docs: give the clone command and size for packs

README.md

@@ -124,7 +124,7 @@ Checked by is the test your platform runs on its own, replicated here from its s
 
 ## Build your own pack
 
-Clone the repo and generate packs for any platform, emulator, or system:
+Packs are built from a clone, and the clone carries the collection: its checkout alone is over 7.7 GB. `git clone --depth 1 https://github.com/Abdess/retrobios.git` leaves the history out. Files kept as release assets are downloaded when a pack needs them. Then generate packs for any platform, emulator, or system:
 
 ```bash
 # The pack of a platform, as released

scripts/generate_readme.py

@@ -40,6 +40,7 @@ from nativemode import check_label
 import release_record
 from siteguides import install_command
 from validate_pr import LARGE_FILE_ISSUE_URL
+from largefiles import load_registered_paths
 from verify import verify_platform
 
 
@@ -447,6 +448,29 @@ REPO = "Abdess/retrobios"
 UPLOAD_URL = f"https://github.com/{REPO}/upload/main/bios"
 
 
+def clone_line(db: dict) -> str:
+    """How to get the clone the commands below run in, and its size.
+
+    Said bare, "clone the repo" started a multi-gigabyte transfer for a user
+    after a 20 MB --required-only pack. The size counts the collection files
+    the repository commits, a floor: release assets are fetched only when a
+    pack needs them.
+    """
+    assets = set(load_registered_paths())
+    committed = sum(
+        entry.get("size", 0) for entry in db.get("files", {}).values()
+        if entry.get("path") not in assets
+    )
+    return (
+        "Packs are built from a clone, and the clone carries the collection:"
+        f" its checkout alone is over {format_size(committed)}."
+        f" `git clone --depth 1 https://github.com/{REPO}.git` leaves the"
+        " history out. Files kept as release assets are downloaded when a"
+        " pack needs them. Then generate packs for any platform, emulator,"
+        " or system:"
+    )
+
+
 def _existing_contributor_block(readme_path: str = "README.md") -> list[str]:
     """The contributors already published, read back from the file.
 
@@ -767,7 +791,7 @@ def generate_readme(db: dict, platforms_dir: str, db_path: str = "database.json"
             "",
             "## Build your own pack",
             "",
-            "Clone the repo and generate packs for any platform, emulator, or system:",
+            clone_line(db),
             "",
             "```bash",
             "# The pack of a platform, as released",

tests/test_doc_pages.py

@@ -2031,3 +2031,22 @@ class RegionFreeRuleNamesWhereItDoesNotHold(unittest.TestCase):
                 if drops:
                     with self.subTest(platform=platform, file=name):
                         self.assertRegex(bullet, rf"(?i)\b{re.escape(platform)}\b")
+
+
+class BuildYourOwnPackSaysWhatTheCloneCosts(unittest.TestCase):
+    """The README sent a user after a minimal pack to "Clone the repo", with no
+    command and no size, and the clone runs to several gigabytes."""
+
+    def test_readme_gives_the_clone_command_and_its_size(self):
+        import generate_readme
+
+        db = json.loads((ROOT / "database.json").read_text(encoding="utf-8"))
+        line = generate_readme.clone_line(db)
+        self.assertIn("git clone --depth 1", line)
+        self.assertRegex(line, r"over \d+(\.\d)? GB")
+        section = _section((ROOT / "README.md").read_text(encoding="utf-8"),
+                           "Build your own pack")
+        self.assertIn("git clone --depth 1", section)
+
+    def test_advanced_usage_names_the_shallow_clone(self):
+        self.assertIn("--depth 1", _page("advanced-usage.md").split("## Custom packs")[0])

wiki/advanced-usage.md

@@ -19,7 +19,8 @@ boot, with no message saying why. It is not a guarantee either: source
 profiles can document files nobody has dumped or that only the user can
 provide, all of them visible in the [gap analysis](../gaps.md).
 
-From a clone of the repository, `python scripts/generate_pack.py --platform
+A clone carries the whole collection, several gigabytes, and `git clone
+--depth 1` leaves its history out. From that clone, `python scripts/generate_pack.py --platform
 retroarch --region us` builds a narrower pack, `--required-only` the bare
 minimum each core needs to start, and `--no-game-data` keeps the firmware and
 leaves out the files a game reads. `--help` lists every way to build your own,