Oha.
There is one essential advice, and it has nothing to do with the technology, it has to do with thinking and disciplne:
start with the lowest fault.
The system -any such system- is very complex. So, only checking if the end-product works or does not work, does not give much information, because the end-product (in this case, the KDE) depends on myriads of sub-systems, and the problem may be anywhere, and may be serious or not.
Most people follow some tutorial, build the thing up unto the end-product, check if that works as expected, and if not, they try a new and different approach.
This is wrong.
Instead, we should be watchful all the time, read the messages, check the logfiles, look for anything that might be an error or a warning, and verify if that might concern us or not. That way we will at least notice the lowest (i.e. first) error in the stack of subsystems, and then can tackle to fix that one.
This may not yet lead to success, but it makes certain that the subsystem stack is okay up to a higher watermark, and we can build on that in the next approach.
So, this is about isolating the errors to their specific subsystem, where they can be fixed, one at a time.
In Your case it means, first build up a proper unix installation that works on the command-line, before starting to approach the network, the graphics, etc. It is about slicing the task to smaller pieces that can be approached and properly tested individually.
Obviousely this is of no concern if things run well up to the end-product - but as soon as errors arise, we should switch our mind to the sliced approach, instead of worrying.
In Your case, the lowest error is that "error 0x31" from the loader. I've never seen that, and searching the internet I find
only one discussion mentioning something similar, and that one is about booting from ZFS-root (which I do not use) and having the boot-code (not) at the right place on disk.
It seems this is not harmful, as the system boots nevertheless (if I understand You right). And consequently it will certainly not be harmful to the desktop installation, and could be ignored for now.
But if I would really want to get rid of it, I would consider that after multiple installs (onto the same disk), there might be remaining artifacts from the former installations. And so, without knowing anything more, I might decide to clean my disk (a clean workplace is always a good thing) by writing zero-bytes onto it, and then go for a new approach.
Now I’m stuck writing this on my iPhone
*laugh* not so bad. Once, when my ISP had rebuilt the infrastructure, I was stuck with hopping on my bicycle, ride down to the village, where a branch bank has a free wlan active at nighttime, and there search for possible solutions.
