Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Enforce correct type-use annotation locations for nested types #1045

Merged
merged 18 commits into from
Oct 9, 2024

Conversation

armughan11
Copy link
Collaborator

Enforce correct type-use annotation locations for nested types as per JSpecify norms. We enforce type-use annotations to be on the inner class and raise an error if they are not. For annotations that are both type use and declaration, we raise an error at an invalid location.

Current behavior

// All three are ok

@Nullable A.B.C foo1 = null;
A.@Nullable B.C foo2 = null;
A.B.@Nullable C foo3 = null;

New behavior

// BUG: Diagnostic contains: Type-use nullability annotations should be applied on inner class
@Nullable A.B.C foo1 = null;
// BUG: Diagnostic contains: Type-use nullability annotations should be applied on inner class
A.@Nullable B.C foo2 = null;
A.B.@Nullable C foo3 = null;


For annotations which are both declaration and annotation and type-use, only foo2 throws an error since the location isn't apt for either scenarios

Copy link

codecov bot commented Sep 25, 2024

Codecov Report

Attention: Patch coverage is 85.48387% with 9 lines in your changes missing coverage. Please review.

Project coverage is 87.61%. Comparing base (9eea2be) to head (cddd77b).
Report is 1 commits behind head on master.

Files with missing lines Patch % Lines
...away/src/main/java/com/uber/nullaway/NullAway.java 80.95% 3 Missing and 5 partials ⚠️
...c/main/java/com/uber/nullaway/NullabilityUtil.java 94.73% 0 Missing and 1 partial ⚠️
Additional details and impacted files
@@             Coverage Diff              @@
##             master    #1045      +/-   ##
============================================
- Coverage     87.62%   87.61%   -0.02%     
- Complexity     2160     2183      +23     
============================================
  Files            85       85              
  Lines          7103     7161      +58     
  Branches       1386     1404      +18     
============================================
+ Hits           6224     6274      +50     
- Misses          450      453       +3     
- Partials        429      434       +5     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

@msridhar msridhar linked an issue Sep 26, 2024 that may be closed by this pull request
Copy link
Collaborator

@msridhar msridhar left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This is a great start! I think we are missing checking on various occurrences of annotated types, like method return types. We should be doing checking in all the places that this checker does it. Can we add that, and also tests?

Comment on lines 1480 to 1484
if (hasNestedClass(state.getTypes(), symbol.type)) {
errorMessage =
new ErrorMessage(
MessageTypes.NULLABLE_ON_WRONG_NESTED_CLASS_LEVEL,
"Type-use nullability annotations should be applied on inner class");
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I can't understand this check and why it occurs here? I think the check should occur independent of whether the initializer assignment is valid. If we need to possibly report multiple error messages from this method, we can use the state.reportMatch API to report the errors and then just always return Description.NO_MATCH.

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Changed the logic here to adhere to the ErrorProne flow.

Comment on lines 1501 to 1503
if (types == null || type == null) {
return false;
}
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

If these null checks are necessary the parameters should be @Nullable. But I don't think they are needed?

@armughan11 armughan11 force-pushed the type-use-location-nested-class branch from 61b0cdc to 213bc44 Compare October 7, 2024 10:45
Copy link
Collaborator

@msridhar msridhar left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Looking really promising! I still need to check the tests more carefully, but here are some comments on the code

|| TYPE_USE_OR_TYPE_PARAMETER.containsAll(elementTypes));
}

private void handleNullabilityOnNestedClass(
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This method should probably be renamed and it needs Javadoc. For renaming, how about:

Suggested change
private void handleNullabilityOnNestedClass(
private void checkNullableAnnotationPositionInType(

Tree typeTree,
Tree errorReportingTree,
VisitorState state) {
if (!(typeTree instanceof JCTree.JCFieldAccess)) {
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The public version of JCTree.JCFieldAccess is com.sun.source.tree.MemberSelectTree; can we just use that here?

@@ -1487,6 +1509,70 @@ public Description matchVariable(VariableTree tree, VisitorState state) {
return Description.NO_MATCH;
}

private static boolean isOnlyTypeAnnotation(Symbol anno) {
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Rename?

Suggested change
private static boolean isOnlyTypeAnnotation(Symbol anno) {
private static boolean isTypeUseAnnotation(Symbol anno) {

I think this will return true for annotations that are both type use and declaration.

nullaway/src/main/java/com/uber/nullaway/NullAway.java Outdated Show resolved Hide resolved
return true;
}
ImmutableSet<ElementType> elementTypes = ImmutableSet.copyOf(target.value());
// Return true only if annotation is not type-use only
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Maybe:

Suggested change
// Return true only if annotation is not type-use only
// Return true for any annotation that is not exclusively a type-use annotation

@@ -340,6 +343,9 @@ private static boolean isDirectTypeUseAnnotation(Attribute.TypeCompound t, Confi
// In JSpecify mode and without the LegacyAnnotationLocations flag, annotations on array
// dimensions are *not* treated as applying to the top-level type, consistent with the JSpecify
// spec.
// Outside of JSpecify mode, annotations which are *not* on the inner type are not treated as
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Do we mean "Outside of JSpecify mode", or independent of JSpecify mode? If it's true no matter what, we can just delete "Outside of JSpecify mode". Or maybe the part about outside JSpecify mode just relates to the use of the legacy annotations mode?

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

That's correct. Since we had earlier blocked using JSpecify mode and Legacy annotations together, this probably is a bit confusing. I'll remove it for the sake of it brevity

Comment on lines 383 to 388
if (!hasNestedClass(symbol.type)) {
if (innerTypeCount > 0) {
return true;
}
}
return innerTypeCount == nestingDepth - 1;
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

need a comment explaining this logic

Comment on lines 158 to 159
" // @Nullable does not apply to the inner type",
" // BUG: Diagnostic contains: @NonNull field f2 not initialized",
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Can we fix up indentation in all the tests for readability?

Copy link
Collaborator

@msridhar msridhar left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

couple more comments on the tests

" // BUG: Diagnostic contains: Type-use nullability annotations should be applied on inner class",
" @Nullable A.B foo4 = null;",
" // BUG: Diagnostic contains: assigning @Nullable expression to @NonNull field",
" A.B.@Nullable C [][] foo5 = null;",
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

If we are handling multi-dim arrays can we add a test where the field is actually @Nullable? I think you need to write A.B.C @Nullable [][]

" // BUG: Diagnostic contains: Type-use nullability annotations should be applied on inner class",
" public @Nullable A.B method4() { return null; }",
" public @Nullable A method5() { return null; }",
" public @Nullable int method6() { return 0; }",
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

remove this one? eventually this should be an error (but not for this PR). @Nullable makes no sense on a primitive type

Copy link
Collaborator

@msridhar msridhar left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

one question

* @param state The visitor state
*/
private void checkNullableAnnotationPositionInType(
List<? extends AnnotationTree> annotations, Tree type, Tree tree, VisitorState state) {
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I don't understand why we need the separate tree parameter. Why just not report errors directly on the type tree? Won't that be better?

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Fixed

Copy link
Collaborator

@msridhar msridhar left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Looks great now, thanks a lot!

@msridhar msridhar enabled auto-merge (squash) October 9, 2024 22:35
@msridhar msridhar merged commit cc5ef65 into uber:master Oct 9, 2024
9 of 11 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

Enforce correct type-use annotation locations for nested types
2 participants