Re: [Haskell-cafe] Property checks and PostgreSQL?

Yes, usually I can ignore extraneous values and this would work. But this is not a local fault. We already check that individual items can always be retrieved. It must be some relation between the items that determines whether an item will be broken. It also depends on the query — any single item can always be retrieved, but some items disappear when a set of items is requested. This is why I want to ensure that the runs of the property check are completely isolated. I am not looking for an example of a single broken item — I am looking for an example of a broken data base as a whole.