set_workspace should iteratively call add_workspace_stack

[DivadNojnarg]

Discussed with @nbenn.

This issues does not seems to have "bad" consequences as it is corrected once shiny takes the lead. However, I believe it does not cost much to fix this.

Currently when we call set_workspace, the list of stacks is only returned at the end. As a result, if there is any result block within a second stack, it has no idea of the first stack, which yields a wrong value at initialisation (corrected by shiny later).

Since we have add_workspace_stack available and we rarely use it, set_workspace could just iterate over the passed elements and add the stacks one by one with add_workspace_stack.

[nbenn]

@DivadNojnarg tbh, I don't think addressing this is worthwhile. You would be able to introduce a fix for a situation like

serve_workspace(
  stack1 = new_stack(new_dataset_block()),
  stack2 = new_stack(new_result_block())
)

In that you have "consistency" at the moment of instantiating the result block and not only after the first post-init update round, but this depends on the order of arguments. For a call as

serve_workspace(
  stack1 = new_stack(new_result_block()),
  stack2 = new_stack(new_dataset_block())
)

you're back to square one. In order to really make this work, we'd need to analyze dependencies, topo-sort stacks, etc.