visit_Index, visit_Slice, visit_Subscript

[JakeAlbee]

I'll split this up into things I've done.

Files Changed

• slice.py
  • I added this for testing the visits that I implemented. An incorrect form of slicing/indexing is commented out.
• Bean.py
  • I added the ability to check if an item was indexable in the NameSpaceBean which pointed to the method that actually did something in the classDefBean
• NameSpaceBean.py
  • I added the index magic method to the int class
• TestAST.py
  • No real changes, only for testing
• Exceptions.py
  • I made an exception called NonIntIndexException which handles the situation where a variable is not an integer and needs to be to index. (See note at bottom)
• Walker.py
  • I added the logic to check Indexing and Slicing.
  • Indexing and slicing both spawn from the Subscript node
  • Index requires the variable/item passed in to be an integer
  • Subscript just requires it to implement the index call.
  • I don't think they needed to return anything but that is worth looking at.

I made a NonIntIndexException because as far as I could tell only visit_Index requires the value to be an integer. visit_Slice only requires you to implement the index magic method. So it's handled two different ways since as far as I could tell, visit_Index wouldn't work even if you implement index but I couldn't find anything conclusive on that so it's worth looking into if someone else can find something.

[JakeAlbee]

Updated and should work now, I see what you were saying.

Though I'm not sure how collections work so I'll have to be briefed on that during the meeting tomorrow and I'll take a look at it tonight to see if I'll have any questions tomorrow

[daviis]

At the end of the test the object in scope that goes with 'y' gets reassigned to have the bean.name 'c'. Also 'c' in scope is a list of type int. It should only be of type int.

I have updated slice.py to have a 'pass' at the end, this way you can set a break point in the debugger on visit_pass and see the state of the scope/namespace

[daviis]

The problem with y will be in visit_assign. The problem with c is likely in visit_index? (maybe)

[JakeAlbee]

visitIndex is correctly returning a varBean without a name, , since it isn't using named things to index with for sure, I'll take a step up and lok at visit_Subscript and see if it's in there

[JakeAlbee]

And it's returning just the type integer, not that it is a list.

[JakeAlbee]

I'm running into issues with the dictionary since the walker can't access what was called before it, so it doesn't know it is a dictionary (which can indeed use a str to index) and throwing an index error. I'm not sure how to get around that.

[JakeAlbee]

As far as I can tell visit_Ellipses works fine and doesn't really do anything. I couldn't get it (or any other extended slice ideas) to not throw a syntax error.

[JakeAlbee]

This should be pretty much good to go.

The only bug I see is that I'm not matching value to key pairs in the dictionary and not handling non-homogenous value types in the dictionary. I wasn't sure how to do that. Otherwise it's working as expected.

[daviis]

I'll merge it in then check it with my changes to see what breaks and hopefully fix all of the unit tests by tomorrow.

[daviis]

I'll merge it in then check it with my changes to see what breaks and hopefully fix all of the unit tests by tomorrow.

[daviis]

I'll merge it in then check it with my changes to see what breaks and hopefully fix all of the unit tests by tomorrow.