Things break. Moving things break much more frequently. In space, where any error can easily be fatal, you generally want to keep things as simple as possible to complete the task at hand. As well, when things break you want them to be as easy to repair as possible. Complexity should be avoided unless absolutely necessary, or in cases where the benefit provided is tremendous. Extremely lengthy spacewalks won't happen for numerous reasons (oxygen, battery, exhaustion, radiation, and so on) so that's not really even a desirable goal.

> Extremely lengthy spacewalks won't happen for numerous reasons (oxygen, battery, exhaustion, radiation, and so on) so that's not really even a desirable goal.

You're not thinking creatively enough. When life support fails and astronauts have to live in their suits for extended periods of time, what is the difference between an extended spacewalk and having to live in your suit for a while?

Well I think there's a couple of issues here. The first is that what you're describing probably could not happen because of some technical reasons. There are generally two types of life support failures. The first is a very slow leak. The ISS, for instance, has had a number of these, and will deal with even more as the station continues to age and deteriorate. Air slowly leaks out, or perhaps your scrubber fails and CO2 levels slowly build up. In these situations you already naturally have many days to solve it.

The other is the catastrophic failure, like a major hull breach. In that case, space stations, bases, and submarines are all designed in modules. You cut off the damaged area and move on with repair. If you happen to be in that damaged area, you're dead. These suits are massive, difficult to get into or out of (requiring multiple people), and also require extensive prep/charging/loading to use. As they become more complex, this all just becomes even more true. In this catastrophic failure scenario, the astronaut would be unconscious in seconds and dead shortly thereafter. It's not like you can just hop into a suit, and heck - even if you could you again would not need to be in it for days.

The other more general issue is that you're adding complexity to try (and probably fail) to solve extremely obscure problems. So the net result is an increased chance of running into an issue. We want to be going the other direction unless there's a very good reason not to. Of course avoiding single point of failures is one of those "very good reasons" but I don't see any single point of failure that this would eliminate.

Don't spacesuits mostly use water for cooling, with a supply of ice that gradually absorbs heat (from the closed loop) and sublimates into space?

What about oxygen though