On Fri, 6 Mar 2026 11:33:52 GMT, Raffaello Giulietti <[email protected]> 
wrote:

>> test/jdk/java/util/Arrays/largeMemory/ParallelPrefix.java line 160:
>> 
>>> 158:         int[] parallelRangeResult = Arrays.copyOfRange(data, 
>>> fromIndex, toIndex);
>>> 159:         Arrays.parallelPrefix(parallelRangeResult, op);
>>> 160:         assertArraysEqual(parallelRangeResult, 
>>> Arrays.copyOfRange(sequentialResult, fromIndex, toIndex));
>> 
>> With JUnit, the first parameter to assertArraysEquals is the "expected" as 
>> migration from TestNG will usually means transposing these parameters. It's 
>> really only an issue if there is failure of course, and only leads to a 
>> confusing message.
>
> The `assertArraysEqual()` in this class are _not_ JUnit methods. It's their 
> implementations that eventually invoke the real JUnit method, with `expected` 
> and `actual` in the conventional JUnit order.

I get that, I'm just saying its a hazard to have these wrappers follow the 
TestNG ordering and then transpose to use the JUnit ordering.

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

PR Review Comment: https://git.openjdk.org/jdk/pull/30111#discussion_r2895311943

Reply via email to