Basic for parsing/interpreting

[Jacco]

Resolved #31

Also implements iter on Dict and List
added main handling on interpreter
Added enumerate global function
Added items function on Dict
tests also check real python

[tusharsadhwani]

Thanks a lot! Most of the work seems good, especially in the Interpreter class.

The parser will probably require some changes, as I think we should reuse parse_expression() for the parsing of the targets here.

Consider this example from Python's AST:

$ python -m ast <<<'for i in x in y in z: ...'
Module(
   body=[
      For(
         target=Name(id='i', ctx=Store()),
         iter=Compare(
            left=Name(id='x', ctx=Load()),
            ops=[
               In(),
               In()],
            comparators=[
               Name(id='y', ctx=Load()),
               Name(id='z', ctx=Load())]),
         body=[
            Expr(
               value=Constant(value=Ellipsis))],
         orelse=[])],
   type_ignores=[])

This seems to show that we can stop at the first in statement for the target, right?

Well, unfortunately that's not the case, for example, if there are commas after that sub-expression:

$ python -m ast <<<'for i in x, j in y in z: ...'
Module(
   body=[
      For(
         target=Name(id='i', ctx=Store()),
         iter=Tuple(
            elts=[
               Name(id='x', ctx=Load()),
               Compare(
                  left=Name(id='j', ctx=Load()),
                  ops=[
                     In(),
                     In()],
                  comparators=[
                     Name(id='y', ctx=Load()),
                     Name(id='z', ctx=Load())])],
            ctx=Load()),
         body=[
            Expr(
               value=Constant(value=Ellipsis))],
         orelse=[])],
   type_ignores=[])

Here, i in x, j became the target and y, z became the iterable.

Doing this correctly will require some more looking into how Python's actual grammar works.

[Jacco]

I did experiment with multiple expressions separated by comma after the in. I think that already works.

The extra ins are comparisson operators or am I missing something :-)

In both case the target is target=Name(id='i', ctx=Store()),

Or do you mean the output of my code :-)

[tusharsadhwani]

You're right, I confused myself.

Making some small changes to the code...

[Jacco]

I think I also forgot to add the 'range' function

[tusharsadhwani]

I made some adjustments:

• Since we don't have a Generator type yet, Items and Enumerate will return a List instead. Unfortunate but it is how it is for now.
• The parser was mostly good! I just made sure that iterable and target were always a Node, instead of a list of Nodes
• Added a parser test to ensure we parse for loops properly.

[tusharsadhwani]

Please feel free to add a range class as well.

[Jacco]

• The parser was mostly good! I just made sure that iterable and target were always a Node, instead of a list of Nodes

I think they can have multiple. There are examples in the tests:

This one has two targets

for k,w in dct.item():
    print(k,w)

[tusharsadhwani]

This one has two targets

for k,w in dct.item():
    print(k,w)

@Jacco In this case target will be a Tuple, check parser_test.py changes.

[Jacco]

Ah ok, I missed the inner workings self.assign :-)

[tusharsadhwani]

I'll merge this PR soon, you can add a range function in a separate one.