feat: a tick keeps looking back 30 days, so an EDITED post is reached (4386)

Operator, 2026-09-23, on a Floppystack post: "this post has been updated as
he implements hot fixes — any chance we have a way to scan for or see updated
posts so we can update ours to match and pull the new attachments and
pictures etc."

The download half already worked: extract_media reads the media list off the
LIVE feed response every walk, so a newly attached hotfix build is a ledger
key we have never seen. Only REACHING the post was missing — a tick stopped
after 20 contiguous already-have-it items, and a post edited three days after
publication sits well below twenty. Not a bug in the early-out; a count
cannot express "recent".

The early-out now needs BOTH conditions: the run of seen items AND a post
published before the horizon. Strictly a widening — window 0 is exactly the
old behaviour, and no window can make a tick stop EARLIER than it used to, so
a source paused for months still walks its whole unseen backlog. The horizon
is a floor on how far to look, never a ceiling.

Inside the window the post-record gate is bypassed too (write_post_record
revisit=True): the body is re-read from the feed response already in hand, so
a revisit costs zero requests, and a body that comes back empty writes
NOTHING rather than blanking one a detail-fetch had filled. Revisits are kept
out of the #862 body-drift canary's sample for the same reason — an empty
revisit is healthy, and counting it would walk the alarm toward firing on
good ticks.

The run summary names what changed ("3 post(s) updated (5 new file(s))") with
a line per post; the ask was to SEE updated posts, not only to end up with
their bytes.

download_revisit_days is a settings row, not a constant (rule 25) — how long
a creator keeps editing is a property of the creator. Default 30, 0 turns it
off. Migration 0108.

Also corrects two stale docstrings: both clients described post_meta as
feeding an Ingester.preview that no longer calls it. It had no consumer at
all until this change.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LVjrnpQjRgHdvq95rASoiR

alembic/versions/0108_download_revisit_days.py

@@ -0,0 +1,54 @@
+"""download_revisit_days — how far back a tick keeps looking for EDITED posts.
+
+Operator, 2026-09-23, pointing at a Floppystack post: *"this post has been
+updated as he implements hot fixes — any chance we have a way to scan for or
+see updated posts so we can update ours to match and pull the new attachments
+and pictures etc."*
+
+A tick stopped after 20 contiguous already-have-it items. That is the right
+instinct and the wrong unit: a post edited three days after publication sits
+well below twenty seen items, so the walk turned around before reaching it. The
+walk now needs BOTH a run of seen items and a post older than this many days
+before it stops.
+
+A settings row rather than a constant (rule 25) because the right window is a
+property of the CREATOR, not of FabledCurator — one artist appends hotfix
+builds for a fortnight, another never touches a post again. 0 turns the revisit
+off entirely and restores the pure count early-out.
+
+30 days is the operator's own number, 2026-09-23.
+
+Revision ID: 0108
+Revises: 0107
+Create Date: 2026-09-23
+
+"""
+from typing import Sequence, Union
+
+import sqlalchemy as sa
+from alembic import op
+
+revision: str = "0108"
+down_revision: Union[str, None] = "0107"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    # server_default so the existing single settings row gets the window without
+    # a data migration — and so an install that predates this column reads 30
+    # rather than 0. 0 is a real, meaningful value here (revisit off), so the
+    # column must never be allowed to arrive at it by omission.
+    op.add_column(
+        "import_settings",
+        sa.Column(
+            "download_revisit_days",
+            sa.Integer(),
+            nullable=False,
+            server_default="30",
+        ),
+    )
+
+
+def downgrade() -> None:
+    op.drop_column("import_settings", "download_revisit_days")

backend/app/api/settings.py

@@ -36,6 +36,7 @@ _EDITABLE_FIELDS = (
     "download_validate_files",
     "download_schedule_default_seconds",
     "download_event_retention_days",
+    "download_revisit_days",
     "download_failure_warning_threshold",
     "series_suggest_enabled",
     "series_suggest_threshold",
@@ -113,6 +114,12 @@ async def update_import_settings():
         v = body["download_schedule_default_seconds"]
         if not isinstance(v, int) or isinstance(v, bool) or v < 60 or v > 86400:
             return _bad_int("download_schedule_default_seconds", 60, 86400)
+    # 0 is a real value (revisit off), so the floor is 0, not 1 — and the
+    # ceiling is a year, past which a "tick" is a backfill wearing a hat.
+    if "download_revisit_days" in body:
+        v = body["download_revisit_days"]
+        if not isinstance(v, int) or isinstance(v, bool) or v < 0 or v > 365:
+            return _bad_int("download_revisit_days", 0, 365)
     if "download_event_retention_days" in body:
         v = body["download_event_retention_days"]
         if not isinstance(v, int) or isinstance(v, bool) or v < 1 or v > 3650:

backend/app/models/import_settings.py

@@ -68,6 +68,16 @@ class ImportSettings(Base):
         Integer, nullable=False, default=90,
         server_default="90",
     )
+    # How far back a routine tick keeps looking after it has run out of new
+    # posts, so a creator who EDITS an older post to attach a hotfix build is
+    # still reached (ingest_core.DEFAULT_REVISIT_DAYS carries the reasoning).
+    # A knob rather than a constant because how long a creator keeps editing is
+    # a property of the creator, not of FabledCurator: 0 turns the revisit off
+    # and restores the pure count early-out.
+    download_revisit_days: Mapped[int] = mapped_column(
+        Integer, nullable=False, default=30,
+        server_default="30",
+    )
     download_failure_warning_threshold: Mapped[int] = mapped_column(
         Integer, nullable=False, default=5,
         server_default="5",

backend/app/services/download_backends.py

@@ -24,6 +24,7 @@ import asyncio
 from pathlib import Path
 
 from .gallery_dl import DownloadResult, ErrorType
+from .ingest_core import DEFAULT_REVISIT_DAYS
 from .patreon_ingester import PatreonIngester
 from .patreon_resolver import extract_vanity, resolve_campaign_id_for_source
 from .platforms import known_platform_keys
@@ -83,6 +84,7 @@ async def run_download(
     mode: str | None,
     gdl,
     sync_session_factory,
+    revisit_days: int = DEFAULT_REVISIT_DAYS,
 ) -> tuple[DownloadResult, str | None]:
     """Uniform download across backends — the download counterpart to
     `verify_source_credential`, so this module is the ONE place that knows how
@@ -106,7 +108,7 @@ async def run_download(
         ), None
     if uses_native_ingester(platform):
         return await _run_native_ingester(
-            ctx, source_config, mode, gdl, sync_session_factory
+            ctx, source_config, mode, gdl, sync_session_factory, revisit_days
         )
     result = await gdl.download(
         url=ctx["url"],
@@ -146,6 +148,7 @@ def _campaign_resolution_error(platform: str, url: str) -> str:
 
 async def _run_native_ingester(
     ctx: dict, source_config, mode: str | None, gdl, sync_session_factory,
+    revisit_days: int = DEFAULT_REVISIT_DAYS,
 ) -> tuple[DownloadResult, str | None]:
     """Run the native ingester for a native platform in a worker thread (sync
     requests/subprocess). Patreon resolves a campaign id from the vanity URL;
@@ -210,6 +213,9 @@ async def _run_native_ingester(
             mode=mode,
             resume_cursor=source_config.resume_cursor,
             time_budget_seconds=source_config.timeout,
+            # How far back a tick keeps looking for EDITED posts. The ingester
+            # applies it to ticks only; a backfill ignores it.
+            revisit_days=revisit_days,
             posts_base=int(overrides.get("_backfill_posts", 0)),
             # plan #709: live progress writes to this running event mid-walk.
             event_id=ctx.get("event_id"),

backend/app/services/download_service.py

@@ -39,6 +39,7 @@ from .gallery_dl import (
     walk_completed,
 )
 from .importer import Importer
+from .ingest_core import DEFAULT_REVISIT_DAYS
 from .platforms import auth_type_for
 from .scheduler_service import set_platform_cooldown
 
@@ -60,6 +61,7 @@ class DownloadService:
         importer: Importer,
         cred_service: CredentialService,
         sync_session_factory=None,
+        revisit_days: int = DEFAULT_REVISIT_DAYS,
     ):
         self.async_session = async_session
         self.sync_session = sync_session
@@ -71,6 +73,12 @@ class DownloadService:
         # the multi-minute walk — see PatreonIngester). Only the patreon branch
         # of phase 2 uses it; gallery-dl sources leave it None.
         self.sync_session_factory = sync_session_factory
+        # ImportSettings.download_revisit_days — how far back a tick keeps
+        # looking for EDITED posts (ingest_core.DEFAULT_REVISIT_DAYS). Passed in
+        # rather than read here because the task already loads the settings row
+        # for rate_limit/validate_files, and a second load on every download
+        # would be the same row twice for one number.
+        self.revisit_days = revisit_days
 
     async def download_source(self, source_id: int) -> int:
         """Returns DownloadEvent.id. Idempotent: in-flight events are returned as-is."""
@@ -178,6 +186,7 @@ class DownloadService:
         return await run_download(
             ctx=ctx, source_config=source_config, skip_value=skip_value, mode=mode,
             gdl=self.gdl, sync_session_factory=self.sync_session_factory,
+            revisit_days=self.revisit_days,
         )
 
     async def _phase1_setup(self, source_id: int) -> dict[str, Any]:

backend/app/services/ingest_core.py

@@ -31,6 +31,7 @@ import json
 import logging
 import time
 from collections.abc import Callable
+from datetime import UTC, datetime, timedelta
 
 from sqlalchemy import delete, func, select, text
 from sqlalchemy.dialects.postgresql import insert as pg_insert
@@ -51,6 +52,33 @@ log = logging.getLogger(__name__)
 # per-file HEADs. Headroom against paywalled/undownloadable items interleaving.
 _TICK_SEEN_THRESHOLD = 20
 
+# How far back a tick keeps looking even once everything is already-have-it —
+# the REVISIT WINDOW. Operator, 2026-09-23, holding up a Floppystack post:
+# *"this post has been updated as he implements hot fixes — any chance we have a
+# way to scan for or see updated posts so we can update ours to match and pull
+# the new attachments and pictures etc."*
+#
+# A creator who edits a three-day-old post to append a hotfix build was
+# structurally unreachable: that post sits twenty-odd already-seen items down
+# the feed, so the count early-out above fired before the walk ever got to it.
+# Not a bug in the early-out — a COUNT cannot express "recent".
+#
+# So the early-out now needs BOTH conditions: the run of already-seen items AND
+# a post published before the horizon. Strictly a widening. Two properties this
+# shape has and a plain "walk the last N days" would not:
+#
+#   * window 0 is exactly the old behaviour, so the feature has an off switch
+#     that costs nothing to reason about;
+#   * no window can make a tick stop EARLIER than it used to. A source paused
+#     for months has an unseen backlog stretching well past any horizon, and
+#     the walk still runs to the end of it — the horizon is a FLOOR on how far
+#     to look, never a ceiling.
+#
+# The live value is `ImportSettings.download_revisit_days` (rule 25 — an
+# operator tuning how far back their creators edit should not need a redeploy).
+# This is the fallback for a caller that passes none.
+DEFAULT_REVISIT_DAYS = 30
+
 # plan #705 #7: after this many failed download/validate attempts a media is
 # "dead-lettered" and skipped on routine tick/backfill walks (recovery still
 # re-attempts it). Stops a permanently-broken media re-erroring forever.
@@ -78,6 +106,28 @@ _LIVE_PROGRESS_INTERVAL = 5.0
 _CANARY_MIN_SAMPLE = 30
 
 
+def _parse_published(raw: object) -> datetime | None:
+    """An ISO-8601 post date from either native client, as aware UTC.
+
+    Patreon's `published_at` is tz-aware with a `Z` or `+00:00` offset;
+    SubscribeStar's is NAIVE (`_parse_ss_datetime` renders a parsed local
+    timestamp with no zone). A naive value is read as UTC — the alternative is
+    discarding it, and a post whose date we refuse to read is a post the revisit
+    window can never reach.
+
+    Anything unparseable returns None, which reads downstream as "not provably
+    recent" and leaves the walk on its pre-revisit behaviour. Never raises: a
+    date we cannot read must not fail a walk that is otherwise working.
+    """
+    if not isinstance(raw, str) or not raw.strip():
+        return None
+    try:
+        parsed = datetime.fromisoformat(raw.strip().replace("Z", "+00:00"))
+    except ValueError:
+        return None
+    return parsed if parsed.tzinfo else parsed.replace(tzinfo=UTC)
+
+
 class Ingester:
     """Generic native-ingest orchestration. Subclass with a platform adapter
     (see the module docstring) — or construct directly with the keyword seams."""
@@ -132,11 +182,17 @@ class Ingester:
         resume_cursor: str | None = None,
         time_budget_seconds: float = 870.0,
         seen_threshold: int = _TICK_SEEN_THRESHOLD,
+        revisit_days: int = DEFAULT_REVISIT_DAYS,
         posts_base: int = 0,
         event_id: int | None = None,
     ) -> DownloadResult:
         """Walk + download for one source, returning a gallery-dl-shaped result.
 
+        `revisit_days` is the tick's revisit window (see DEFAULT_REVISIT_DAYS):
+        inside it a tick neither early-outs nor trusts the post-record gate, so
+        a post edited after we first captured it is re-read and its new
+        attachments downloaded. 0 turns the window off.
+
         `mode` is "tick" | "backfill" | "recovery" | "recapture". Recovery
         bypasses the tier-1 seen-ledger AND the dead-letter ledger (tier-2 disk
         still skips kept files). Recapture (#830) is the cheap "re-grab post
@@ -179,6 +235,21 @@ class Ingester:
         # no media download, no post-record stub. Absent on stub/not-yet-migrated
         # clients → nothing is ever treated as gated.
         post_is_gated = getattr(self.client, "post_is_gated", None)
+        # The revisit window (see DEFAULT_REVISIT_DAYS). `post_meta` is an
+        # existing client seam — both native clients already implement it, for
+        # a preview sample whose caller has since gone, so this needed no new
+        # contract, only a live consumer for one. Absent seam, an unreadable
+        # date or a window of 0 → `horizon` never matches and the walk behaves
+        # exactly as it did before 2026-09-23.
+        #
+        # The window applies to TICKS only. A backfill is gated on purpose
+        # (capture each post once) and `recapture` mode already exists for the
+        # operator-driven "re-read every body" pass; a horizon there would be a
+        # third overlapping answer to a question that has two.
+        post_meta = getattr(self.client, "post_meta", None)
+        horizon: datetime | None = None
+        if mode == "tick" and revisit_days > 0 and post_meta is not None:
+            horizon = datetime.now(UTC) - timedelta(days=revisit_days)
         start = time.monotonic()
         last_live = start  # plan #709: last live-progress write timestamp
         log_lines: list[str] = []
@@ -210,6 +281,12 @@ class Ingester:
         # absolute across chunks instead of an inflating sum. posts_processed
         # stays the gross per-chunk count used for the run summary.
         chunk_new_posts = 0
+        # Posts inside the revisit window that we had already captured, and the
+        # media those revisits turned up. Reported in the run summary — the
+        # operator's ask was to SEE the updated posts, not only to end up with
+        # their files ("so we can update ours to match").
+        revisited = 0
+        revisit_downloads = 0
         consecutive_seen = 0
         emitted_cursor: str | None = None
         reached_bottom = False
@@ -322,6 +399,20 @@ class Ingester:
                 # resume_cursor None, so everything counts.
                 if not (resume_cursor and page_cursor == resume_cursor):
                     chunk_new_posts += 1
+                # Inside the revisit window? Computed per post rather than
+                # "stop once one post is old" because the feed is only MOSTLY
+                # date-ordered — a pinned or re-pinned post can sit above older
+                # ones, and one such post must not end the walk.
+                in_window = False
+                if horizon is not None:
+                    published = _parse_published((post_meta(post) or {}).get("date"))
+                    in_window = published is not None and published >= horizon
+                # Set by the post-record block below when this post was already
+                # captured on an earlier walk. Stays False when the platform has
+                # no post-record seam, so the revisit accounting simply reports
+                # nothing rather than guessing.
+                post_already_recorded = False
+                downloaded_before = downloaded
                 # Tier-gated post (#874): the account can't fully view it, so
                 # Patreon serves only blurred locked-preview media. Skip it
                 # ENTIRELY — no media download AND no post-record stub (operator
@@ -353,11 +444,31 @@ class Ingester:
                             set() if recapture_records
                             else self._seen_keys(source_id, [pkey])
                         )
-                        if pkey not in already:
+                        post_already_recorded = pkey in already
-                            rec = write_post_record(post, artist_slug)
+                        # A post inside the revisit window is re-read even
-                            posts_recorded += 1
+                        # though the gate has it: that gate's whole job is to
-                            if rec.body_chars:
+                        # stop us paying for a post twice, and an EDITED post is
-                                posts_with_body += 1
+                        # not the same post. `revisit=True` keeps the cost at
+                        # zero requests — the downloader re-reads the body from
+                        # the feed response already in hand and declines to
+                        # write at all if that body came back empty, so a
+                        # detail-fetched body is never overwritten by a blank.
+                        if not post_already_recorded or in_window:
+                            rec = write_post_record(
+                                post, artist_slug, revisit=post_already_recorded,
+                            )
+                            if not post_already_recorded:
+                                # FIRST captures only feed the #862 body canary.
+                                # A revisit legitimately comes back empty — a
+                                # post whose body only ever arrived from the
+                                # detail endpoint has none in the feed, and the
+                                # downloader declines to write it. Counting
+                                # those into the sample would walk the canary
+                                # toward firing on healthy ticks, which is the
+                                # one thing a drift alarm must never do.
+                                posts_recorded += 1
+                                if rec.body_chars:
+                                    posts_with_body += 1
                             if rec.path is not None:
                                 post_records.append(str(rec.path))
                                 self._mark_seen(source_id, [(pkey, ppid)])
@@ -367,7 +478,8 @@ class Ingester:
                             # a 0-char body is the "why is this one empty" answer.
                             log_lines.append(
                                 f"  post {ppid} [{rec.post_type or '?'}] "
-                                f"body: {rec.body_chars} chars"
+                                + ("re-read, " if post_already_recorded else "")
+                                + f"body: {rec.body_chars} chars"
                                 + ("" if rec.body_chars else " — EMPTY")
                                 + (f" — {rec.title}" if rec.title else "")
                             )
@@ -451,7 +563,15 @@ class Ingester:
                         to_fail.append((key, media_item.post_id, outcome.error or "error"))
                         # An error neither advances nor resets the run-of-seen.
 
-                    if mode == "tick" and consecutive_seen >= seen_threshold:
+                    # `not in_window` is the revisit window's half of the
+                    # early-out: a run of already-seen items is only permission
+                    # to stop once the walk is BELOW the horizon. Both halves,
+                    # never either alone — see DEFAULT_REVISIT_DAYS.
+                    if (
+                        mode == "tick"
+                        and not in_window
+                        and consecutive_seen >= seen_threshold
+                    ):
                         early_out = True
                         break
 
@@ -465,6 +585,21 @@ class Ingester:
                 if to_fail:
                     self._record_failures(source_id, to_fail)
 
+                # An already-captured post that yielded NEW media is an edited
+                # post — the operator's Floppystack case, and the one thing in
+                # this walk worth naming individually in the run log. The media
+                # half needed no new detection: `extract_media` reads the media
+                # list off the live feed response, so a hotfix build appended
+                # last night is simply a ledger key we have never seen.
+                new_here = downloaded - downloaded_before
+                if post_already_recorded and new_here:
+                    revisited += 1
+                    revisit_downloads += new_here
+                    log_lines.append(
+                        f"  post {post.get('id')} — updated: "
+                        f"{new_here} new file(s)"
+                    )
+
                 # plan #709: time-throttled live progress to the running event so
                 # the Downloads view ticks ~every 5s, independent of page size.
                 now = time.monotonic()
@@ -527,6 +662,12 @@ class Ingester:
             # visible in the Raw stdout (e.g. "bodies 3/180" reads as off).
             + (f", bodies {posts_with_body}/{posts_recorded}" if posts_recorded else "")
             + (f", {gated_skipped} gated-skipped" if gated_skipped else "")
+            # Only when it happened: on a quiet tick this is 0 and saying so
+            # every run would bury the times it is not.
+            + (
+                f", {revisited} post(s) updated ({revisit_downloads} new file(s))"
+                if revisited else ""
+            )
             + (", reached end" if reached_bottom else "")
             + (", time-boxed" if budget_hit else "")
         )

backend/app/services/patreon_client.py

@@ -483,8 +483,14 @@ class PatreonClient:
 
     @staticmethod
     def post_meta(post: dict) -> dict:
-        """Title + published date for a post — for the preview sample (plan #708
+        """Title + published date for a post. Part of the client contract.
-        B4). Part of the client contract `ingest_core.Ingester.preview` calls."""
+
+        Written for a preview sample (plan #708 B4) whose caller has since gone;
+        as of 2026-09-23 its consumer is the core's REVISIT WINDOW, which needs
+        a post's date to know whether a tick is still inside it. Both native
+        clients answer in the same shape — an ISO-8601 string under `date`, or
+        None — so the core reads a date without knowing the platform.
+        """
         attrs = post.get("attributes") or {}
         title = attrs.get("title")
         published = attrs.get("published_at")

backend/app/services/patreon_downloader.py

@@ -340,7 +340,7 @@ class PatreonDownloader(BaseNativeDownloader):
 
     def _write_sidecar_data(
         self, post: dict, sidecar_path: Path, *, source_url: str | None = None,
-        minimal: bool = False,
+        minimal: bool = False, detail_fetch: bool = True,
     ) -> Path:
         """Serialize the post's metadata to `sidecar_path`. The post-only record
         (`write_post_record`) writes the FULL post (body/title/date/url); the
@@ -364,7 +364,13 @@ class PatreonDownloader(BaseNativeDownloader):
         # dict — so a multi-image post fetches detail at most once, the post-record
         # body-length read reuses it, and a fully-seen post (no fresh download → no
         # sidecar write) never pays the extra GET.
-        if (not content or not content.strip()) and self._content_fetcher:
+        # `detail_fetch=False` on a REVISIT (a post inside the tick's revisit
+        # window that we already captured): re-read the body from the feed
+        # response we are holding and pay nothing. Without this a 30-day window
+        # would buy one detail GET per body-less post per tick, forever — a
+        # per-creator cost that grows with how prolific they are, to re-fetch a
+        # body we already stored.
+        if (not content or not content.strip()) and self._content_fetcher and detail_fetch:
             fetched = self._content_fetcher(str(post.get("id") or ""))
             if fetched:
                 content = fetched
@@ -386,7 +392,9 @@ class PatreonDownloader(BaseNativeDownloader):
         sidecar_path.write_text(json.dumps(data, indent=2))
         return sidecar_path
 
-    def write_post_record(self, post: dict, artist_slug: str) -> PostRecordOutcome:
+    def write_post_record(
+        self, post: dict, artist_slug: str, *, revisit: bool = False,
+    ) -> PostRecordOutcome:
         """Write a post-ONLY sidecar (no media file) for a media-less post, so
         the importer can still upsert the Post + its body — text posts often hold
         the only copy of an external <a href> link. Named `_post.json`: the
@@ -397,6 +405,18 @@ class PatreonDownloader(BaseNativeDownloader):
         Returns a PostRecordOutcome (path None when the post has no id) carrying
         the captured body's shape — post_type + final char count — so the engine
         can log per-post handling without re-reading the post itself.
+
+        `revisit=True` is the tick re-reading a post it already captured
+        (ingest_core's revisit window, #...). Two differences, both about not
+        making an update cost more than it is worth:
+
+        * no detail-fetch — the body comes from the feed response already in
+          hand, so a revisit costs zero requests;
+        * a body that comes back empty writes NOTHING and returns `path=None`.
+          On a first capture an empty body is the truth about the post; on a
+          revisit it usually just means this post's body only ever came from
+          the detail endpoint we just declined to call, and writing it would
+          blank a stored body to say something we never learned.
         """
         attrs = post.get("attributes") or {}
         title = attrs.get("title") if isinstance(attrs.get("title"), str) else None
@@ -406,9 +426,17 @@ class PatreonDownloader(BaseNativeDownloader):
             return PostRecordOutcome(
                 path=None, post_type=post_type, title=title, body_chars=0,
             )
+        if revisit:
+            feed_body = post_body_html(attrs)
+            if not (isinstance(feed_body, str) and feed_body.strip()):
+                return PostRecordOutcome(
+                    path=None, post_type=post_type, title=title, body_chars=0,
+                )
         post_dir = self.images_root / artist_slug / "patreon" / post_dir_name(post)
         post_dir.mkdir(parents=True, exist_ok=True)
-        path = self._write_sidecar_data(post, post_dir / "_post.json")
+        path = self._write_sidecar_data(
+            post, post_dir / "_post.json", detail_fetch=not revisit,
+        )
         # _write_sidecar_data has by now memoized any detail-fetched body onto
         # post["attributes"]["content"], so re-read it for the FINAL char count.
         body = (post.get("attributes") or {}).get("content")

backend/app/services/subscribestar_client.py

@@ -742,8 +742,15 @@ class SubscribeStarClient:
 
     @staticmethod
     def post_meta(post: dict) -> dict:
-        """Title + date for the preview sample. Title is synthesized from the body
+        """Title + date. Title is None — SubscribeStar has no title field, and
-        (SubscribeStar has no title field)."""
+        the importer synthesizes one from the body.
+
+        `date` is the contract the core's REVISIT WINDOW reads (2026-09-23):
+        ISO-8601 or None. NAIVE here, because `_parse_ss_datetime` renders a
+        parsed local timestamp with no zone; the core reads a naive date as UTC
+        rather than discarding it, since a date it refuses to read is a post
+        the window can never reach.
+        """
         attrs = post.get("attributes") or {}
         return {"title": None, "date": attrs.get("published_at")}
 

backend/app/services/subscribestar_downloader.py

@@ -184,10 +184,20 @@ class SubscribeStarDownloader(BaseNativeDownloader):
         sidecar_path.write_text(json.dumps(data, indent=2))
         return sidecar_path
 
-    def write_post_record(self, post: dict, artist_slug: str) -> PostRecordOutcome:
+    def write_post_record(
+        self, post: dict, artist_slug: str, *, revisit: bool = False,
+    ) -> PostRecordOutcome:
         """Write the post-first `_post.json` (body/links/metadata) — the sole
         writer of the post record on the native path. SubscribeStar's body is
-        already in the feed HTML, so no detail-fetch is needed."""
+        already in the feed HTML, so no detail-fetch is needed.
+
+        `revisit=True` is the tick re-reading a post it already captured
+        (ingest_core's revisit window). The no-detail-fetch half of that
+        contract is free here — there is no detail endpoint — but the
+        don't-blank-a-stored-body half still applies: a chunk that parsed with
+        no content must not overwrite a body we already have. Same guarantee as
+        the Patreon downloader, for the same reason, so a walk behaves the same
+        on both platforms."""
         attrs = post.get("attributes") or {}
         title = attrs.get("title") if isinstance(attrs.get("title"), str) else None
         post_type = attrs.get("post_type") if isinstance(attrs.get("post_type"), str) else None
@@ -196,6 +206,12 @@ class SubscribeStarDownloader(BaseNativeDownloader):
             return PostRecordOutcome(
                 path=None, post_type=post_type, title=title, body_chars=0,
             )
+        if revisit:
+            body = attrs.get("content")
+            if not (isinstance(body, str) and body.strip()):
+                return PostRecordOutcome(
+                    path=None, post_type=post_type, title=title, body_chars=0,
+                )
         post_dir = self.images_root / artist_slug / "subscribestar" / post_dir_name(post)
         post_dir.mkdir(parents=True, exist_ok=True)
         path = self._write_sidecar_data(post, post_dir / "_post.json")

backend/app/tasks/download.py

@@ -177,6 +177,10 @@ def download_source(self, source_id: int, _serialize_waits: int = 0) -> int:
                 settings = ImportSettings.load_sync(sync_session)
                 rate_limit = settings.download_rate_limit_seconds
                 validate_files = settings.download_validate_files
+                # How far back a tick keeps looking for EDITED posts. Read here
+                # with the other downloader knobs, off the row this block is
+                # already holding open.
+                revisit_days = settings.download_revisit_days
 
             gdl = GalleryDLService(
                 images_root=IMAGES_ROOT,
@@ -207,6 +211,7 @@ def download_source(self, source_id: int, _serialize_waits: int = 0) -> int:
                         # the walk). Same factory the importer's sync session
                         # comes from — a different DB connection per checkout.
                         sync_session_factory=SyncFactory,
+                        revisit_days=revisit_days,
                     )
                     return await svc.download_source(source_id)
         finally:

frontend/src/components/subscriptions/SettingsTab.vue

@@ -153,6 +153,24 @@
             </div>
           </v-col>
         </v-row>
+        <v-row>
+          <v-col cols="12" sm="4">
+            <v-text-field
+              v-model.number="dl.download_revisit_days"
+              label="Re-check edited posts (days)"
+              type="number" :min="0" :max="365"
+              density="compact" hide-details
+              @blur="saveDownloader"
+            />
+            <div class="fc-help">
+              A routine check keeps looking back this far after it runs out of
+              new posts, so a creator who edits an older post to add a file is
+              still picked up. New attachments download; the post text is
+              re-read from the same page, costing nothing extra. 0 turns this
+              off. Default 30.
+            </div>
+          </v-col>
+        </v-row>
         <v-alert v-if="importStore.settingsError" type="error" variant="tonal" class="mt-2" closable>
           {{ importStore.settingsError }}
         </v-alert>
@@ -209,6 +227,7 @@ const dl = reactive({
   download_schedule_default_seconds: 28800,
   download_event_retention_days: 90,
   download_failure_warning_threshold: 5,
+  download_revisit_days: 30,
   extdl_mega_enabled: true,
   extdl_gdrive_enabled: true,
   extdl_mediafire_enabled: true,

tests/test_api_settings_downloader.py

@@ -65,3 +65,41 @@ async def test_extdl_toggle_rejects_non_bool(client):
         "/api/settings/import", json={"extdl_gdrive_enabled": "nope"}
     )
     assert resp.status_code == 400
+
+
+# -- download_revisit_days: how far back a tick looks for EDITED posts --------
+
+
+@pytest.mark.asyncio
+async def test_revisit_window_defaults_to_thirty_days(client):
+    body = await (await client.get("/api/settings/import")).get_json()
+    assert body["download_revisit_days"] == 30
+
+
+@pytest.mark.asyncio
+async def test_revisit_window_is_settable(client):
+    resp = await client.patch(
+        "/api/settings/import", json={"download_revisit_days": 7}
+    )
+    assert resp.status_code == 200
+    assert (await resp.get_json())["download_revisit_days"] == 7
+
+
+@pytest.mark.asyncio
+async def test_zero_is_accepted_because_it_is_the_off_switch(client):
+    """0 turns the revisit off and restores the pure count early-out. Pinned
+    because the obvious bounds check for a "days" field is `>= 1`, and that
+    would take the off switch away without anything failing."""
+    resp = await client.patch(
+        "/api/settings/import", json={"download_revisit_days": 0}
+    )
+    assert resp.status_code == 200
+    assert (await resp.get_json())["download_revisit_days"] == 0
+
+
+@pytest.mark.asyncio
+async def test_a_negative_window_is_refused(client):
+    resp = await client.patch(
+        "/api/settings/import", json={"download_revisit_days": -1}
+    )
+    assert resp.status_code == 400

tests/test_patreon_downloader.py

@@ -658,6 +658,53 @@ def test_write_post_record_writes_enriched_post_only_sidecar(tmp_path):
     assert rec.body_chars == len("<p>text post body</p>")
 
 
+# The revisit contract (2026-09-23): a tick re-reading a post it already
+# captured, so an edit made after first capture reaches us. Both halves exist to
+# stop an update costing more than it is worth.
+
+
+def test_a_revisit_re_reads_the_body_without_paying_for_a_detail_fetch(tmp_path):
+    """A 30-day window would otherwise buy one detail GET per body-less post per
+    tick, forever — a per-creator cost that grows with how prolific they are, to
+    re-fetch a body we already stored."""
+    calls: list[str] = []
+
+    def _fetcher(post_id: str) -> str:
+        calls.append(post_id)
+        return "<p>detail body</p>"
+
+    dl = PatreonDownloader(
+        images_root=tmp_path, cookies_path=None, validate=False,
+        session=_FakeSession(), content_fetcher=_fetcher,
+    )
+    post = _post()
+    post["attributes"]["content"] = "<p>feed body</p>"
+    rec = dl.write_post_record(post, "artist-x", revisit=True)
+
+    assert calls == []
+    assert json.loads(rec.path.read_text())["content"] == "<p>feed body</p>"
+
+
+def test_a_revisit_with_an_empty_body_writes_nothing_at_all(tmp_path):
+    """The data-loss guard. On a FIRST capture an empty body is the truth about
+    the post; on a revisit it usually means the body only ever came from the
+    detail endpoint we just declined to call. Writing it would blank a stored
+    body to say something we never learned."""
+    dl = PatreonDownloader(
+        images_root=tmp_path, cookies_path=None, validate=False,
+        session=_FakeSession(), content_fetcher=lambda _pid: "<p>detail</p>",
+    )
+    post = _post()
+    post["attributes"]["content"] = ""
+    rec = dl.write_post_record(post, "artist-x", revisit=True)
+
+    assert rec.path is None
+    assert rec.body_chars == 0
+    # Not "wrote an empty file" — nothing was written, so a record captured on
+    # an earlier walk is still exactly what it was.
+    assert not list(tmp_path.rglob("_post.json"))
+
+
 def test_write_post_record_none_without_post_id(tmp_path):
     dl = PatreonDownloader(
         images_root=tmp_path, cookies_path=None, validate=False,

tests/test_patreon_ingester.py

@@ -7,6 +7,8 @@ real CDN. The ledger is real (a sync sessionmaker bound to the test engine), so
 the tier-1 skip and the idempotent mark-seen run against actual rows.
 """
 
+from datetime import UTC, datetime, timedelta
+
 import pytest
 from sqlalchemy import func, select
 from sqlalchemy.orm import sessionmaker
@@ -50,9 +52,15 @@ class _FakeClient:
     """Stub PatreonClient. `pages` is a list of (page_cursor, [posts]); each post
     is (post_id, [MediaItem]). `raise_on_first` lets a test trip drift."""
 
-    def __init__(self, pages, raise_exc=None, empty_body=False, gated=None):
+    def __init__(self, pages, raise_exc=None, empty_body=False, gated=None,
+                 published=None):
         self._pages = pages
         self._raise_exc = raise_exc
+        # {post_id: ISO-8601 published_at} for the revisit window. Absent → the
+        # date reads as None, which is how EVERY test written before the window
+        # existed keeps its old behaviour: an unreadable date is never inside
+        # the horizon, so the count early-out stands alone.
+        self._published = dict(published or {})
         # empty_body simulates a body-field schema break: every post comes back
         # with no content (the #862 canary's trip condition).
         self._empty_body = empty_body
@@ -89,7 +97,10 @@ class _FakeClient:
         return post["_media"]
 
     def post_meta(self, post):
-        return {"title": post.get("id"), "date": None}
+        return {
+            "title": post.get("id"),
+            "date": self._published.get(str(post.get("id") or "")),
+        }
 
     @staticmethod
     def post_is_gated(post):
@@ -113,6 +124,7 @@ class _FakeDownloader:
         self.error = set(error or ())
         self.download_calls = 0
         self.post_records = 0
+        self.post_revisits = 0
 
     def download_post(self, post, media_items, artist_slug, *, is_seen,
                       should_stop=lambda: False, recapture=False):
@@ -146,8 +158,19 @@ class _FakeDownloader:
                 outcomes.append(MediaOutcome(media=m, status="downloaded", path=p, error=None))
         return outcomes
 
-    def write_post_record(self, post, artist_slug):
+    def write_post_record(self, post, artist_slug, *, revisit=False):
         self.post_records += 1
+        if revisit:
+            self.post_revisits += 1
+        attrs_ = post.get("attributes") or {}
+        body_ = attrs_.get("content")
+        # Mirrors the real downloaders' revisit contract: a re-read whose body
+        # came back empty writes NOTHING rather than blanking a stored one.
+        if revisit and not (isinstance(body_, str) and body_.strip()):
+            return PostRecordOutcome(
+                path=None, post_type=attrs_.get("post_type"),
+                title=attrs_.get("title"), body_chars=0,
+            )
         p = self.tmp_path / f"{post.get('id')}__post.json"
         p.write_text("{}")
         attrs = post.get("attributes") or {}
@@ -1114,3 +1137,289 @@ async def test_body_canary_silent_below_min_sample(source_id, sync_engine, tmp_p
     )
     assert result.success is True
     assert result.error_type is None
+
+
+# --- the revisit window: reaching posts that were EDITED after capture -------
+#
+# Operator, 2026-09-23, on a Floppystack post: *"this post has been updated as
+# he implements hot fixes — any chance we have a way to scan for or see updated
+# posts so we can update ours to match and pull the new attachments and
+# pictures etc."*
+#
+# A tick stopped after N contiguous already-have-it items, so an edit to a
+# three-day-old post was structurally unreachable: it sits well below twenty
+# seen items. The early-out now needs BOTH that run AND a post published before
+# the horizon.
+
+
+def _iso(days_ago):
+    return (datetime.now(UTC) - timedelta(days=days_ago)).isoformat()
+
+
+def _seed_seen(sync_engine, source_id, media):
+    factory = sessionmaker(sync_engine, expire_on_commit=False)
+    with factory() as s:
+        for m in media:
+            s.add(PatreonSeenMedia(
+                source_id=source_id, filehash=_ledger_key(m), post_id=m.post_id,
+            ))
+        s.commit()
+
+
+@pytest.mark.asyncio
+async def test_a_run_of_seen_items_does_not_stop_a_tick_inside_the_window(
+    source_id, sync_engine, tmp_path,
+):
+    """The bug itself. Three all-seen posts, threshold 2 — the old walk turned
+    around at the second and never looked at the third, which is exactly where
+    an edited post lives."""
+    seen = [_media(f"p{i}", 1) for i in range(1, 4)]
+    _seed_seen(sync_engine, source_id, seen)
+
+    client = _FakeClient(
+        [(None, [(m.post_id, [m]) for m in seen])],
+        published={m.post_id: _iso(3) for m in seen},
+    )
+    ing = _ingester(sync_engine, tmp_path, client, _FakeDownloader(tmp_path))
+
+    result = ing.run(
+        source_id=source_id, campaign_id="c1", artist_slug="ingest",
+        url="https://patreon.com/ingest", mode="tick", seen_threshold=2,
+        revisit_days=30,
+    )
+
+    assert result.success is True
+    assert client.consumed_posts == 3
+
+
+@pytest.mark.asyncio
+async def test_the_early_out_still_fires_once_the_walk_is_below_the_horizon(
+    source_id, sync_engine, tmp_path,
+):
+    """The other half, and the one that keeps a tick cheap. Same three posts,
+    published outside the window → the count early-out stands exactly as it
+    did. Asserted beside the test above because the window is only correct if
+    BOTH conditions are required; either one alone is a different feature."""
+    seen = [_media(f"p{i}", 1) for i in range(1, 4)]
+    _seed_seen(sync_engine, source_id, seen)
+
+    client = _FakeClient(
+        [(None, [(m.post_id, [m]) for m in seen])],
+        published={m.post_id: _iso(90) for m in seen},
+    )
+    ing = _ingester(sync_engine, tmp_path, client, _FakeDownloader(tmp_path))
+
+    result = ing.run(
+        source_id=source_id, campaign_id="c1", artist_slug="ingest",
+        url="https://patreon.com/ingest", mode="tick", seen_threshold=2,
+        revisit_days=30,
+    )
+
+    assert result.success is True
+    assert client.consumed_posts == 2
+
+
+@pytest.mark.asyncio
+async def test_a_window_of_zero_is_the_behaviour_that_shipped_before_it(
+    source_id, sync_engine, tmp_path,
+):
+    """The off switch. Recent posts, a window of 0 → the walk stops on the
+    count alone, so an operator who wants the old cheap tick has one."""
+    seen = [_media(f"p{i}", 1) for i in range(1, 4)]
+    _seed_seen(sync_engine, source_id, seen)
+
+    client = _FakeClient(
+        [(None, [(m.post_id, [m]) for m in seen])],
+        published={m.post_id: _iso(1) for m in seen},
+    )
+    ing = _ingester(sync_engine, tmp_path, client, _FakeDownloader(tmp_path))
+
+    ing.run(
+        source_id=source_id, campaign_id="c1", artist_slug="ingest",
+        url="https://patreon.com/ingest", mode="tick", seen_threshold=2,
+        revisit_days=0,
+    )
+
+    assert client.consumed_posts == 2
+
+
+@pytest.mark.asyncio
+async def test_a_backfill_ignores_the_window_because_it_never_early_outs(
+    source_id, sync_engine, tmp_path,
+):
+    """Stated so the window cannot quietly acquire a second job. A backfill
+    walks to the bottom regardless, and `recapture` is the mode that re-reads
+    every body — a horizon there would be a third answer to a two-answer
+    question."""
+    seen = [_media(f"p{i}", 1) for i in range(1, 4)]
+    _seed_seen(sync_engine, source_id, seen)
+
+    client = _FakeClient(
+        [(None, [(m.post_id, [m]) for m in seen])],
+        published={m.post_id: _iso(90) for m in seen},
+    )
+    downloader = _FakeDownloader(tmp_path)
+    ing = _ingester(sync_engine, tmp_path, client, downloader)
+
+    ing.run(
+        source_id=source_id, campaign_id="c1", artist_slug="ingest",
+        url="https://patreon.com/ingest", mode="backfill", seen_threshold=2,
+        revisit_days=30,
+    )
+
+    assert client.consumed_posts == 3
+    assert downloader.post_revisits == 0
+
+
+@pytest.mark.asyncio
+async def test_an_edited_post_inside_the_window_downloads_its_new_attachment(
+    source_id, sync_engine, tmp_path,
+):
+    """The operator's case, end to end.
+
+    Walk one captures a post with one file. The creator then edits it to attach
+    a hotfix build. Walk two must download that file AND name the post as
+    updated — the ask was to SEE which posts changed, not only to end up with
+    their bytes.
+
+    The detection half needed nothing new: `extract_media` reads the media list
+    off the LIVE feed response every walk, so a newly attached file is simply a
+    ledger key we have never seen. Only reaching the post was missing.
+    """
+    original = _media("p1", 1)
+    client1 = _FakeClient(
+        [(None, [("p1", [original])])], published={"p1": _iso(3)},
+    )
+    ing1 = _ingester(sync_engine, tmp_path, client1, _FakeDownloader(tmp_path))
+    ing1.run(
+        source_id=source_id, campaign_id="c1", artist_slug="ingest",
+        url="https://patreon.com/ingest", mode="tick", revisit_days=30,
+    )
+
+    hotfix = _media("p1", 2)
+    client2 = _FakeClient(
+        [(None, [("p1", [original, hotfix])])], published={"p1": _iso(3)},
+    )
+    downloader2 = _FakeDownloader(tmp_path)
+    ing2 = _ingester(sync_engine, tmp_path, client2, downloader2)
+    result = ing2.run(
+        source_id=source_id, campaign_id="c1", artist_slug="ingest",
+        url="https://patreon.com/ingest", mode="tick", revisit_days=30,
+    )
+
+    assert result.files_downloaded == 1
+    assert "1 post(s) updated (1 new file(s))" in result.stdout
+    assert "post p1 — updated: 1 new file(s)" in result.stdout
+    # The post record was re-read, not skipped by the capture gate.
+    assert downloader2.post_revisits == 1
+
+
+@pytest.mark.asyncio
+async def test_a_post_below_the_horizon_keeps_its_capture_gate(
+    source_id, sync_engine, tmp_path,
+):
+    """The gate still does its job everywhere the window does not reach — else
+    the window would have quietly become "re-read every post forever"."""
+    old = _media("pold", 1)
+    client1 = _FakeClient([(None, [("pold", [old])])], published={"pold": _iso(90)})
+    ing1 = _ingester(sync_engine, tmp_path, client1, _FakeDownloader(tmp_path))
+    ing1.run(
+        source_id=source_id, campaign_id="c1", artist_slug="ingest",
+        url="https://patreon.com/ingest", mode="tick", revisit_days=30,
+    )
+
+    client2 = _FakeClient([(None, [("pold", [old])])], published={"pold": _iso(90)})
+    downloader2 = _FakeDownloader(tmp_path)
+    ing2 = _ingester(sync_engine, tmp_path, client2, downloader2)
+    ing2.run(
+        source_id=source_id, campaign_id="c1", artist_slug="ingest",
+        url="https://patreon.com/ingest", mode="tick", revisit_days=30,
+    )
+
+    assert downloader2.post_records == 0
+
+
+@pytest.mark.asyncio
+async def test_a_revisit_that_finds_nothing_new_is_not_reported_as_an_update(
+    source_id, sync_engine, tmp_path,
+):
+    """Most revisits find nothing — that is the normal case, and a run summary
+    claiming "1 post(s) updated" on every tick would train the operator to stop
+    reading it (lesson: a signal that is always on is not a signal)."""
+    m = _media("p1", 1)
+    client1 = _FakeClient([(None, [("p1", [m])])], published={"p1": _iso(3)})
+    ing1 = _ingester(sync_engine, tmp_path, client1, _FakeDownloader(tmp_path))
+    ing1.run(
+        source_id=source_id, campaign_id="c1", artist_slug="ingest",
+        url="https://patreon.com/ingest", mode="tick", revisit_days=30,
+    )
+
+    client2 = _FakeClient([(None, [("p1", [m])])], published={"p1": _iso(3)})
+    downloader2 = _FakeDownloader(tmp_path)
+    ing2 = _ingester(sync_engine, tmp_path, client2, downloader2)
+    result = ing2.run(
+        source_id=source_id, campaign_id="c1", artist_slug="ingest",
+        url="https://patreon.com/ingest", mode="tick", revisit_days=30,
+    )
+
+    assert result.files_downloaded == 0
+    assert "updated" not in result.stdout
+    assert downloader2.post_revisits == 1
+
+
+@pytest.mark.asyncio
+async def test_revisits_do_not_feed_the_body_drift_canary(
+    source_id, sync_engine, tmp_path,
+):
+    """#862's canary fails a run that recorded a meaningful sample of posts and
+    got a body from NONE of them. A revisit legitimately comes back empty — the
+    post's body only ever arrived from the detail endpoint, which a revisit
+    declines to call — so counting revisits into that sample would walk the
+    alarm toward firing on healthy ticks. First captures only."""
+    posts = [(f"c{i}", []) for i in range(_CANARY_MIN_SAMPLE)]
+    published = {pid: _iso(3) for pid, _ in posts}
+
+    client1 = _FakeClient([(None, posts)], published=published)
+    ing1 = _ingester(sync_engine, tmp_path, client1, _FakeDownloader(tmp_path))
+    first = ing1.run(
+        source_id=source_id, campaign_id="c1", artist_slug="ingest",
+        url="https://patreon.com/ingest", mode="tick", revisit_days=30,
+    )
+    assert first.success is True
+
+    # Second walk: every post is a revisit, and every body comes back empty.
+    client2 = _FakeClient([(None, posts)], published=published, empty_body=True)
+    ing2 = _ingester(sync_engine, tmp_path, client2, _FakeDownloader(tmp_path))
+    second = ing2.run(
+        source_id=source_id, campaign_id="c1", artist_slug="ingest",
+        url="https://patreon.com/ingest", mode="tick", revisit_days=30,
+    )
+
+    assert second.success is True
+    assert second.error_type is not ErrorType.API_DRIFT
+
+
+@pytest.mark.asyncio
+async def test_a_post_whose_date_will_not_parse_falls_back_to_the_count(
+    source_id, sync_engine, tmp_path,
+):
+    """A date we cannot read must not fail the walk, and must not be guessed
+    into the window. It reads as "not provably recent", which leaves that post
+    on the behaviour it had before the window existed."""
+    seen = [_media(f"p{i}", 1) for i in range(1, 4)]
+    _seed_seen(sync_engine, source_id, seen)
+
+    client = _FakeClient(
+        [(None, [(m.post_id, [m]) for m in seen])],
+        published={m.post_id: "last Tuesday" for m in seen},
+    )
+    ing = _ingester(sync_engine, tmp_path, client, _FakeDownloader(tmp_path))
+
+    result = ing.run(
+        source_id=source_id, campaign_id="c1", artist_slug="ingest",
+        url="https://patreon.com/ingest", mode="tick", seen_threshold=2,
+        revisit_days=30,
+    )
+
+    assert result.success is True
+    assert client.consumed_posts == 2

tests/test_subscribestar_native.py

@@ -453,6 +453,29 @@ def test_write_post_record(tmp_path):
     assert rec.body_chars > 0
 
 
+def test_a_revisit_with_an_empty_body_writes_nothing(tmp_path):
+    """The revisit contract's second half, which applies here even though the
+    first (no detail-fetch) is free on SubscribeStar — there is no detail
+    endpoint. A chunk that parsed with no content must not overwrite a body we
+    already have, so a walk behaves the same on both platforms."""
+    dl = _downloader(tmp_path)
+    post = _post("222")
+    post["attributes"]["content"] = ""
+    rec = dl.write_post_record(post, "artist-x", revisit=True)
+
+    assert rec.path is None
+    assert rec.body_chars == 0
+    assert not list(tmp_path.rglob("_post.json"))
+
+
+def test_a_revisit_that_still_has_a_body_writes_it(tmp_path):
+    dl = _downloader(tmp_path)
+    rec = dl.write_post_record(_post("333"), "artist-x", revisit=True)
+
+    assert rec.path is not None
+    assert "body text" in json.loads(rec.path.read_text())["content"]
+
+
 def test_skip_seen_does_not_download(tmp_path):
     session = _FakeSession()
     dl = _downloader(tmp_path, session=session)