fix: scrape targets at the pinned platform release

scripts/scraper/targets/__init__.py

@@ -7,6 +7,7 @@ from __future__ import annotations
 
 import importlib
 import pkgutil
+import re
 from abc import ABC, abstractmethod
 from pathlib import Path
 
@@ -22,6 +23,35 @@ class BaseTargetScraper(ABC):
         """Fetch targets and their core lists. Returns dict matching target YAML format."""
         ...
 
+    @staticmethod
+    def pinned_revision(platform: str, platforms_dir: str | None = None) -> str:
+        """The upstream revision platforms/<platform>.yml was scraped at.
+
+        A target list describes the same release as the platform's BIOS
+        list: read on its own "latest stable tag", a Recalbox 10.2 board list
+        filtered 10.1.1 systems. Moving to a new release is a rescrape of the
+        platform first.
+        """
+        import yaml
+
+        directory = Path(platforms_dir) if platforms_dir else (
+            Path(__file__).resolve().parents[3] / "platforms"
+        )
+        with open(directory / f"{platform}.yml", encoding="utf-8") as handle:
+            source = str((yaml.safe_load(handle) or {}).get("source", ""))
+        for pattern in (
+            r"^https://raw\.githubusercontent\.com/[^/]+/[^/]+/([^/]+)/",
+            r"^https://github\.com/[^/]+/[^/]+/(?:blob|tree)/([^/]+)/",
+            r"^https://gitlab\.com/.+?/-/(?:raw|blob|tree)/([^/]+)/",
+        ):
+            match = re.match(pattern, source)
+            if match:
+                return match.group(1)
+        raise RuntimeError(
+            f"platforms/{platform}.yml names no pinned revision in its source "
+            f"({source!r}): rescrape the platform first"
+        )
+
     def write_output(self, data: dict, output_path: str) -> None:
         """Write target data to YAML file."""
         try: