On Fri, 20 Mar 2026 17:38:47 GMT, Andy Goryachev <[email protected]> wrote:

>> This PR fixes the bug where list changes could be wrong (incorrect index). 
>> Nearly everything was already in place, but one usecase that had a minor 
>> flaw.
>> 
>> Fix is simple, the index we use for iterating must be adjusted in case we 
>> remove a change from `addRemoveChanges`.
>> 
>> All tests are green and I also added the Tests that were in the ticket. 
>> Adjusted to JUnit 5 and with a bit of explaination which helped me a lot to 
>> better understand the issue and fix it.
>
> modules/javafx.base/src/test/java/test/javafx/collections/ListChangeBuilderTest.java
>  line 665:
> 
>> 663:         list.remove(1);
>> 664: 
>> 665:         list.doEndChange();
> 
> ... and, just for fun, perhaps check that the right content remains in the 
> list after the change?

But does it make sense? I mean, we literally `add` and `remove` the items. And 
want to test the change, so I think this will just make the test bigger.

> modules/javafx.base/src/test/java/test/javafx/collections/ListChangeBuilderTest.java
>  line 710:
> 
>> 708:         list.doEndChange();
>> 709:     }
>> 710: 
> 
> Question: do the new 4 tests provide a complete coverage as far as the bug is 
> concerned?
> 
> For example, do we need to cover the following scenarios: `add-remove-add`, 
> and `remove-add-remove`?

I checked the test coverage and the method in question is covered completely. 
The whole `ListChangeBuilder` is nearly completely covered, so I did think we 
are good here.

-------------

PR Review Comment: https://git.openjdk.org/jfx/pull/2117#discussion_r2967253047
PR Review Comment: https://git.openjdk.org/jfx/pull/2117#discussion_r2967259269

Reply via email to