On Tue, 30 Jun 2026 02:30:36 GMT, Chen Liang <[email protected]> wrote:

>> As noted in https://bugs.openjdk.org/browse/JDK-8386802, `Util.entryList` 
>> and `Util.moduleEntryList` copy their input lists using indexed access. This 
>> is fine for `RandomAccess` lists but can be inefficient for 
>> sequentially-accessed implementations.
>> 
>> This patch keeps the existing indexed path for `RandomAccess` lists and 
>> falls back to iteration otherwise. Regression coverage is added for 
>> non-`RandomAccess` inputs.
>> 
>> Testing:
>> - `make CONF_CHECK=ignore java.base-java images`
>> - `make CONF_CHECK=ignore test-only 
>> TEST="test/jdk/jdk/classfile/UtilTest.java"`
>> - `make CONF_CHECK=ignore test-only TEST="test/jdk/jdk/classfile"`
>> - `make CONF_CHECK=ignore test-only TEST="test/jdk:tier1" 
>> JTREG="VERBOSE=summary"`
>> 
>> ---------
>> - [x] I confirm that I make this contribution in accordance with the 
>> [OpenJDK Interim AI Policy](https://openjdk.org/legal/ai).
>
> Let's just keep the iterator loop and remove the `get` loop for clarity.

@liach
> the iterator can throw NSEE if the list shrinks unexpectedly

That likely won’t happen, as the `Iterator` traversal ends when `hasNext()` 
returns `false`.
The fix for that case is to check that `i == result.length` after the end of 
the loop, and throw an exception[^1] if not.

[^1]: [`ConcurrentModificationException`] is probably appropriate here

[`ConcurrentModificationException`]: 
https://docs.oracle.com/en/java/javase/26/docs/api/java.base/java/util/ConcurrentModificationException.html

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

PR Comment: https://git.openjdk.org/jdk/pull/31719#issuecomment-4845126649

Reply via email to