Test that the shipped data stays under 100 MiB and that publish_release.py stops on a tag at another commit, both in the gate

.github/workflows/test.yml

@@ -27,7 +27,7 @@ jobs:
           else
             base='${{ github.event.before }}'
           fi
-          if git diff --quiet "$base"...HEAD -- '*.go' go.mod go.sum Dockerfile .dockerignore .github data testdata README.md; then
+          if git diff --quiet "$base"...HEAD -- '*.go' go.mod go.sum Dockerfile .dockerignore .github data release-tooling testdata README.md; then
             echo "code=false" >> "$GITHUB_OUTPUT"
           else
             echo "code=true" >> "$GITHUB_OUTPUT"

Dockerfile

@@ -11,6 +11,12 @@ RUN go vet ./... && \
 	go run github.com/fzipp/gocyclo/cmd/gocyclo@v0.6.0 -over 14 -ignore _test . && \
 	go test -race ./...
 
+FROM python:3.14.8-slim AS release-tooling
+COPY release-tooling/ /release-tooling/
+RUN python3 -m unittest discover -s /release-tooling -p '*_test.py' && touch /tested
+
 FROM portable
+# BuildKit skips a stage nothing copies from.
+COPY --from=release-tooling /tested /tmp/release-tooling-tested
 RUN unformatted="$(gofmt -l .)"; test -z "$unformatted" || \
 	{ echo "unformatted files:"; echo "$unformatted"; exit 1; }

release-tooling/publish_release_test.py

@@ -0,0 +1,78 @@
+import io
+import os
+import tempfile
+import unittest
+import urllib.error
+from contextlib import redirect_stderr, redirect_stdout
+from unittest import mock
+
+import publish_release
+
+SHA = "a" * 40
+OTHER = "b" * 40
+
+
+def not_found(path):
+	return urllib.error.HTTPError(path, 404, "Not Found", None, None)
+
+
+class TagGuard(unittest.TestCase):
+	def setUp(self):
+		cwd = os.getcwd()
+		dir = tempfile.TemporaryDirectory()
+		self.addCleanup(dir.cleanup)
+		os.chdir(dir.name)
+		self.addCleanup(os.chdir, cwd)
+		with open("CHANGELOG.md", "w", encoding="utf-8") as f:
+			f.write("# Changelog\n\n## [1.2.3]\n\n- A change.\n")
+		env = mock.patch.dict(os.environ, {"SHA": SHA})
+		env.start()
+		self.addCleanup(env.stop)
+
+	def publish(self, responses):
+		created = []
+
+		def request(path, data=None):
+			if data is not None:
+				created.append(data)
+				return {"html_url": "https://forge.example/release"}
+			response = responses.get(path)
+			if response is None:
+				raise not_found(path)
+			return response
+
+		with mock.patch.object(publish_release, "request", request), \
+			redirect_stdout(io.StringIO()), redirect_stderr(io.StringIO()):
+			code = publish_release.main()
+		return code, created
+
+	def test_a_tag_at_another_commit_publishes_nothing(self):
+		code, created = self.publish({"git/ref/tags/v1.2.3": {"object": {"sha": OTHER, "type": "commit"}}})
+		self.assertEqual((code, created), (1, []))
+
+	def test_an_annotated_tag_at_another_commit_publishes_nothing(self):
+		code, created = self.publish({
+			"git/ref/tags/v1.2.3": {"object": {"sha": "c" * 40, "type": "tag"}},
+			"git/tags/" + "c" * 40: {"object": {"sha": OTHER, "type": "commit"}},
+		})
+		self.assertEqual((code, created), (1, []))
+
+	def test_an_annotated_tag_at_this_commit_publishes(self):
+		code, created = self.publish({
+			"git/ref/tags/v1.2.3": {"object": {"sha": "c" * 40, "type": "tag"}},
+			"git/tags/" + "c" * 40: {"object": {"sha": SHA, "type": "commit"}},
+		})
+		self.assertEqual((code, [c["tag_name"] for c in created]), (0, ["v1.2.3"]))
+
+	def test_no_tag_publishes_one_at_this_commit(self):
+		code, created = self.publish({})
+		self.assertEqual(code, 0)
+		self.assertEqual(created, [{"body": "- A change.", "name": "v1.2.3", "tag_name": "v1.2.3", "target_commitish": SHA}])
+
+	def test_a_published_release_publishes_nothing(self):
+		code, created = self.publish({"releases/tags/v1.2.3": {"html_url": "https://forge.example/release"}})
+		self.assertEqual((code, created), (0, []))
+
+
+if __name__ == "__main__":
+	unittest.main()

shipped_data_test.go

@@ -2,6 +2,7 @@ package fejkdata
 
 import (
 	"io/fs"
+	"os"
 	"reflect"
 	"slices"
 	"strings"
@@ -18,6 +19,27 @@ func TestABareNewFailsNamingWithDataFS(t *testing.T) {
 	}
 }
 
+func TestShippedDataStaysUnder100MiB(t *testing.T) {
+	var size int64
+	err := fs.WalkDir(os.DirFS("data"), ".", func(path string, d fs.DirEntry, err error) error {
+		if err != nil || d.IsDir() {
+			return err
+		}
+		info, err := d.Info()
+		if err != nil {
+			return err
+		}
+		size += info.Size()
+		return nil
+	})
+	if err != nil {
+		t.Fatal(err)
+	}
+	if size > 100<<20 {
+		t.Errorf("data/ holds %d bytes, over 100 MiB: carry out todo.md item 193, moving the largest packages into nested Go modules", size)
+	}
+}
+
 func TestEveryShippedModuleLoads(t *testing.T) {
 	f, err := New(withShipped(), WithSeed(1))
 	if err != nil {