > start:
-                t["end_s"] = round(end, 3)
         # A trim with only a start (open-ended) is valid (cut to the end).
         if t.get("start_s") or "end_s" in t:
             out["trim"] = t

backend/app/downloads/formats.py

@@ -62,7 +62,12 @@ def normalize(spec: dict) -> dict:
     if s["mode"] not in ("av", "v", "a"):
         s["mode"] = "av"
     if s["max_height"] is not None:
-        s["max_height"] = int(s["max_height"])
+        # A malformed max_height (non-numeric inline spec / stored profile) falls back to "best"
+        # rather than raising an unhandled 500 at enqueue.
+        try:
+            s["max_height"] = int(s["max_height"])
+        except (TypeError, ValueError):
+            s["max_height"] = None
     for b in ("embed_subs", "embed_chapters", "embed_thumbnail", "sponsorblock"):
         s[b] = bool(s[b])
     # Audio-only files can't embed subtitles/thumbnails as video; keep the flags meaningful.

backend/app/downloads/service.py

@@ -46,6 +46,14 @@ def get_or_create_asset(
     )
     asset = db.execute(stmt).scalar_one_or_none()
     if asset is not None:
+        # A previously-FAILED shared asset must be re-attempted for the new job about to hold it —
+        # reset it to pending (+clear the error) so the worker actually re-downloads instead of
+        # short-circuiting the new job with the asset's stale error. (Mirrors resume_download; without
+        # this a fresh enqueue of a once-failed (source,format) pair is permanently poisoned, since
+        # errored assets carry no expires_at and GC never clears them.)
+        if asset.status == "error":
+            asset.status = "pending"
+            asset.error = None
         return asset
     asset = MediaAsset(
         source_kind=source_kind,

backend/app/routes/downloads.py

@@ -15,6 +15,7 @@ from fastapi.responses import FileResponse
 from sqlalchemy import func, select
 from sqlalchemy.orm import Session
 
+from app import sysconfig
 from app.auth import admin_user, require_human
 from app.config import settings
 from app.db import get_db
@@ -499,9 +500,18 @@ def _release_asset(db: Session, job: DownloadJob) -> None:
     """Drop this job's hold on its asset when it leaves a holding state. Once NO job holds the
     asset anymore, delete the file + row immediately — a deleted download should free its disk,
     and the shared cache only needs to span *overlapping* holders (a later re-add just downloads
-    again). Idempotent: a job that's already left holding (canceled/error) is a no-op, so delete
+    again). `error` counts as holding: enqueue always +1'd ref_count and the worker never
-    after cancel doesn't double-release."""
+    decrements on failure (ref_count is the route's job), so an errored job releases here on
-    if not job.asset_id or job.status not in ("queued", "running", "paused", "done"):
+    delete — else its +1 leaks and a later resume→ready keeps the file pinned past its last holder.
+    Idempotent for the cancel path: cancel already released while the job was holding, then set it
+    `canceled` (∉ the set below), so delete-after-cancel is a no-op and never double-releases.
+
+    We DON'T delete an errored asset row here even at ref==0: another request may be concurrently
+    re-enqueuing the same (source,format) — get_or_create_asset would reset that row to `pending` and
+    +1 it, and deleting it under that would FK-null the new job's asset (a lost update on ref_count).
+    A ready asset is still freed at ref==0 (its file is the point); an orphaned errored row is cheap
+    (no file) and gets reused+reset by the next enqueue, or reclaimed by a later GC pass."""
+    if not job.asset_id or job.status not in ("queued", "running", "paused", "done", "error"):
         return
     asset = db.get(MediaAsset, job.asset_id)
     if asset is None:
@@ -876,7 +886,7 @@ def admin_storage(
         "ready_files": ready[0],
         "total_bytes": int(ready[1]),
         "total_assets": total_assets,
-        "total_cap_bytes": settings.download_total_max_bytes,
+        "total_cap_bytes": sysconfig.get_int(db, "download_total_max_bytes"),
         "per_user": [
             {"user_id": uid, "email": emails.get(uid), "footprint_bytes": quota.footprint(db, uid)}
             for uid, _ in per_user