On Tue, 19 Aug 2025 09:25:28 GMT, Daniel Fuchs <dfu...@openjdk.org> wrote:
> I'm not sure saying "this method doesn't change the channel position" is
> appropriate - indeed it's not the method that would potentially change the
> channel position but the body publisher it returns.
That's a good point. I had overlooked the part that the method implementation
merely returns a BodyPublisher which then uses the FileChannel. So perhaps we
should say:
This method and the returned {@code BodyPublisher} do not modify the {@code
channel}'s position.
> Also - though uploading non-overlapping portions of the file is how you would
> usually use the publisher, it is not required that the portions are
> non-overlapping for concurrency to work correctly.
Right. The mention about non-overlapping was more to suggest that an entire
file can be split into smaller independent parts as part of the upload. But
yes, as you note, it's not mandatory and using overlapping portions too does
support concurrent usage of the FileChannel. That mention can be removed if it
creates confusion.
-------------
PR Comment: https://git.openjdk.org/jdk/pull/26155#issuecomment-3201249751
