docs: build RetroPie in the release pipeline run

The release built and checked the RetroPie pack with two commands after the pipeline, which left it out of the verify and pack consistency check. pipeline.py --include-archived builds and checks it with the rest.

tests/test_audit_regressions.py

@@ -5587,5 +5587,15 @@ class ThePipelineRunsTheSchemaCheckCIRuns(unittest.TestCase):
             self.assertIn(f'"{flag}"', source)
 
 
+class TheReleaseBuildsEveryServedPackInOneRun(unittest.TestCase):
+    """The release ran the pipeline, then built and checked RetroPie by hand:
+    the consistency check between verify and the pack never covered it."""
+
+    def test_the_pipeline_step_includes_archived_platforms(self):
+        page = (ROOT / "wiki" / "release-process.md").read_text(encoding="utf-8")
+        self.assertIn("python scripts/pipeline.py --include-archived", page)
+        self.assertNotIn("generate_pack.py --platform retropie", page)
+
+
 if __name__ == "__main__":
     unittest.main()

wiki/release-process.md

@@ -188,27 +188,25 @@ Releasing is deliberate and local. Nothing on GitHub builds a pack: the
 pipeline runs here, the archives are checked here, and `gh` uploads them.
 
 ```bash
-# 1. Full pipeline, online, so data directories and MAME/FBNeo hashes are fresh
-python scripts/pipeline.py
+# 1. Full pipeline, online, so data directories and MAME/FBNeo hashes are
+#    fresh. --include-archived builds RetroPie too, archived but still served,
+#    and holds its pack to the same consistency and integrity checks.
+python scripts/pipeline.py --include-archived
 
-# 2. RetroPie, which is archived but still served
-python scripts/generate_pack.py --platform retropie --output-dir dist/
-python scripts/generate_pack.py --platform retropie --verify-packs --output-dir dist/
-
-# 3. A release file must be under 2 GiB. A larger pack becomes parts that are
+# 2. A release file must be under 2 GiB. A larger pack becomes parts that are
 #    each a ZIP of whole files (Pack.part1of2.zip): any tool opens one, and
 #    the parts extracted into one folder are the pack. Each part is read back
 #    and the set compared with the pack before the pack is removed.
 python scripts/split_pack.py dist/
 
-# 4. Checksums of the files as published, then sign the list. The checksums
+# 3. Checksums of the files as published, then sign the list. The checksums
 #    answer corruption, and the signature answers a rewritten release, which is
 #    the one thing a checksum published beside its own artifacts cannot
 #    answer.
 (cd dist && sha256sum *.zip > SHA256SUMS.txt)
 ssh-keygen -Y sign -f ~/.ssh/retrobios_signing -n file dist/SHA256SUMS.txt
 
-# 5. Record what each pack holds, read from the archives just checked:
+# 4. Record what each pack holds, read from the archives just checked:
 #    file count, extracted size, download size, published files. The README
 #    and the site print these beside the download links, and the notes table
 #    takes its Size and Files columns from the same lines. The command stops
@@ -218,7 +216,7 @@ python scripts/release_record.py dist/ --tag "v${DATE}"
 python scripts/generate_readme.py --db database.json --platforms-dir platforms
 python scripts/generate_site.py
 
-# 6. Write the notes: install commands, the parts sentence and the pack table
+# 5. Write the notes: install commands, the parts sentence and the pack table
 #    come from release.json, and what changed since the previous tag is the
 #    maintainer's own text in tmp/whats-new.md, appended as written.
 python scripts/release_notes.py --whats-new tmp/whats-new.md --output tmp/notes.md
@@ -243,7 +241,7 @@ python scripts/check_release.py --tag "v${DATE}" --sums dist/SHA256SUMS.txt \
 git add release.json README.md mkdocs.yml
 git commit -m "chore: record release v${DATE}" && git push
 
-# 7. Keep only the new release plus large-files: an older pack carries hashes
+# 6. Keep only the new release plus large-files: an older pack carries hashes
 #    the platforms no longer check, so it misleads more than it helps
 gh release list --json tagName,createdAt \
   --jq 'sort_by(.createdAt) | reverse | .[].tagName' | grep -v '^large-files$' \
@@ -301,13 +299,13 @@ The release notes are written by `release_notes.py`: the install commands,
 the pack table and its parts sentence come from `release.json`, then what
 changed since the previous tag and the contributors of the closed issues,
 written by hand. A pack has two sizes and they are far apart: Batocera
-downloads as 2.4 GB and extracts to 4.0 GB. Step 5 prints both, and whichever
+downloads as 2.4 GB and extracts to 4.0 GB. Step 4 prints both, and whichever
 one the table carries, the header names it. Someone sizing a USB drive is
 reading that column. The README table is the extracted size of the released
 pack, from `release.json`. `SHA256SUMS.txt` lists the checksums of the files
 as published, parts included.
 
-`release.json` is the record step 5 writes: per pack, the files it holds, its
+`release.json` is the record step 4 writes: per pack, the files it holds, its
 two sizes and the files it is published in. It exists because the pages are
 regenerated on every push while a release is cut every few weeks. Read from
 the install manifests, the table gave the count main would build that day:
@@ -322,9 +320,9 @@ opened alone is not an archive and no tool says a part is missing, so five
 reports in six months took one for a broken download (issues 45, 51, 66, 77
 and 79). The README, the download page and the troubleshooting page describe
 both layouts for as long as a release cut that way is still published. Once
-step 7 has deleted it, the `.zip.001` paragraph leaves those three pages.
+step 6 has deleted it, the `.zip.001` paragraph leaves those three pages.
 
-The table carries a Files column, the count step 5 prints, and the
+The table carries a Files column, the count step 4 prints, and the
 notes never open on the size of the collection. That total covers every
 platform and emulator together and no pack holds it: v2026.09.04 led with
 "10,330 files" right after "one pack per platform", and someone who extracted