fix(downloads): a converted upload too big says so, and a suspended owner's waiting rebuild is not "preparing"

- A converted upload over the space left (not zero) fails with CONVERTED_TOO_BIG_MESSAGE, not
  "storage is full"; it classifies as quota_exceeded by the phrase it shares with the clip's.
- The public watch page says "preparing" only for a job a worker will take: a suspended owner's
  queued job waits for as long as the suspension lasts, so the link says the video is gone, as
  for a paused rebuild.
- The worker's claim docstring names the restart case: a running job put back in the queue waits.

backend/tests/test_download_errors.py

@@ -64,6 +64,7 @@ def test_our_own_worker_refusals_classify_as_themselves():
     # from silently orphaning the category — the UI would then fall back to "Download failed".
     assert classify(dlerrors.QUOTA_FULL_MESSAGE) == "quota_exceeded"
     assert classify(dlerrors.CLIP_TOO_BIG_MESSAGE) == "quota_exceeded"
+    assert classify(dlerrors.CONVERTED_TOO_BIG_MESSAGE) == "quota_exceeded"
 
 
 def test_the_redacted_network_text_classifies_back_to_network():

backend/tests/test_upload_conversion.py

@@ -15,6 +15,7 @@ import pytest
 
 from app import sysconfig
 from app.config import settings
+from app.downloads import errors as dlerrors
 from app.downloads import gc as gcmod
 from app.downloads import pipeline, quota, storage
 from app.downloads import probe as probemod
@@ -124,6 +125,8 @@ def test_a_conversion_that_outgrows_the_cap_is_refused_and_keeps_the_original(
     db.expire_all()
     job = db.get(DownloadJob, job.id)
     assert (job.status, job.error_code) == ("error", "quota_exceeded")
+    # 90 of the 100 are still free: the converted file is too big, the storage is not full.
+    assert job.error == dlerrors.CONVERTED_TOO_BIG_MESSAGE
     assert original.exists()
     assert [p.name for p in _uploads_dir(_root, u.id).iterdir()] == [original.name]
 
@@ -558,6 +561,33 @@ def test_a_link_says_preparing_only_while_its_video_is_being_made(
     assert r.json()["detail"]["reason"] == reason
 
 
+def test_a_link_to_a_suspended_owners_waiting_rebuild_says_it_is_gone(
+    db, as_user, make_user, _root, monkeypatch, upload_headers
+):
+    # The worker leaves a suspended user's queued jobs where they are, maybe for good: like a paused
+    # rebuild, nothing is making the video, so "try again shortly" would never come true.
+    _probe(monkeypatch, vcodec="h264")
+    owner = make_user()
+    c = as_user(owner)
+    job = c.post(
+        "/api/downloads/upload", content=b"h264 bytes", headers=upload_headers("clip.mp4")
+    ).json()
+    link = c.post(f"/api/downloads/{job['id']}/links", json={"allow_download": False}).json()
+    db.get(MediaAsset, job["asset"]["id"]).status = "pending"
+    db.get(DownloadJob, job["id"]).status = "queued"
+    owner.is_suspended = True
+    db.commit()
+    r = c.get(f"/api/public/watch/{link['token']}")
+    assert r.status_code == 404
+    assert r.json()["detail"]["reason"] == "video_gone"
+    # A rebuild already running when the owner was suspended still finishes.
+    db.get(DownloadJob, job["id"]).status = "running"
+    db.commit()
+    assert c.get(f"/api/public/watch/{link['token']}").json()["detail"]["reason"] == (
+        "video_preparing"
+    )
+
+
 def test_the_admin_storage_view_counts_a_waiting_upload(db, as_user, make_user, _root, monkeypatch):
     # The same disk the owner's footprint and the GC's cap count (`quota.ON_DISK`).
     _probe(monkeypatch)