mhebes / cs4500-a1-part2 Goto Github PK

Would it be possible to remove sorted list from the specification? We don't believe it is needed for this, and there is also no way to compare two Objects, so the idea of sorting them is somewhat nonsensical.

Please add delete map1 to the test function. It's segfaulting.

Based on the tests, it seems that the only thing that maps can own is the key and value lists. I think that's too forceful a decision to make.

The reason that I say key and value lists are the only things that can be owned by the map is that the deletes in the clone test are only freeing those two lists for map1. Other heap memory used by the map is not freed.

If a map can only own two lists, it'll be too vulnerable to the result of hash collisions, and/or not get the expected performance.

The simple fix is to simply change all map declaration to map pointer declarations in the test file. Calling delete on the pointers will free all the heap memory used by the map.

In changing the interface in #9 we probably broke the tests. Working on that :)
Related to changes I'm making for #4

Chase posted his repo for a String and Object class on piazza, so that we can all come together on the same specs for those classes. Would you all mind adopting those specs/do you have any objections? Link to repo for convenience: https://github.com/chasebish/cwc_object_string

It seems a bad idea to have these two methods when we potentially could be putting millions of key-value pairs into the map.

We will be doing a lot of copying for little value since we usually just want to find out if a key is among the keys in the map, or if a value is among the values.

According to the design, the clone method makes a shallow copy and does not copy keys and values. However, the test implies that the clone method also clones keys and values. Please clarify.

In line 43 of the tests, we are calling the get() method of List in a loop. In the specs for list, the get method returns a copy of the object. But we are never freeing these copies in the tests.

Hey guys, I think you're missing an #include in your spec.
#include <assert.h> in order to be able to use assert()

Can you please add an includes for list.h in the map.h spec for compilation? Thanks!

Hey all - wanted to ask your opinions on the spec.

Map.keys() and Map.values( ) return a List object - but List is just an alias for Array.

Since you're implementing Array, I don't see any reason you need to implement List as well.

I want to change those methods to return your Array instead of my List, but I don't know what the best way to do that is. Thoughts?

These repositories contain the APIs you should implement

1. Queue

https://github.com/apooch68/cs4500A1-part2.git

2. Map

https://github.com/DTxCode/SWDev-A1-Part2.git

3. Array

https://github.com/dorothysyu/part2.git

Hey guys, after your recent changes to the test file, should lines 379 and 384 have the opposite condition? Since d1 should now be expected to be in both keys and values?

Thanks.

If map2 is a shallow copy of map1, who has ultimate ownership of the keys and values since they need to point to the same objects per last comment by gyroknight on #4? In other words, which object is responsible for freeing them?

Map is required to return List objects when asked about its keys and values. This brings List within the scope of map specification.

According to the comment, it is required to not allocate any additional memory for potential list contents. It also requires the size of the list to be 1.

This seems mutually incompatible. Please change the initial List size to 0, or allow the allocation of some memory at construction.

For string.h and list.h we included our implementations from previous assignments. These should be removed. Doing this now, stand by.

Hi guys. Would you mind adding documentation that specifies how your map methods should behave when give a key or value == nullptr? Thanks!

Hi guys, just a small thing. I think you're missing #include <stdlib.h> in your test file. It's needed to get size_t.

Please create tests or description to specify what should be expected when we remove a key-value pair.

Since the new Object and String APIs no longer include a print function, it should probably also be removed from Map.

I can't figure out how to populate the list object, maybe the constructor could take an array or could you provide an add(...) function?

We saw that you removed the methods that List inherits from the Object class (i.e. equals and hash) so could you just add those back, since those should be stubs in any class that inherits from Object?

There are a few assertions currently in your test file that won't pass.

On line 341: the size of map1 should be equal to 2, not 3, since the third put call is replacing the pairing of 'c' to 'd' with 'c' to 'a', not adding a new entry.

On line 368: the same issue as on line 341.

On lines 373 and 374: the keys and values lists for the map should both be of size 2 because the map itself only has two entries.

Could you all remove the stub implementations included in the spec since this should only be an interface? Thanks!

Please create specifications of a getter for the char* wrapped by the String class. Otherwise, we cannot use the value protected by the class.