fix(e12): unbreak the gate, and stop printing a command that can strand the restore

Round five. 47 -> 34 -> 15 -> 7 -> 6 raw findings, and this round's sharpest was the
loop's own blind spot: **the branch did not pass the gate.** `prettier --check` is one of
its eight lanes and I had been running the other seven, so an unformatted `onError` from
round one sat on the branch for four rounds — `siftlode publish` would have refused it,
with no --skip flag to get past.

**The restart hint could still strand the restored data.** Round four said the scripts no
longer print a command they cannot resolve; they did. The COMPOSE_FILE fallback only fires
when `docker-compose.yml` is ABSENT, so in a source checkout it never fires and the hint
expanded to a bare, pasteable `docker compose up -d` — which reads the other file, whose
volumes are `pgdata`/`downloads` rather than `siftlode_pgdata`/`siftlode_downloads`, and
recreates the database on an empty one. When the file is unresolved the scripts now print
`-f <the compose file you started with>`: not runnable, which is the point.

**A comment naming the wrong owner.** `formats.py` still said a zero budget is "a refusal
for enqueue to make" — the premise round one disproved, three files away from the two
docstrings that now say the opposite. Anyone acting on it would delete the worker's refusal
and re-open the original inversion, so it now says where the refusal actually lives.

Also: the reset-flag rule was rewritten in three consecutive rounds and pinned by nothing —
two tests now cover a fragment losing its token, both by a pasted verify link and by
navigating away entirely; `backup.ps1` deletes a partially-copied dump before throwing, the
guard round four gave the docs but not the script it mirrors; and one comment that read as
if the asset were untouched by a line four below it.

Gate green, all eight lanes: ruff, ruff format, 3x tsc, eslint, prettier, knip,
vitest 687, pytest 950.

backend/app/downloads/formats.py

@@ -164,9 +164,12 @@ def build_ydl_opts(
         opts.update(extra)
 
     # SEC-13: hard ceiling on a single download. AFTER the adapter merge, for the same reason
-    # `cookiefile` is — a source adapter must not be able to raise a limit the quota layer set. A
-    # non-positive budget is left unset: "0 bytes left" is a refusal for enqueue to make, not
-    # something to express here as an impossible yt-dlp filter.
+    # `cookiefile` is — a source adapter must not be able to raise a limit the quota layer set.
+    #
+    # A non-positive budget is left unset because 0 is not an expressible yt-dlp filter, NOT because
+    # nothing needs to happen: the worker refuses that job outright before extraction (`_OwnerRefused`
+    # → `quota_exceeded`). Enqueue cannot make that call — it ran while the user was still under the
+    # cap. Delete the worker's refusal and this branch silently uncaps the one user already over it.
     if max_bytes and max_bytes > 0:
         opts["max_filesize"] = max_bytes
 

backend/app/downloads/pipeline.py

@@ -567,9 +567,9 @@ def _download(
         _reset_asset_pending(asset_id)
         log.info("job %s aborted (pause/cancel)", job_id)
     except _OwnerRefused as exc:
-        # Fail this job only. The asset is untouched and goes back to `pending`, so another user's
-        # job for the same video is claimed and downloaded normally instead of inheriting a reason
-        # that is about someone else's account.
+        # Fail this job only. The asset is never marked `error`; it goes back to `pending` (below),
+        # so another user's job for the same video is claimed and downloaded normally instead of
+        # inheriting a reason that is about someone else's account.
         msg = str(exc)[:500]
         with SessionLocal() as db:
             service.finalize_if_running(

frontend/src/components/AdminUsers.tsx

@@ -350,10 +350,9 @@ function PendingAccess() {
     onError: (err) =>
       notify({
         level: "error",
-        message:
-          isEmailOff(err)
-            ? t("users.pendingAccess.resendEmailOff")
-            : t("users.pendingAccess.resendFailed"),
+        message: isEmailOff(err)
+          ? t("users.pendingAccess.resendEmailOff")
+          : t("users.pendingAccess.resendFailed"),
       }),
   });
 

frontend/src/components/accountPanel.dom.test.tsx

@@ -111,9 +111,7 @@ describe("the account panel reports a refused recovery action", () => {
 
     await clickAction("users.account.resend");
 
-    await waitFor(() =>
-      expect(errorMessages()).toContain("users.pendingAccess.resendEmailOff"),
-    );
+    await waitFor(() => expect(errorMessages()).toContain("users.pendingAccess.resendEmailOff"));
   });
 
   it("reports a genuine send failure as a failure, not as a sent mail", async () => {
@@ -137,8 +135,6 @@ describe("the account panel reports a refused recovery action", () => {
 
     await clickAction("users.account.activate");
 
-    await waitFor(() =>
-      expect(errorMessages()).toContain("users.pendingAccess.activateFailed"),
-    );
+    await waitFor(() => expect(errorMessages()).toContain("users.pendingAccess.activateFailed"));
   });
 });

frontend/src/lib/verifyEntry.dom.test.tsx

@@ -158,6 +158,39 @@ describe("confirming an emailed verification link", () => {
     expect(screen.getByTestId("reset").textContent).toBe("reset-pending");
   });
 
+  // The flag must not outlive the token it describes. Navigating away from a `#reset=` fragment —
+  // a pasted verify link, the back button — leaves the address bar with no reset token, so a
+  // still-true flag promises a recovery that signing out can no longer complete AND (being checked
+  // first by VerifyBanner) hides whatever arrived in its place.
+  it("drops the reset notice once the fragment carrying it is gone", async () => {
+    const verifyEmail = vi.fn().mockResolvedValue({ ok: true, email: "someone@example.com" });
+
+    const outcome = await mountAt("/#reset=going-away", verifyEmail);
+    expect(screen.getByTestId("reset").textContent).toBe("reset-pending");
+
+    await act(async () => {
+      window.history.replaceState(window.history.state, "", "/#verify=tok-next");
+      window.dispatchEvent(new HashChangeEvent("hashchange"));
+    });
+
+    expect(screen.getByTestId("reset").textContent).toBe("-");
+    expect(outcome()).toBe("ok:someone@example.com");
+  });
+
+  // …and a fragment change that carries neither token drops it too (the back button out of a
+  // reset link), which a one-way "set it when you see one" listener would not.
+  it("drops the reset notice when the fragment changes to no token at all", async () => {
+    await mountAt("/#reset=going-away", vi.fn());
+    expect(screen.getByTestId("reset").textContent).toBe("reset-pending");
+
+    await act(async () => {
+      window.history.replaceState(window.history.state, "", "/#somewhere-else");
+      window.dispatchEvent(new HashChangeEvent("hashchange"));
+    });
+
+    expect(screen.getByTestId("reset").textContent).toBe("-");
+  });
+
   // A reset link can arrive over an unread confirmation (and a slow /auth/verify can resolve after
   // one did), so both states can be set at once. Neither may eat the other: `dismiss` used to clear
   // both, which is how the reset guidance could be destroyed before it was ever rendered.

scripts/backup.ps1

@@ -46,7 +46,12 @@ if ($dumpCode -ne 0) {
 docker compose cp "db:$remote" $out
 $cpCode = $LASTEXITCODE
 Remove-RemoteDump $remote
-if ($cpCode -ne 0) { throw "BACKUP FAILED — could not copy the dump out of the container (exit $cpCode)." }
+# Remove the partial file: `cp` streams, so a mid-stream failure leaves a truncated dump under
+# the final name — non-zero length, so the size check below would not catch it either.
+if ($cpCode -ne 0) {
+    Remove-Item $out -ErrorAction SilentlyContinue
+    throw "BACKUP FAILED — could not copy the dump out of the container (exit $cpCode)."
+}
 if (-not (Test-Path $out) -or (Get-Item $out).Length -eq 0) {
     if (Test-Path $out) { Remove-Item $out }
     throw "BACKUP FAILED — the dump came out empty; no backup was written."

scripts/restore.ps1

@@ -63,10 +63,13 @@ if ($restoreCode -ne 0) {
     throw "pg_restore reported errors while restoring $DumpFile (exit $restoreCode) — read the output above before treating this instance as restored."
 }
 Write-Output "Restored $DumpFile"
-$cf = if ($env:COMPOSE_FILE) { " -f $env:COMPOSE_FILE" } else { "" }
-Write-Output "api and worker are stopped. Start them again the way you normally start Siftlode:"
-Write-Output "  docker compose$cf up -d"
-Write-Output "Use the SAME compose file you started with — 'up -d' recreates from whatever file it reads,"
+# Deliberately not a runnable line when we could not resolve the file: both compose files can be
+# present (a source checkout), and the two declare DIFFERENT volume names — so a pasted bare
+# 'docker compose up -d' would recreate the database on an empty volume and strand the restore.
+$cf = if ($env:COMPOSE_FILE) { $env:COMPOSE_FILE } else { "<the compose file you started with>" }
+Write-Output "api and worker are stopped. Start them again with:"
+Write-Output "  docker compose -f $cf up -d"
+Write-Output "It must be the SAME file you started with: 'up -d' recreates from whatever file it reads,"
 Write-Output "and a different one points at different volumes. Rolling a release back? Pin the older"
 Write-Output "IMAGE_TAG in .env FIRST: every start re-runs the migrations, and 'start' would relaunch the"
 Write-Output "existing container on the image you are rolling back FROM."

scripts/restore.sh

@@ -42,9 +42,16 @@ fi
 # not swallowed and not dressed up as success.
 if docker compose exec -T db pg_restore -U "$USER_NAME" -d "$DB_NAME" --clean --if-exists < "$FILE"; then
   echo "Restored $FILE"
-  echo "api and worker are stopped. Start them again the way you normally start Siftlode:"
-  echo "  docker compose ${COMPOSE_FILE:+-f $COMPOSE_FILE }up -d"
-  echo "Use the SAME compose file you started with — 'up -d' recreates from whatever file it reads,"
+  echo "api and worker are stopped. Start them again with:"
+  if [ -n "${COMPOSE_FILE:-}" ]; then
+    echo "  docker compose -f $COMPOSE_FILE up -d"
+  else
+    # Deliberately NOT a runnable line. Both compose files can be present (a source checkout), and
+    # the two declare DIFFERENT volume names — so a pasted bare 'docker compose up -d' would recreate
+    # the database on an empty volume and strand what was just restored.
+    echo "  docker compose -f <the compose file you started with> up -d"
+  fi
+  echo "It must be the SAME file you started with: 'up -d' recreates from whatever file it reads,"
   echo "and a different one points at different volumes. Rolling a release back? Pin the older"
   echo "IMAGE_TAG in .env FIRST: every start re-runs the migrations, and 'start' would relaunch the"
   echo "existing container on the image you are rolling back FROM."