Recognise that CircuitBuilder::namespace can't fail

[matthiasgoergens]

At least CircuitBuilder::namespace won't fail on its own. But it might fail, if the circus construction function (or so) you are passing to fails.

The new type reflects that.

Extracted from #466 and addressing https://github.com/scroll-tech/ceno/pull/455/files#r1818907561

This will lead to a lot of clean up down the line, but I want to keep this PR small.

[hero78119]

I have some plan to formalized our namespace to make it more organized.
Which means namespace string should meet with some format, or it will return failed
With that, please onhold for attempt to change on namespace api (and all the namespace/debug map related) first.

[matthiasgoergens]

Yes, that's why I'm holding off on #435 for now. @hero78119 This PR is literally just recognising that instead of specialising our return type to Result<X, ZKVMError>, we can use just Y and get the old behaviour by setting Y === Result<T, ZKVMError>. The implementation of fn namespace doesn't change at all, and all uses of the old API are still valid here, it's not really a change at all.

How does your plan look like?

I would suggest going for a more functional approach. If you are still interested in suggestions, I can sketch that out a bit more. But I would also be curious to see what you are cooking up.

[matthiasgoergens]

@lispc Here's a simple realisation: the namespace function as written actually works for any type T, not just for Result.

Realising that, lets us remove plenty of unnecessary OK(..) and .unwrap() later down.

It's useful to do this, because Result types should be used when we actually have something that can go wrong and should be handled; instead of being dealt with via .unwrap() thoughout. (If all the 'error handling' you ever gonna do is via .unwrap(), you might as well panic!() instead of creating an Err in the first place, at least that gives more useful stack traces.)

[matthiasgoergens]

I have some plan to formalized our namespace to make it more organized. Which means namespace string should meet with some format, or it will return failed With that, please onhold for attempt to change on namespace api (and all the namespace/debug map related) first.

Looks like the namespace reorganisation plan isn't happening. At least not anytime soon. Can we please merge this?

[matthiasgoergens]

@hero78119 Any news on the promised refactoring?