RFR: 8316660: Use a MemorSegment for file mapping in FileChannelLinesSpliterator

[Alan Bateman]

On Thu, 21 Sep 2023 10:50:49 GMT, Alan Bateman <alanb at openjdk.org> wrote:

>> This PR proposes using a `MemorySegment` instead of a `ByteBuffer` in `FileChannelLinesSpliterator`.
>> 
>> The old solution deterministically unmapped the ByteBuffer if closed. If not closed, the ByteBuffer would be GC:ed.
>> 
>> This proposal mimics the old behavior but may hold the mapped memory region slightly longer.
>> 
>> Tested and passed tier1, tier2, tier3
>
> test/jdk/java/nio/file/FileChannelLinesSpliterator/Parallel.java line 45:
> 
>> 43: import static org.junit.jupiter.api.Assertions.*;
>> 44: 
>> 45: public class Parallel {
> 
> This test should needs to move to test/jdk/java/nio/file/Files and be named something like LinesParallel. The description can also say that it's a test for Files.lines.

I also changed the JBS issue to make it clear that this is about the Files.lines implementation, do you mind adjusting the PR to align with that as "FileChannelLinesSplitator" is an internal class.

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

PR Review Comment: https://git.openjdk.org/jdk/pull/15814#discussion_r1332880813