This makes zasm suitable to run from ROM.

Otherwise, we overflow into the RAM mirror, which means we
overwrite the beginning of the RAM.

Yup, that's ultimately why I've just made this whole big zasm
refactoring in the previous commits. To allow for this.

But also, zasm is in much better shape now...

This will allow interesting override scenarios, adding flexibility.

Forward reference of SHELL_MEM_PTR wasn't working (nor should it
have been...).

Instead of strings of variable length driving the iteration of the
registry, we do so through records that keep track of lengths and
counts.

Things are a bit more straightforward now.

This will allow me to make the ".org" treatment a bit less murky.

*sob*...

I'm about to split the global registry in two (labels and consts)
and the previous state of registry selection made things murky.
Now it's much better.

It allows to see much more easily where the action happens on the screen.

Also, remove zasm/test7 because it changes too much all time time
(whenever zasm changes) and isn't precise enough. Too much noise,
not worth it.

The dual scraptchpad thing doesn't work. Things become very
complicated when it's time to write that back to the file. We
overwrite our contents and end up with garbage.

This hard-binds ed to the filesystem (I liked the idea of working
only with blockdevs though...), but this is necessary for the
upcoming `w` command. We need some kind of way to tell the
destination to write to truncate itself.

This only has a meaning in the filesystem, but it's necessary to
let the file know that its registered file size has possibly
shrunk.

I thought of alternatives that would have allowed me to keep ed
blkdev-centered, but they were all too hackish to my own taste.

Hence, this new hard-bind on files.

During expression parsing, if a local label was parsed, it would
select the local registry and keep that selection, making
subsequent global labels register in the wrong place.

I'm about to break compatibility with scas. Before I do that, I
need to adjusts tests. Instead of running scas to compare results,
we commit expected result as binaries directly in the repo.

Looping inside ReadC broke the shell's hook logic. We have to
return immediately when nothing is typed.

Also, make tests run in a *BSD environment

It was mostly wrong.

Sure, it's a bit slower, but it prevents a lot of hard to debug
problems. I don't have to want to remember "don't use IX if you
have any blk* calls". Let's optimize I/O later.