Is it necessary to use IO dispatcher for Firestore calls in Kotlin Coroutines?

I’m using Firebase Firestore to store some data on the cloud, and I realized that I’m using the Main (UI) dispatcher to call suspend and non-suspend functions. Can this cause me to freeze my UI?

There is a call to Firestore:

  suspend fun saveTask(task: Task) {
    tasksCollection
      .document(task.id)
      .set(task.toTaskDataMap())
      .await() // This is the suspending function.
  }

Call to the repository from the view model:

  fun onSaveTaskClicked() = viewModelScope.launch(saveTaskExceptionHandler) {
    val task: Task = /* Building the task class */

    /* Validating the task data */

    taskRepository.saveTask(task)
    
    /* Navigation logic after task is successfully saved */
  }

Should I call the saveTask() function with the IO dispatcher or not?

>Solution :

Absolutely not. The point of a suspend function (like await()) is that it doesn’t block. Dispatchers.IO is for calling blocking IO functions.

It is possible to write a suspend function that calls blocking code without using an appropriate dispatcher. This would be a blocking suspend function, but it breaks convention and is incorrect. You can trust libraries from reputable sources like Google and Square won’t expose incorrectly blocking suspend functions.