Mar 23, 2009 12:56 AM|Paul Linton|LINK
I have been mainly interested in the way Rob approached having testable data access. The "is it a repository according to Martin Fowler" question never bothered me. I justed wanted to be able to test as much of the code as possible without hitting an actual
database. I have really only spent 10 minutes looking at this latest code but it is very obviously quite different in the data access area (that looks like SQL strings in there, how quaintly 2007[:)] )
I put a fair bit of effort into pausing the last few videos at strategic points so I could write down the code that scrolled past (took a while to find a pen, but I borrowed one from my kids). From what I saw I put together a fairly satisfactory repository
style setup. I ended with a single repository for each aggregate which made use of an ISession implementation that I could swap using StructureMap for 'live' or 'testing' purposes. It was certainly not ideal as I found several times that things tested OK
against my InMemorySession but that the SQLSession gagged (LinqToSql puts some tough restrictions on what you can do). But over the last few weeks I have found ways to work around the scenarios that caused problems and it is now full speed ahead developing
In all likelihood I may not look too closely at this latest incarnation of the StoreFront as it seems to be too different in the areas that I have put a lot of effort in to. You are 100% right that some new screencasts will be needed to explain what is
going on in this latest code drop and why it was decided on.