Ben Simmons Homework 2 #9

LinuxSBC · 2024-10-04T04:52:33Z

No description provided.

pcantrell · 2024-10-04T15:07:38Z

Good work!

For the final step of this assignment:

Open up the “Files changed” tab of this pull request.
Open up my own solution to the homework so that you can compare the two:
- Solution without bonus problems
- Solution with bonus problems

Walk through the two solutions, comparing one small chunk of code at a time. Things to think about as you go:

Remember: different doesn’t necessarily mean wrong. Sometimes one way is clearly better than another, and sometimes you can just say, “Yup, that works too!”
In the java-type-checker, did you ever rely on using strings to compare types? Would your solution work if there are two different types that happen to have the same names?
Are there differences between our approaches that would result in different language behavior — different type checking or different member lookup — but the directions and tests simply didn’t specify the desired behavior enough to make one clearly the correct one? (Think about the poor language implementers who have to deal with this sort of question all the time!)
Are there Python or Java language features you notice that you’d like to investigate further?
Are there Python or Java language features that I should learn from you and investigate further? (This often happens with student work!)
Did you forget to delete any commented-out or otherwise unused code, or extraneous debug statements?

Now, to complete the assignment:

Write a brief note about one or two things that stood out to you as you compare the two solutions.

You can ask a question, or just name something notable. Post these comments here in this comment thread and/or as comments on individual lines of your code in the “Files changed” tab.

LinuxSBC · 2024-10-07T16:18:15Z

java-type-checker/java_type_checker/expressions.py

+        for arg in self.args:
+            arg.check_types()
+
+        self.receiver.static_type().method_named(self.method_name)


This line does nothing. I put it in while figuring things out and forgot to remove it.

LinuxSBC · 2024-10-07T16:18:55Z

java-type-checker/java_type_checker/expressions.py

+        self.lhs.check_types()
+        self.rhs.check_types()
+        if not self.rhs.static_type().is_subtype_of(self.lhs.static_type()):
+            raise JavaTypeMismatchError("Cannot assign " + self.rhs.static_type().name + " to variable " + self.lhs.name + " of type " + self.lhs.static_type().name)


This is way worse than the string formatting in your solution. It works, but it's messy.

LinuxSBC · 2024-10-07T16:27:32Z

java-type-checker/java_type_checker/expressions.py

+        for i in range(len(self.args)):
+            args_types.append(self.args[i].static_type().name)
+            if not self.args[i].static_type().is_subtype_of(parameter_types[i]):
+                raise_error = True


I made this weird solution so I could find the list of all type names, but I totally missed that there was a _names() function that does it for me.

LinuxSBC · 2024-10-07T16:29:13Z

java-type-checker/java_type_checker/types.py

@@ -89,6 +91,11 @@ class JavaPrimitiveType(JavaType):
    Primitive types are not object types and do not have methods.
    """

+    def is_subtype_of(self, other):
+        if other == self:


Of course this is just return other == self. I did this late at night after working for a long time on some other homework, so I wasn't thinking very clearly.

LinuxSBC · 2024-10-07T16:31:19Z

java-type-checker/java_type_checker/types.py

+            return True
+        for supertype in self.direct_supertypes:
+            if supertype.is_subtype_of(other):
+                return True


You did this on one line. I think mine is a bit more clear, but they work the same, and it could totally be argued that yours is better. However, the any() function you used is really useful.

LinuxSBC · 2024-10-07T16:33:13Z

java-type-checker/java_type_checker/types.py

    def __init__(self):
        super().__init__("void")

+    def is_subtype_of(self, other):
+        if other == self:


I'm not sure why yours didn't include this, but anyway, this is another example of return other == self being better.

Mine does include it. Maybe you were looking at the “without bonus solutions” link?

Yep, I missed that that's a bonus thing.

LinuxSBC · 2024-10-07T16:35:30Z

java-type-checker/java_type_checker/types.py

+        return False
+
+    def method_named(self, method_name):
+        raise NoSuchJavaMethod("Cannot invoke method " + method_name + "() on null")


I used "null", while you used self.name. self.name should always be null, so I don't think there's a difference. Also, the is_subtype_of(self, other) method should have done return other.is_object_type.

Always beware of repeating the same constant in multiple places in the code, even if you think the redundancy won’t hurt anything. (What if the language decides it’s called nil instead?)

LinuxSBC · 2024-10-07T16:41:27Z

python-attr-lookup/src/plang/PythonObject.java

@@ -51,7 +51,10 @@ public List<PythonObject> getMRO() {
     * result (i.e. it remembers the list buildMRO() returned and keeps returning it).
     */
    protected List<PythonObject> buildMRO() {
-        throw new UnsupportedOperationException("not implemented yet");
+        mro = new ArrayList<PythonObject>();


The assumption with this is that buildMRO() is only ever called by getMRO(), which I did not realize. As a result, I set mro in here instead of using a separate variable. It won't cause any problems, but it's a bit weird.

The side effect here does make the code a bit brittle. You’re creating a brief window where the mro var is a mutable collection, then immediately gets replaced with an immutable one — and making it possible to leave the mutable one there permanently by calling buildMRO directly from somewhere else. I agree that won’t cause any problems as the code exists now, but it does open up a window for future bugs.

LinuxSBC · 2024-10-07T16:43:10Z

python-attr-lookup/src/plang/PythonObject.java

-        throw new UnsupportedOperationException("not implemented yet");
+        mro = new ArrayList<PythonObject>();
+        mro.add(this);
+        mro.addAll(type.buildMRO());


You used getType().getMRO(), which is better because you're using the actual getters. getType() won't be any different from type at the moment, but it might change, and using buildMRO() will be slower than getMRO() and return a mutable list.

LinuxSBC · 2024-10-07T17:06:23Z

python-attr-lookup/src/plang/PythonType.java

+        mro.add(this);
+        if (base != null)
+            mro.add(base);
+        return mro;


Again, I shouldn't have used mro here. Also, yours is significantly more robust. Yours is this:

if(getBase() != null) result.addAll(getBase().getMRO()); if(getType() != null) result.addAll(getType().getMRO());

Adding the MRO of the base is obviously better than just adding the base, and it will probably break things to not add the MRO of the type. Also, using getters is generally a good idea.

LinuxSBC added 2 commits September 30, 2024 22:39

Complete part 1

65ec2a7

Complete assignment (without bonuses)

7fe1000

LinuxSBC commented Oct 7, 2024

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Ben Simmons Homework 2 #9

Ben Simmons Homework 2 #9

LinuxSBC commented Oct 4, 2024

pcantrell commented Oct 4, 2024

LinuxSBC Oct 7, 2024

LinuxSBC Oct 7, 2024

LinuxSBC Oct 7, 2024

LinuxSBC Oct 7, 2024

LinuxSBC Oct 7, 2024

LinuxSBC Oct 7, 2024

pcantrell Oct 7, 2024

LinuxSBC Oct 7, 2024

LinuxSBC Oct 7, 2024

pcantrell Oct 7, 2024

LinuxSBC Oct 7, 2024

pcantrell Oct 7, 2024

LinuxSBC Oct 7, 2024 •

edited

Loading

LinuxSBC Oct 7, 2024 •

edited

Loading

Ben Simmons Homework 2 #9

Are you sure you want to change the base?

Ben Simmons Homework 2 #9

Conversation

LinuxSBC commented Oct 4, 2024

pcantrell commented Oct 4, 2024

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

LinuxSBC Oct 7, 2024 • edited Loading

Choose a reason for hiding this comment

LinuxSBC Oct 7, 2024 • edited Loading

Choose a reason for hiding this comment

LinuxSBC Oct 7, 2024 •

edited

Loading

LinuxSBC Oct 7, 2024 •

edited

Loading