Fix TYPE_CHECKING stub defaults in envs.py to match actual runtime defaults (#35645)

vllm/envs.py

@@ -35,7 +35,7 @@ if TYPE_CHECKING:
     VLLM_USAGE_STATS_SERVER: str = "https://stats.vllm.ai"
     VLLM_NO_USAGE_STATS: bool = False
     VLLM_DO_NOT_TRACK: bool = False
-    VLLM_USAGE_SOURCE: str = ""
+    VLLM_USAGE_SOURCE: str = "production"
     VLLM_CONFIGURE_LOGGING: bool = True
     VLLM_LOGGING_LEVEL: str = "INFO"
     VLLM_LOGGING_PREFIX: str = ""
@@ -48,7 +48,7 @@ if TYPE_CHECKING:
     VLLM_USE_FLASHINFER_SAMPLER: bool | None = None
     VLLM_PP_LAYER_PARTITION: str | None = None
     VLLM_CPU_KVCACHE_SPACE: int | None = 0
-    VLLM_CPU_OMP_THREADS_BIND: str = ""
+    VLLM_CPU_OMP_THREADS_BIND: str = "auto"
     VLLM_CPU_NUM_OF_RESERVED_CPU: int | None = None
     VLLM_CPU_SGL_KERNEL: bool = False
     VLLM_XLA_CACHE_PATH: str = os.path.join(VLLM_CACHE_ROOT, "xla_cache")
@@ -89,7 +89,7 @@ if TYPE_CHECKING:
     VLLM_LORA_RESOLVER_CACHE_DIR: str | None = None
     VLLM_LORA_RESOLVER_HF_REPO_LIST: str | None = None
     VLLM_USE_AOT_COMPILE: bool = False
-    VLLM_USE_BYTECODE_HOOK: bool = False
+    VLLM_USE_BYTECODE_HOOK: bool = True
     VLLM_FORCE_AOT_LOAD: bool = False
     VLLM_USE_MEGA_AOT_ARTIFACT: bool = False
     VLLM_USE_TRITON_AWQ: bool = False