AI Implementation feature(1087): Mastodon Post Composition and Publishing (#6)

docs/architecture/pipeline-runner.md

@@ -10,9 +10,8 @@ timestamp: 2026-08-04T17:51:00Z
 `tenbackward.main` is the entry point executed by cron. The pipeline first
 ensures that the configured blog repository is available and current, then
 scans Jekyll posts for today's ten-year anniversary, filters already-recorded
-paths, and persists newly discovered identifiers. Job 1086 wires the
-anniversary matcher into the runner; Mastodon publishing remains a future
-step.
+paths, publishes a single combined Mastodon status for the new matches, and
+finally persists the freshly published identifiers.
 
 
 # Call Flow
@@ -28,8 +27,9 @@ main.main()
   ├── result = _run_with_retry(config)
   │       ├── for attempt in 1 .. max_retries+1:
   │       │     try: return _run_once(config)
-  │       │     └── _run_once → ensure_repo → iter_anniversary_paths
-  │       │         → PostedStore deduplication → mark_posted_many
+  │       │     └── _run_once → ensure_repo → find_anniversary_matches
+  │       │         → PostedStore deduplication → publish_mastodon
+  │       │         → mark_posted_many   # state only after publish OK
   │       ├── log_error("pipeline_error", exc=exc, attempt=attempt, max_attempts=attempts)
   │       ├── log_error("pipeline_failed", exc=last_exc, attempts=attempts)
   │       └── return None
@@ -40,12 +40,16 @@ main.main()
 # Extension Point: `_iter_candidates`
 
 ```python
-def _iter_candidates(config: Config) -> Iterable[str]: ...
+def _iter_candidates(config: Config) -> Iterable[MatchedPost]: ...
 ```
 
-Job 1086 wires `_iter_candidates` to the anniversary matcher. The matcher
-is now active; it discovers matching Jekyll posts after the repository is
-synchronized.
+`_iter_candidates` delegates to `find_anniversary_matches()` with
+`config.blog_dir / "_posts" / "blog"` and `config.site_url`. The matcher
+returns full `MatchedPost` values (relative path, title, date, canonical
+URL); the runner uses `MatchedPost.path` as a stable deduplication ID and
+passes the full objects to `publish_mastodon()` for status composition.
+See [Anniversary Matching](/architecture/anniversary-matching.md) for the
+file and front matter rules.
 
 ## Blog Synchronization
 
@@ -56,14 +60,6 @@ fail immediately with `BlogRepoError`. The same synchronization occurs once
 in `main()` before `startup` and again inside `_run_once()` as the retryable
 pipeline boundary.
 
-
-`_iter_candidates(config)` delegates to `iter_anniversary_paths()` with
-`config.blog_dir / "_posts" / "blog"` and `config.site_url`. The matcher
-returns relative Markdown paths, such as
-`2016/2016-08-04-example.md`, which are used as stable deduplication IDs.
-See [Anniversary Matching](/architecture/anniversary-matching.md) for the
-file and front matter rules.
-
 # `_run_once` — Summary Counters
 
 `_run_once` returns a 5-tuple:
@@ -86,6 +82,19 @@ The store persists the list under a `"posted"` key
 (`{"posted": ["2014/2014-08-04-foo.md", ...]}`) and serialises
 concurrent runs with an `fcntl.flock`.
 
+# Publish Boundary
+
+`publish_mastodon()` is the single success boundary. The runner calls it
+**after** dedupe and **before** `mark_posted_many`. If the call raises a
+`PublishError`, the runner treats it like any other pipeline exception —
+the retry loop in `_run_with_retry` re-attempts, and no identifier is
+written to `posted.json`. Identical-day matches are published as **one**
+combined status: prefix line + one `title\nurl` block per post + hashtags
+line, sorted by `(date, path)` for deterministic ordering. The default
+prefix is `Heute vor 10 Jahren:` (overridable via `THROWNBACK_PREFIX`).
+Generated statuses are validated against `MASTODON_STATUS_LIMIT` (500
+characters); the runner fails safely (raises) rather than truncating.
+
 # Retry Behaviour (`_run_with_retry`)
 
 * `attempts = max(1, config.max_retries + 1)` — at least one attempt