Rupert Swarbrick

Ah, sorry for not replying more quickly. I think you'll need to tweak the commit message (there are some notes in my review) and then force-push to the branch.

(Annoyingly, the test still fails after this change, but it gets much further and it's exploding on something unrelated, so I thought this was worth doing as a first step!)

Oops: it looks like I forgot about this PR (sorry). I've just rebased and force-pushed, and will merge if everything looks good in an hour or two.

I've just re-pushed. I didn't realise, but `dv_base_reg` shouldn't be registered with the factory (it's a base class with a bunch of arguments to its constructor). I've dropped that but...

*sigh* and that applies to `dv_base_mem` too... Try again! :-)

Well, that's rather silly. It turns out that almost everything I was trying to add has either too many arguments in its constructor or is a "fake name" class that...

Force push is just a rebase: no other change.

The force-push slightly rejigs things to have a bit less duplicated code. It was surprisingly hard to get right, but I *think* everything should be wired up properly now.

Thanks @Razer6 for the very thorough review. I've addressed the first few items but there's quite a long list of missing template calls. Oops! I'll try to sort the rest...

Hi @Razer6! I think I've now responded to the items you raised. For ones that were obvious ("oh yeah, I'll do that..."), I've addressed them in the version that I've...