debug(phase2): 细化插桩区分 CreateProcessA 卡点与 join 等待行为

App/lib/core/BuildPipeline.vp

@@ -532,13 +532,22 @@ def _run_cmd_wait(cmd: str) -> int:
     sa.bInheritHandle = 1
     si.cb = win32process.STARTUPINFOA.__sizeof__()
     # create_flags=0：子进程继承父控制台（输出直接显示，同 system 语义）
+    #region debug-point LLC4
+    stdio.printf("[LLC-DBG] cp: 调用 CreateProcessA\n")
+    #endregion
     ok: int = win32process.CreateProcessA(
         t.CPtr(0), cmd_line, t.CPtr(0), t.CPtr(0), 1, 0,
         t.CPtr(0), t.CPtr(0), t.CPtr(c.Addr(si)), t.CPtr(c.Addr(pi)))
     stdlib.free(cmd_line)
+    #region debug-point LLC5
+    stdio.printf("[LLC-DBG] cp: CreateProcessA 返回 ok=%d pid=%d\n", ok, pi.dwProcessId)
+    #endregion
     if ok == 0:
         return -1
     win32sync.WaitForSingleObject(pi.hProcess, w32.win32base.INFINITE)
+    #region debug-point LLC6
+    stdio.printf("[LLC-DBG] cp: 等待完成\n")
+    #endregion
     exit_code: t.CUInt32T = 0
     win32process.GetExitCodeProcess(pi.hProcess, t.CPtr(c.Addr(exit_code)))
     w32.win32base.CloseHandle(pi.hProcess)
@@ -614,7 +623,13 @@ def run_llc_parallel(tasks: bytes, task_count: int, max_threads: int) -> int:
         while j < batch_n:
             th_j: vthreading.Thread | t.CPtr = (vthreading.Thread | t.CPtr)(t.CPtr(c.Deref(
                 t.CUInt64T(t.CUInt64T(th_ptrs) + t.CSizeT(j) * 8, t.CPtr))))
+            #region debug-point LLC7
+            stdio.printf("[LLC-DBG] join 前 %d\n", j)
+            #endregion
             th_j.join(600000)
+            #region debug-point LLC8
+            stdio.printf("[LLC-DBG] join 后 %d alive=%d\n", j, th_j.is_alive())
+            #endregion
             th_j.close()
             j += 1
         # 本批结果检查