[Bugfix] Fix to compile when NVSHMEM is ON (#142)

I am not sure if it is still possible to get this PR picked in 24.02. But it could not compile with the cmd `bash build.sh libwholegraph pylibwholegraph -v --allgpuarch --enable-nvshmem` on 24.02 release branch.

Authors:
  - Chang Liu (https://github.com/chang-l)
  - Brad Rees (https://github.com/BradReesWork)

Approvers:
  - Chuang Zhu (https://github.com/chuangz0)
  - Brad Rees (https://github.com/BradReesWork)

URL: #142

cpp/src/wholememory_ops/functions/nvshmem_scatter_func_impl_integer_data_int64_indices.cu

@@ -89,7 +89,7 @@ wholememory_error_code_t nvshmem_scatter_integer_int64_func(
                        embedding_entry_count_per_rank,
                        p_env_fns,
                        stream,
-                       scatter_sms;
+                       scatter_sms);
   } catch (const wholememory::cuda_error& wle) {
     WHOLEMEMORY_ERROR("scatter CUDA LOGIC Error %s\n", wle.what());
     return WHOLEMEMORY_LOGIC_ERROR;