ci: a release publishes a changelog, not an image (318 step 7)

Step 2 took the build consequence away from a `v*` tag — `main` has already
built and published the commit by the time anyone tags it, and rebuilding
would re-push `:c-<sha>`, which rule 145 forbids even when the source
matches. That left the tag with nothing to do at all.

This is the job it has instead. Step 6 put the derived version in the
Settings footer, so an operator can say WHICH build they are running; this
says what is in it that was not in the one they ran last month. Both halves
of one question (note #3127 §5).

The previous release is found by walking ANCESTRY, not by sorting a list.
That is load-bearing here specifically: rule 148 moved the tag shape from
`v26.05.22.0` to `v2026.08.28.2208`, and lexicographically `v2026...` sorts
BEFORE `v26...` — the third character is `0` against `6`. A sorted
implementation would reach back past every new-shape tag to the newest
old-shape one and publish months of commits as "changes since", looking
entirely correct while doing it. `git describe --exclude` is immune to the
shape change, and reachability is the more honest question anyway.

The publisher GETs and PATCHes rather than POSTing and recovering the id
from a 409 — note #3127 §6.7, which is ThoughtSync #2182's bug. A `v*` tag
is created once so the conflict path is rare, but "rare" is how that one
survived to be found somewhere else.

Cross-checks are reported on the release, not enforced. The tag is already
pushed by the time this runs, so failing would leave the operator with a tag,
no release, and a red lane to explain it — while the release is still the
useful object. It says so at the top when the tag names a version the web
image does not report, or when the commit is not on `main` and the `:c-`
rollback refs it lists were never published.

Nothing runs on a schedule and nothing auto-tags on merge. Release tags are
bookmarks (note #3127 §0); FC went twelve weeks without one and nothing was
wrong.

Also here:
- `scripts/` joins the ruff lane. release_notes.py runs only on a tag push,
  so a syntax error there would otherwise surface at the one moment nobody
  wants to be debugging a workflow.
- version.spec.js reads the workflow directory instead of listing three
  files by hand. Its own comment says the assertion should survive consumers
  coming and going; the hardcoded list was the part that could not, and
  release.yml would have joined the directory without joining the check.

Tests build a synthetic history spanning the tag-shape change rather than
leaning on this repo's tags, so the span assertion holds whether or not a
checkout brought the tags along — a span test that quietly skips is worse
than one that fails.

tests/test_release_notes.py

@@ -0,0 +1,149 @@
+"""What the release changelog promises, and the way it would lie quietly.
+
+A changelog has no consumer that checks it. If it lists the wrong span nothing
+fails — the release publishes, reads perfectly, and tells the operator that a
+month of work landed in a build that never contained it. That is the same
+silent-and-plausible failure class as a revision that identifies the wrong
+content (`test_artifact_identity.py` guards the other side of it), so the span
+selection is asserted rather than eyeballed.
+
+Everything runs the script the way `release.yml` runs it — as a subprocess,
+through `--dry-run`. That is the same code path as a real publish right up to
+the HTTP call, so these exercise the interface CI uses instead of a Python
+re-implementation of it.
+"""
+from __future__ import annotations
+
+import subprocess
+from pathlib import Path
+
+import pytest
+
+ROOT = Path(__file__).resolve().parent.parent
+SCRIPT = ROOT / "scripts" / "release_notes.py"
+
+
+def notes(*args: str, cwd: Path | None = None) -> str:
+    return subprocess.run(
+        ["python3", str(SCRIPT), "--dry-run", *args],
+        capture_output=True, text=True, check=True, cwd=cwd or ROOT,
+    ).stdout
+
+
+def body_of(out: str) -> str:
+    assert "--- body ---" in out, f"no body was rendered:\n{out}"
+    return out.split("--- body ---", 1)[1]
+
+
+def git(repo: Path, *args: str) -> str:
+    return subprocess.run(
+        ["git", "-c", "user.email=ci@example.invalid", "-c", "user.name=ci",
+         "-c", "commit.gpgsign=false", *args],
+        capture_output=True, text=True, check=True, cwd=repo,
+    ).stdout.strip()
+
+
+@pytest.fixture
+def shaped_history(tmp_path: Path) -> Path:
+    """Three releases spanning the rule 148 tag-shape change.
+
+    Ancestry order is `v26.06.04.0` → `v2026.08.28.2208` → `v2026.08.29.1000`,
+    which is the exact arrangement where walking ancestry and sorting a list
+    disagree — see the test below. Synthetic rather than taken from this repo's
+    own tags so it holds whether or not CI's checkout brought the tags along:
+    a span test that quietly skips is the one outcome worse than a failing one.
+    """
+    repo = tmp_path / "shaped"
+    repo.mkdir()
+    git(repo, "init", "-q", "-b", "main")
+    for i, tag in enumerate(("v26.06.04.0", "v2026.08.28.2208", "v2026.08.29.1000")):
+        (repo / "f.txt").write_text(f"{i}\n")
+        git(repo, "add", "f.txt")
+        git(repo, "commit", "-q", "-m", f"work landing in {tag}")
+        git(repo, "tag", tag)
+    # One more commit and a merge, so the merge-exclusion test has something to
+    # exclude that a first-parent listing would otherwise show.
+    git(repo, "checkout", "-q", "-b", "side")
+    (repo / "g.txt").write_text("side\n")
+    git(repo, "add", "g.txt")
+    git(repo, "commit", "-q", "-m", "feat: work done on the side branch")
+    git(repo, "checkout", "-q", "main")
+    git(repo, "merge", "-q", "--no-ff", "side", "-m", "Merge pull request #999 from side")
+    git(repo, "tag", "v2026.08.30.0900")
+    return repo
+
+
+def test_the_previous_release_is_found_by_ancestry_not_by_sorting(shaped_history):
+    """The trap this repo is standing in right now.
+
+    Rule 148 moved the tag shape from `v26.05.22.0` to `v2026.08.28.2208`.
+    Lexicographically `v2026...` sorts BEFORE `v26...` — the third character is
+    `0` against `6` — so a sorted-list implementation reaches back past every
+    new-shape tag to the newest OLD-shape one and emits months of commits as
+    "changes since". It looks entirely correct on any repo whose tags share a
+    single shape, which is every repo until the day the shape changes.
+
+    Here, ancestry says `v2026.08.28.2208` and sorting says `v26.06.04.0`.
+    """
+    out = notes("v2026.08.29.1000", cwd=shaped_history)
+    assert "previous=v2026.08.28.2208" in out
+    assert "v26.06.04.0" not in out
+
+
+def test_the_body_names_the_span_it_actually_listed(shaped_history):
+    """A body whose heading says "since X" over commits computed from Y is
+    unfalsifiable from outside — both halves read fine on their own."""
+    body = body_of(notes("v2026.08.29.1000", cwd=shaped_history))
+    assert "## Changes since v2026.08.28.2208" in body
+    assert "v2026.08.28.2208..v2026.08.29.1000" in body
+    assert "work landing in v2026.08.29.1000" in body
+    assert "work landing in v2026.08.28.2208" not in body
+
+
+def test_merges_are_excluded_so_the_list_is_the_work(shaped_history):
+    """Rule 153 merges dev into main with a plain merge commit, so `main`'s
+    first-parent view is nothing but "Merge pull request #N". Including those
+    would publish a changelog of PR numbers over the actual changes."""
+    body = body_of(notes("v2026.08.30.0900", cwd=shaped_history))
+    assert "feat: work done on the side branch" in body
+    assert "Merge pull request #999" not in body
+
+
+def test_the_first_release_still_renders_with_nothing_behind_it(shaped_history):
+    """No previous tag is reachable from the oldest one. That is a real state,
+    not an error, and it must not take the release down with it."""
+    out = notes("v26.06.04.0", cwd=shaped_history)
+    assert "previous=<none>" in out
+    assert "## Changes" in body_of(out)
+
+
+def test_a_non_tag_ref_renders_but_refuses_to_claim_it_published():
+    """`--dry-run HEAD` is the operator's preview before deciding to cut a tag
+    at all. It must not describe itself as a release that happened."""
+    out = notes("HEAD")
+    assert "which is not a tag" in out
+    body_of(out)
+
+
+def test_the_rollback_refs_name_all_three_images():
+    """Rule 145: `:c-<sha>` is the rollback unit, and the three images move
+    together. A release listing only the web image sends an operator into a
+    rollback that leaves ml and agent on the newer build — the exact mismatch
+    build.yml builds all three on every push to avoid."""
+    body = body_of(notes("HEAD"))
+    for image in ("fabledcurator", "fabledcurator-ml", "fabledcurator-agent"):
+        assert f"bvandeusen/{image}:c-" in body, f"{image} missing from the rollback refs"
+
+
+def test_an_unbounded_span_is_truncated_and_says_so():
+    """With no reachable previous tag the span is the whole history. Emitting
+    eleven hundred lines would bury the one line explaining why there are
+    eleven hundred of them, so the cap is part of the message, not a silent
+    slice."""
+    out = notes("HEAD")
+    if "previous=<none>" not in out:
+        pytest.skip("a previous tag is reachable from HEAD in this checkout")
+    body = body_of(out)
+    listed = [ln for ln in body.split("\n") if ln.startswith("- ")]
+    assert len(listed) <= 200
+    assert "more than a changelog is for" in body