MiniOberon Updates

Mar 16, 4:30PM: I've made the changes to AbstractSyntaxPkg which were
  discussed in class.  I've also make some minor updates in the grammar
  (fixing the empty productions in type_list and var_list).  I also
  realized that my proc_or_forward_seq had the semi-colons in the
  wrong place.

Mar 16, 4:50PM: I realized that I also had the semi-colons in the
  wrong place for type_list, field_list, and var_list.

Mar 18, 1:30PM: TypeVarList is designed to hold entries of type
  TVEntry.  This was a superclass for VarEntry and TypeEntry, but
  these last two classes are never used anywhere in the grammar!
  Since the entries of TypeVarList are really VarList or TypeList,
  I've changed things so that TVEntry is now the superclass
  of VarList and TypeList (so either one can go into TypeVarList).
  I've also removed VarEntry and TypeEntry from the Visitor
  interfaces, since they weren't really being used.

Mar 31, 12:15PM: The Return class was defined to have one
  attribute which was an ExprList.  However, you can only
  return one value at a time, so this should be Expr.
  I updated AbstractSyntax/Result.java to reflect this.

Apr 6, 4:30PM: There was one semicolon missing in the
  module production in TinyObe_BNF.  It was inserted
  after " proc_body ID ".

Apr 6, 5:00PM: I've place a sample program using arrays
  into the pascal directory ~mikes/170/TinyObe_progs .

Apr 7, Noon: The TinyObe_progs/Orbits.to sample program
  was incorrectly using "NOT" as the logical not operator.
  It has been changed to the correct "~".

Apr 7, Noon: The tree for CallStmt didn't have a field
  to store the name of the procedure being called.
  I've added a name field of type String.

Apr 10, 9:45PM: The original grammar for TinyObe required that
  at least one variable be declared.  That was an oversight
  on my part.  To fix this I replace "VAR var_list" in the
  production rule for proc_body with "opt_vars" and added
  a new production for opt_vars. (If you never changed
  from the MiniOberon rules for type_or_var_seq, then you
  don't have to worry about this at all.)

Apr 12, 11:00PM: In the FOR loops of the sample program
  170/TinyObe_progs/Counting.to I had accidentally used
  "to" as a keyword, but it should be "TO" and has been
  corrected.

Apr 19, 10:15PM: In my codedemo directory, the MiniJava
  compiler actually produced bad code if you tried to use
  a multiply or divide operator.  I just found this out.
  So, I've replaced the code for those cases.  Basically
  the "mull" operator needed to be "imull" and the divide
  code needed to be a bit more complicated (see
  codedemo/codegen/Translator.java).

Apr 21, 12:10AM: The two of the test programs in the
  TinyObe_progs directory had (unintentional) syntax errors
  in them.  Diamond.to had two spots where I used = instead
  of := in assignments and Primes.to had div instead of DIV
  and a missing semicolon.  I've repaired both of the
  programs.  It is my intention that the programs in that
  directory have no errors, so, if you find any, please
  let me know.

Apr 21, 3:45PM: Even though we're mostly done with it, I
  corrected an error in the TinyObe grammar.  I had an
  extra VAR in the production for proc_body.

Apr 22, 1:00PM: The second loop in TinyObe_progs/Diamond.to
  should have the test left>8 , it accidentally was left<8.

~about Apr 30:  I forgot to mention that I added a few more
  test programs to the MiniOberon_progs directory sometime
  around here.

May 2, 3:30PM: I just added the e -> (e) production to
  MiniOberon.  It should have been there all along.  I believe
  you simply need to add one production to your .cup file
  in order to implement this change.

May 3, 3:00PM: In the BubbleSort.mod sample program, the
  variable 'size' was never initialized.  I've added
  'size := 10' to the bottom of the Init routine (where
  it always should have been.

May 5, 3:15PM: In the PrimeTest.mod sample program, the
  variable 'flag' was accidentally written 'FLAG' at one
  point.

May 5, 3:30PM: The sample program BubbleSort.mod also had
  all sorts of problems.  It seems to be correct now.