In this case, you are not counting the soup---you are counting the container which holds the soup. :)
Here, "soup" is an uncountable object---what do it mean to have one soup? Three soups? 19.21 soups? We measure soup by its container, but then we are counting containers which happen to be filled with soup---we are not counting the actual soup itself! This is the same for anything that takes a measure word (for example, how much rice vs. how many grains of rice; I have three sticks of chalk (how many, countable), not three chalks (how much, uncountable); etc.). I hope that helped!
But Kiryo---from a quick Google search, I don't think Esperanto makes this distinction, but it would be nice if someone could answer this definitively.
The 'do' is really something that belongs to English, rather than Esperanto (or any other language really).
Try looking up things on 'Do-support'. Most languages just invert the subject and verb, but English does this weird little thing with the verb 'to do' to assist in forming questions.