feat(mcfold): P5 mod jar 解析（三种加载器+多加载器诊断+方块名提取，含诚实边界标注）

tests/test_modjar.py

@@ -0,0 +1,126 @@
+"""mod jar 解析回归 —— 用**真实 mod jar** 验证（三种加载器 + 边界）。"""
+import sys, os, json, zipfile, io, tempfile
+sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+from mcfold.modjar import parse_mod, scan_mods, summarize, _parse_toml, detect_loaders
+
+FORGE   = "/home/mc-test/mods/curtain-mc1.20.1-1.3.2.jar"
+NEOFORGE= "/home/minecraft/mods/bettercombat-neoforge-2.3.2+1.21.1.jar"
+MULTI   = "/home/minecraft/mods/Explorify v1.6.4 f15-88.mod.jar"
+BLOCKS  = "/home/minecraft/mods/[扩充乐事] expandeddelight-0.1.4.jar"
+
+def exists(p): return os.path.isfile(p)
+
+def test_toml_parser_comment_headers():
+    """★ 踩过的坑：表头带行内注释（'[[mods]] #mandatory'）必须正确识别。"""
+    text = '''
+loaderVersion = "[47,)"
+[[mods]] #mandatory
+modId = "curtain" #mandatory
+version = "${file.jarVersion}"
+displayName = "Curtain"
+[[dependencies.curtain]]
+modId = "forge"
+'''
+    m = _parse_toml(text)
+    assert len(m["mods"]) == 1, f"mods 段数应为 1，实际 {len(m['mods'])}"
+    assert m["mods"][0]["modId"] == "curtain", m["mods"][0]
+    assert m["mods"][0]["displayName"] == "Curtain"
+    # 依赖段的 modId 不应污染 mods[0]
+    assert m["mods"][0]["modId"] != "forge"
+    print("  ✓ TOML: 带行内注释的 [[mods]] 正确识别，依赖段不污染")
+
+def test_forge_real():
+    if not exists(FORGE):
+        print("  - forge 样本不存在，跳过"); return
+    m = parse_mod(FORGE)
+    assert m.loader == "forge", m.loader
+    assert m.mod_id == "curtain", m.mod_id
+    assert m.name == "Curtain", m.name
+    # curtain 只加命令、不新增方块 —— 0 方块是正确结果
+    assert m.block_names == [], m.block_names
+    print(f"  ✓ Forge 真实: id={m.mod_id} name={m.name} 方块=0（它只加命令）")
+
+def test_neoforge_real():
+    if not exists(NEOFORGE):
+        print("  - neoforge 样本不存在，跳过"); return
+    m = parse_mod(NEOFORGE)
+    assert m.loader == "neoforge", m.loader
+    assert m.mod_id == "bettercombat", m.mod_id
+    assert m.version == "2.3.2+1.21.1", m.version
+    print(f"  ✓ NeoForge 真实: id={m.mod_id} ver={m.version}")
+
+def test_multi_loader_jar():
+    """多加载器 jar（同时含 forge/neoforge/fabric 元数据）必须能诊断。"""
+    if not exists(MULTI):
+        print("  - 多加载器样本不存在，跳过"); return
+    m = parse_mod(MULTI)
+    assert len(m.loaders_present) >= 2, m.loaders_present
+    assert m.mod_id == "explorify", m.mod_id
+    # 结构 mod：用原版方块生成地形，自己不新增方块 ⇒ 0 方块正确
+    assert m.block_names == [], m.block_names
+    print(f"  ✓ 多加载器 jar: 含 {m.loaders_present}，id={m.mod_id}（结构 mod，0 方块）")
+
+def test_block_extraction_real():
+    """★ 核心：从真实 mod 提取方块名清单（这是 mod 支持的可见部分）。"""
+    if not exists(BLOCKS):
+        print("  - 方块 mod 样本不存在，跳过"); return
+    m = parse_mod(BLOCKS)
+    assert m.mod_id == "expandeddelight", m.mod_id
+    assert len(m.block_names) > 20, f"方块数偏少: {len(m.block_names)}"
+    # 已知方块（前面实测过）
+    for want in ("cask", "cheese_wheel", "asparagus_crate"):
+        assert want in m.block_names, f"缺方块 {want}"
+    q = m.qualified_blocks()
+    assert "expandeddelight:cask" in q, q[:5]
+    # 物品也应提取到
+    assert len(m.item_names) > 20, len(m.item_names)
+    print(f"  ✓ 方块提取: {len(m.block_names)} 方块 / {len(m.item_names)} 物品，含 cask/cheese_wheel")
+
+def test_no_state_id_claim():
+    """★ 诚实性：解析结果里**不含** state id（因为拿不到），摘要必须明说。"""
+    m = parse_mod(BLOCKS) if exists(BLOCKS) else None
+    if m is None:
+        print("  - 跳过"); return
+    s = summarize([m])
+    assert "state id" in s["note"], s["note"]
+    assert "block_names" in s and s["block_name_count"] > 0
+    # 摘要里不应有任何形如 state id 的映射承诺
+    assert "state_ids" not in s and "stateId" not in json.dumps(s)
+    print(f"  ✓ 诚实标注: 摘要含 {s['block_name_count']} 方块名，并注明无 state id")
+
+def test_bad_jar_no_crash():
+    """损坏的 jar 不能中断整目录扫描。"""
+    with tempfile.TemporaryDirectory() as d:
+        p = os.path.join(d, "broken.jar")
+        open(p, "wb").write(b"this is not a zip")
+        m = parse_mod(p)
+        assert not m.ok and "zip" in (m.error or "").lower(), m.error
+        good = os.path.join(d, "ok.jar")
+        with zipfile.ZipFile(good, "w") as zf:
+            zf.writestr("fabric.mod.json", json.dumps({"id":"t","name":"T","version":"1"}))
+        mods = scan_mods(d)
+        assert len(mods) == 2, len(mods)
+        s = summarize(mods)
+        assert s["loaded"] == 1 and s["failed"] == 1, s
+        print(f"  ✓ 损坏 jar 不中断扫描: loaded={s['loaded']} failed={s['failed']}")
+
+def test_json_meta_without_assets():
+    """纯 JSON 元数据、无 assets 的 mod 也要能拿到 id（边界）。"""
+    with tempfile.TemporaryDirectory() as d:
+        p = os.path.join(d, "t.jar")
+        with zipfile.ZipFile(p, "w") as zf:
+            zf.writestr("fabric.mod.json", json.dumps(
+                {"id":"mymod","name":"My Mod","version":"2.0",
+                 "depends":{"minecraft":">=1.20","fabric-api":"*"}}))
+        m = parse_mod(p)
+        assert m.loader == "fabric" and m.mod_id == "mymod", (m.loader, m.mod_id)
+        assert m.depends.get("minecraft") == ">=1.20", m.depends
+        assert m.block_names == []
+        print("  ✓ fabric 纯元数据: id/depends 正确，无 assets 不报错")
+
+if __name__ == "__main__":
+    for fn in [test_toml_parser_comment_headers, test_forge_real, test_neoforge_real,
+               test_multi_loader_jar, test_block_extraction_real, test_no_state_id_claim,
+               test_bad_jar_no_crash, test_json_meta_without_assets]:
+        fn()
+    print("\nmod jar 解析：全部通过")