Also, fixed broken AT-XY in Grid...

Working in "blk/" folder from a modern system is harder than it
should be. Moving blocks around is a bit awkward, grepping is a
bit less convenient than it could be, git blame has troubles
following, etc.

In this commit, we modify blkpack and blkunpack to work with single
text files with blocks being separated by a special markup.

I think this will make the code significantly more convenient to
work into.

I wanted to make CVM's forth use the Grid subsystem, but doing so
would break the stage binary. Hence, this split.

NetBSD doesn't have ncurses. Linking to curses doesn't seem to change
anything. Tried on OpenBSD and Ubuntu.

The stage binary is now there.

This will soon replace the libz80 based "forth" and
"stage" executables. This is much, much faster.

All recipe makesfiles would previously assume a built emul/.

Fixes #113

Writing directly to forth.bin seems to cause problems in some
environments such as Travis CI.

Read the contents of forth.bin at runtime. This allows us to get
rid of the bin2c tooling.

Instead, embed absolute path to blkfs. Having to rebuild the stage
binary at every change in blkfs is getting tedious.

It doesn't change anything featurewise, but this change is in
preparation for the addition of an eventual AT-XY.

XPACKed source needing to be kept in bootstrap bin didn't make
much sense. With the removal of stripfc, it makes more sense to keep
the compiled form of stage 1 in repository as a bootstrap binary.

Also, rename it to forth.bin.

Big bootstrapping change which greatly reduces the role of the modern
operating system in the bootstrapping process.

This will allow us to read from blkfs in tests.

Less noisy this way

The way is clear for complete stage1 bootstrapping on the RC2014
target!

It doesn't serve much purpose and takes up many precious bytes. And,
after all, parsed didn't do it either...

This allows us to remove forth/z80a.fs. Another big step towards
self-hosting!

Also, fix some PSP leaks related to LOAD.

It's not used

The former was only used in the peculiar context of "/emul". The
regular case is actually HERE pointing to RAMEND on boot.

Let's revisit the relinking approach, with LITA this time.
This reverts commit 6652125d47.

... instead of building boot.fs, and then z80c.fs/icore.fs in separate
passes. This strengten xcomp by making it support 2 completely separate
dicts (that is, the X dict is *not* connected to the system dict. The chain
now stops at EXIT def in boot.fs).

Memory mapping is a dead end too, solution has to be at a higher
level. xcomp is my new approach.

This was a dead end.

Not quite functional yet, but works in toy situations. The goal is to
concatenate RC2014's bootstrapped dict to the boot binary.

`xxd' is not available on all systems, and on others does not support
the `-i' flag. Since bin2c.sh relied on a tool that I can't seem to find
a compatible version of, I have included a simple, portable replacement in C.

Usage remains the same:
bin2c ARRAYNAME < inputfile > outputfile.

This change is also reflected in emul/Makefile.

This should help with the bootstrapping of non-emulated environment.
For example, I have a problem with the RC2014: I can't send it
bootstrap info until the ACIA is up. I need to find a way...