linkenneth / furry-bear Goto Github PK

NAMES:
----------------------------------------------------------------------

Kenneth Lin
cs61a-kn

Michael Lin Yang
cs61a-ia


WORK SPLIT:
----------------------------------------------------------------------

We never explicitly determined Partner A and Partner B, nor did we
explicitly split the work for any part. We also did a lot of
cross-checking and helping each other write their parts, so we can't
really determine where the work split comes.


BUGS:
----------------------------------------------------------------------

Although our interpreter works for arbitrarily long tail recursions,
it does not work for arbitrariliy long lists of arguments. This is
because the structure of scm_read() lends itself to be recursive for
every argument. Therefore, our interpreter cannot handle more than
~900 arguments, and fewer depending on the recursion
limits. Therefore, the autograder testing will not pass with some of
the stress tests we introduced in our tests.scm file. This is not
necessarily a bug but more a flaw in design that would require major
refactoring to fix.


NOTES:
----------------------------------------------------------------------

* In our tests.scm, we introduced many stress tests that would not
  pass the autograder tests simply because of the number of arguments
  it requires that the reader must read. We have provided a
  tests_nostress.scm for running tests with higher recursion
  limits. Note that this does not mean that our interpreter does not
  handle tail recursion (see BUGS).

* For efficiency, we refactored a lot of for loops using
  self.expr.nth() to use while loops. Some functions may look
  different as a result.

* We defined three new utility functions to simply our code:
  make_single_body(), evaluate_expr_seq_and_set_expr_as_last(), and
  make_list(). The first two are defined in scheme.py, while the last
  is defined in scheme_primitives.py.

* For easier testing, we altered our scheme_test.py to treat all
  memory locations such as 0x23fa6b etc. to be interpreted to same.

* We customized our prompt a bit :)