force unique/unseen questions in quiz retakes

This test deals with starting a quiz when there is randomisation. Both random selection of questions, and internal randomisation like with calculated questions. Random questions Create a category in the question bank containing 4 true-false questions. (Making the question texts A, B, C and D so you can easily tell them apart will make the rest of this easier.) Create a quiz that allows unlimited attempts. To that quiz add two random questions from the category you created, and make sure they are both on the same page. Preview the quiz. Click Start a new preview several times, and check you get a different set of two questions each time. (Well, it is random, so you may occasionally see the same as before.) Now attempt the quiz as a student. Note the questions you get in the first attempt (this should be a random selection). Submit that attempt, and start a second attempt. This time you should get the two question you did not see in the first attempt. Submit that attempt, and start a third attempt. This time you should get another random selection of two of the questions from all four. Internal randomisation Create a calculated question (e.g. {a} + {b}) https://docs.moodle.org/28/en/Calculated_question_type . Make all the datasets shared, and add 2 dataset items. Create a new quiz containing that calculated question. Attempt the quiz as a student. The values you get in the first attempt should be random. Finish that attempt and start a second one. This time, you should see the values you did not get the first time. Badly set up quiz. Create a quiz that tries to select 2 or more questions from the category that only contains the one question you created. Preview the quiz, and verify that you get an error. Shared datasets Create a second calculated (e.g. {a} - {b}) question in the same category as the first one. Use the same shared datasets as before. Create a new quiz containing both calculated questions on the same page. Attempt the quiz as a student. The values you get in both questions should be the same. Finish that attempt and start a second one. This time, you should see the set values you did not get the first time, but again the same values should be used in both questions. (Since this involves randomisation, ideally you should really do the all tests a few times, to verify that they are not passing just by chance the first time.)