The goal is to avoid mixing those routines with "character devices"
(acia, vpd, kbd) which aren't block devices and have routines that
have different expectations.

This is a first step to fixing #64.

Things are a bit more straightforward now.

This huge refactoring remove the Seek and Tell routine from blockdev
implementation requirements and change GetC and PutC's API so that they
take an address to read and write (through HL/DE) at each call.

The "PTR" approach in blockdev implementation was very redundant from
device to device and it made more sense to generalize. It's possible
that future device aren't "random access", but we'll be able to add more
device types later.

Another important change in this commit is that the "blockdev handle" is
now opaque. Previously, consumers of the API would happily call routines
directly from one of the 4 offsets. We can't do that any more. This
makes the API more solid for future improvements.

This change forced me to change a lot of things in fs, but overall,
things are now simpler. No more `FS_PTR`: the "device handle" now holds
the active pointer.

Lots, lots of changes, but it also feels a lot cleaner and solid.

What used to be `tools/emul/user.h` was in fact specific to zasm, so I
moved it there.

To avoid name confusion, I renamed what used to be kernel.h and user.h
to kernel-bin.h and user-bin.h.

For now, top-level only

Up until now, invalid source input had undefined behavior. We're now
beginning to define that behavior so that zasm can be a bit more usable.

It was a bad idea to remove it. Now that I'm introducing the concept of
a per-app glue file, it becomes much easier to build emulated zasm as a
userspace app.

Made things complicated for nothing.

Make symRegister's logic pointer-based so we can break through the 0x100
limit.

To straighten out includes and to pave the way into zasm being part of
the same "include CFS" as parts, we make zasm includes namespaced.

Couldn't get rid of instrUpcode though, too complicated.

They serve no purpose and make the code less flexible.

Instead of buffering input in memory one line at a time, we go in "just
in time" mode and always read contents directly from I/O, without
buffering.

It forces us to implement a `ioPutback` scheme, but on the other hand it
greatly simplifies cases where multiple tokens are on the same line
(when a label is directly followed by an instruction).

The end result feels much more solid and less hackish.

Will need it soon.

It was previously skipped on a local label.

That is, labels with code following right afterwards.

Facilitates debugging

Gives clearer output on errors. Easier debugging...

This will soon allow use to seek and tell on input, which is necessary
for a second pass which is necessary for forward symbol references.

This require making `blkSel` a bit more flexible. Rather than having one
global selected blkdev, each app can select its own, in its own memory
space.

It always reads in the scratchpad with the same max size. No need for
DE-dancing anymore.