SquirrelJME/SquirrelJME

# 2014/10/21

***DISCLAIMER***: _These notes are from the defunct k8 project which_
_precedes SquirrelJME. The notes for SquirrelJME start on 2016/02/26!_
_The k8 project was effectively a Java SE 8 operating system and as such_
_all of the notes are in the context of that scope. That project is no_
_longer my goal as SquirrelJME is the spiritual successor to it._

## 14:39

Busy day today. An idea I have is to split the rt jar into two bits, the other
bit being an uncompressed sup jar. The supplement JAR will have all of the
special my own stuff in there, which will be loaded by the virtual machine.
The stuff in there would be decompression for deflate support along with
generators for native code. This way, most of the stuff can remain being in
Java while the very few front end bits are in C. This would then make later
transition to pure Java for my own kernel much easier. After the C VM is
compiled I can use it to generate native kernels when needed, which are
designed to run on itself. The deflate stuff would be very minimal, but once
it is loaded and able to be ran it can then be used to decompress the runtime
JAR and start execution of standard code. Then that decompression could be
later then be used by the internal library's decompression code. In fact I
have this idea where the C virtual machine will just be a basic bootstrap
where a state is initialized then it becomes fully Java. So generated code
would either be static for kernels (or programs) or it can be linked into the
C runtime. In general I would always use recompilation and native code
execution rather than byte code stuff, there would be an initial speed penalty
before everything is cached but once that is done things will speed up
greatly.