Have begin implementation of Jive, working on state 8.
Through Tuesday, 9/11 made good progress, but still
some problems.
JiveVPL.save can be copied to JiveVPL.class to
allow running my VPL interpreter (version that
takes the memory produced by Jive.java) on
a given VPL file produced by Jive.
As of Sunday, 9/16/2018, at noon, Jive is initially
implemented, testing on a few test cases
(may still have errors, of course, but is useable
for demonstrating the desired translation from
Jive to VPL)
All the tests in the folder Tests seem to be working.
To run my VPL simulator on a VPL program produced by
Jive, you need to compile IntPair.java and
then
copy VPL.save to VPL.class
(I'm hiding my VPL source code so Project 1 isn't ruined).
September 17:
in class we wrote a little Jive program, named "sept17",
that was supposed to ask the user for n and
then store in an array n followed by the first n
Fibonacci numbers (we left off the first 1, which was
fine for our purposes), and then
display that array.
When we translated the Jive program to VPL and then
ran the VPL program, the behavior was weird.
After class I found the following errors in the
Jive program (yay---not in Jive.java, my translator):
first, we forgot to put a Halt command in the
"main" at the end, so after doing whatever work
was in "main", it continued into the next function,
or something. At any rate, it was bad, and was
cured by putting in the Halt command.
second, nothing was displaying. I figured out
after a while that we had forgotten to store
the value n at offset 0 in the array. When I
put in the code:
size -> Put array 0,
the program behaved as desired, except at the end
it crashed. After some work, I realized that I
(with no one else to blame), should have done
n -> Put array 0
With that change, sept17 worked perfectly
(up to n==45---for larger n, integer
overflow produced nonsensical results at the
end the sequence).
The punchline is, we didn't find a problem with
Jive.java!
