fix: lay out both builds of a dual-mode emulator

Batocera, RetroBat, Recalbox and ROCKNIX offer some emulators both as a
libretro core and standalone. Filed as standalone only, the libretro
layout was dropped: libretro Dolphin had no dolphin-emu/Sys/GC files.
Platforms now name dual-mode emulators and each gets both layouts.

scripts/scraper/batocera_scraper.py

@@ -18,6 +18,7 @@ from pathlib import Path
 
 
 from common import yaml_load
+from safeparse import yaml_load_upstream
 
 from .base_scraper import (
     BaseScraper,
@@ -98,7 +99,7 @@ def es_systems_emulators(text: str) -> tuple[set[str], set[str]]:
     reading it alone left DOSBox Staging, melonDS DS and nine other built
     cores outside the pack.
     """
-    data = yaml_load(text)
+    data = yaml_load_upstream(text)
     if not isinstance(data, dict):
         raise ValueError("es_systems.yml is not a mapping")
     cores: set[str] = set()
@@ -126,6 +127,20 @@ def es_systems_emulators(text: str) -> tuple[set[str], set[str]]:
     return cores, standalone
 
 
+def es_systems_libretro(text: str) -> set[str]:
+    """The cores es_systems.yml offers under `libretro:` for some system."""
+    data = yaml_load_upstream(text)
+    if not isinstance(data, dict):
+        raise ValueError("es_systems.yml is not a mapping")
+    found: set[str] = set()
+    for system in data.values():
+        emulators = system.get("emulators") if isinstance(system, dict) else None
+        libretro = emulators.get("libretro") if isinstance(emulators, dict) else None
+        if isinstance(libretro, dict):
+            found.update(str(core) for core, spec in libretro.items() if isinstance(spec, dict))
+    return found
+
+
 SYSTEM_SLUG_MAP = {
     "advision": "entex-advision",
     "amiga500": "commodore-amiga",
@@ -253,11 +268,12 @@ class Scraper(BaseScraper):
         self.tag = fetch_stable_tag()
         super().__init__(url=SOURCE_URL.format(tag=self.tag))
 
-    def _fetch_cores(self) -> tuple[list[str], list[str]]:
+    def _fetch_cores(self) -> tuple[list[str], list[str], list[str]]:
         """Extract core names and standalone cores from configgen-defaults.yml.
 
-        Returns (all_cores, standalone_cores) where standalone_cores are
-        those with emulator != "libretro".
+        Returns (all_cores, standalone_cores, dual_mode_cores): standalone
+        cores are those with emulator != "libretro", dual-mode ones those
+        es_systems.yml also offers under libretro (dolphin for GameCube).
         """
         url = CONFIGGEN_DEFAULTS_URL.format(tag=self.tag)
         try:
@@ -271,10 +287,10 @@ class Scraper(BaseScraper):
             raise ConnectionError(
                 f"Failed to fetch {url}: {e}"
             ) from e
-        data = yaml_load(raw)
-        cores, standalone = es_systems_emulators(
-            self._fetch_text(ES_SYSTEMS_URL.format(tag=self.tag))
-        )
+        data = yaml_load_upstream(raw)
+        es_systems = self._fetch_text(ES_SYSTEMS_URL.format(tag=self.tag))
+        cores, standalone = es_systems_emulators(es_systems)
+        libretro = es_systems_libretro(es_systems)
         if not cores:
             raise RuntimeError("es_systems.yml names no emulator")
         for system, cfg in data.items():
@@ -289,7 +305,8 @@ class Scraper(BaseScraper):
                 if core and core != emulator:
                     standalone.add(core)
         dropped = _not_cores()
-        return sorted(cores - dropped), sorted(standalone - dropped)
+        dual = (standalone & libretro) - dropped
+        return sorted(cores - dropped), sorted(standalone - dropped), sorted(dual)
 
     def _fetch_text(self, url: str) -> str:
         req = urllib.request.Request(url, headers={"User-Agent": "retrobios-scraper/1.0"})
@@ -449,7 +466,7 @@ class Scraper(BaseScraper):
 
             systems[req.system]["files"].append(requirement_entry(req))
 
-        cores, standalone = self._fetch_cores()
+        cores, standalone, dual = self._fetch_cores()
         result = {
             "platform": "Batocera",
             "version": self.tag.removeprefix("batocera-"),
@@ -463,6 +480,9 @@ class Scraper(BaseScraper):
         }
         if standalone:
             result["standalone_cores"] = standalone
+        if dual:
+            # Launched both ways: each layout gets its files.
+            result["dual_mode_cores"] = dual
         return result
 
 

scripts/scraper/recalbox_scraper.py

@@ -153,6 +153,18 @@ def _core_names(values: dict[str, str]) -> tuple[set[str], set[str]]:
     return names, set(names)
 
 
+def system_ini_libretro(text: str) -> set[str]:
+    """The libretro cores one Recalbox system.ini offers."""
+    import re
+
+    return {
+        values.get("core", "")
+        for section, values in _ini_sections(text).items()
+        if re.fullmatch(r"core\.\d+(\.[\w-]+)?", section)
+        and values.get("emulator") == "libretro" and values.get("core")
+    }
+
+
 def system_ini_cores(text: str) -> tuple[set[str], set[str]]:
     """The cores one Recalbox system.ini offers, and those run standalone.
 
@@ -227,8 +239,9 @@ class Scraper(BaseScraper):
             raise RuntimeError(f"no system.ini at {self.tag} for {', '.join(missing[:5])}")
         return [texts[path] for path in paths]
 
-    def _fetch_cores(self) -> tuple[list[str], list[str]]:
-        """Every core Recalbox's systems offer, and those it runs standalone.
+    def _fetch_cores(self) -> tuple[list[str], list[str], list[str]]:
+        """Every core Recalbox's systems offer, those it runs standalone, and
+        those it offers both as a libretro core and standalone.
 
         Read from each system.ini of recalbox-romfs2 at the scraped tag: the
         systems ES lists and the cores a user can pick for each.
@@ -250,13 +263,15 @@ class Scraper(BaseScraper):
             raise RuntimeError(f"no Recalbox system found at {self.tag}")
         cores: set[str] = set()
         standalone: set[str] = set()
+        libretro: set[str] = set()
         for text in self._system_inis(systems):
             found, alone = system_ini_cores(text)
             cores |= found
             standalone |= alone
+            libretro |= system_ini_libretro(text)
         if not cores:
             raise RuntimeError(f"no core in Recalbox's system.ini files at {self.tag}")
-        return sorted(cores), sorted(standalone)
+        return sorted(cores), sorted(standalone), sorted(libretro & standalone)
 
     def fetch_requirements(self) -> list[BiosRequirement]:
         """Parse es_bios.xml and return BIOS requirements."""
@@ -350,7 +365,7 @@ class Scraper(BaseScraper):
 
             systems[req.system]["files"].append(requirement_entry(req))
 
-        cores, standalone = self._fetch_cores()
+        cores, standalone, dual = self._fetch_cores()
         return {
             "platform": "Recalbox",
             "version": self.tag,
@@ -361,6 +376,8 @@ class Scraper(BaseScraper):
             "verification_mode": "md5",
             "cores": cores,
             "standalone_cores": standalone,
+            # Launched both ways: each layout gets its files.
+            **({"dual_mode_cores": dual} if dual else {}),
             "systems": systems,
         }
 

scripts/scraper/retrobat_scraper.py

@@ -44,6 +44,21 @@ ES_SYSTEMS_URL = (
 )
 
 
+def es_systems_dual(text: str) -> set[str]:
+    """Emulators es_systems.cfg offers both as a libretro core and standalone."""
+    from common import parse_untrusted_xml
+
+    libretro: set[str] = set()
+    standalone: set[str] = set()
+    for emulator in parse_untrusted_xml(text, "es_systems.cfg").iter("emulator"):
+        name = (emulator.get("name") or "").strip()
+        if name == "libretro":
+            libretro |= {(core.text or "").strip() for core in emulator.iter("core")} - {""}
+        elif name:
+            standalone.add(name)
+    return libretro & standalone
+
+
 def es_systems_cores(text: str) -> tuple[set[str], set[str]]:
     """Every core RetroBat's ES offers, and the emulators it runs standalone.
 
@@ -207,7 +222,9 @@ class Scraper(BaseScraper):
         req = urllib.request.Request(es_url, headers={"User-Agent": "retrobios-scraper/1.0"})
         try:
             with urllib.request.urlopen(req, timeout=30) as resp:
-                cores, standalone = es_systems_cores(resp.read().decode("utf-8"))
+                es_text = resp.read().decode("utf-8")
+            cores, standalone = es_systems_cores(es_text)
+            dual = es_systems_dual(es_text)
         except urllib.error.URLError as e:
             raise ConnectionError(f"Failed to fetch {es_url}: {e}") from e
         if not cores:
@@ -223,6 +240,8 @@ class Scraper(BaseScraper):
             "verification_mode": "md5",
             "cores": sorted(cores),
             "standalone_cores": sorted(standalone),
+            # Launched both ways: each layout gets its files.
+            **({"dual_mode_cores": sorted(dual)} if dual else {}),
             "systems": systems,
         }
 

scripts/scraper/rocknix_scraper.py

@@ -236,6 +236,9 @@ class Scraper(BaseScraper):
             # emulators ROCKNIX packages beside it.
             "cores": ["retroarch", *standalone],
             "standalone_cores": standalone,
+            # RetroArch carries every libretro core beside them, so an
+            # emulator with a libretro build is launched both ways.
+            "dual_mode_cores": list(standalone),
             "systems": systems,
         }
 

scripts/scraper/targets/batocera_targets_scraper.py

@@ -23,7 +23,7 @@ from datetime import datetime, timezone
 
 import yaml
 
-from common import yaml_load
+from safeparse import yaml_load_upstream
 
 from . import BaseTargetScraper
 
@@ -244,7 +244,7 @@ def _parse_es_systems(text: str) -> dict[str, list[str]]:
           <core_name>: {requireAnyOf: [BR2_PACKAGE_FOO]}
     """
     try:
-        data = yaml_load(text)
+        data = yaml_load_upstream(text)
     except yaml.YAMLError:
         return {}