On Wed, 29 Apr 2026 18:11:53 GMT, Sergey Bylokhov <[email protected]> wrote:
>> I don't think we've changed the meaning of size.
>> Before this PR was begun we had
>>
>> /**
>> * Constructs a byte-based {@code DataBuffer} with a single bank using
>> the
>> * specified array, size, and offset. {@code dataArray} must have at
>> least
>> * {@code offset} + {@code size} elements.
>> ...
>> */
>> public DataBufferByte(byte[] dataArray, int size, int offset){
>>
>> and
>> public int getElem(int i) {
>> return (int)(data[i+offset]) & 0xff;
>> }
>>
>> The constructor docs still say the same now and the getElem() call still
>> uses [i+offset]
>
> Then the code(and test), and the error message should be updated to follow
> the spec "0 <= index < size"
I think I see the confusion.
There's no point adding offset to both sides.
And the constructor already validated offset+size (the bank one even has a
comment on this!)
So the code was correct (and I'll revert) and the question was about the error
message.
It should have said too large for offset+size .. or just remove the mention of
offset.
I'll do the latter. Although removing it from the method that takes a bank
loses a little bit of info.
I've also added some tests that ensure we can access index=(size-1) with a
non-zero offset in place
-------------
PR Review Comment: https://git.openjdk.org/jdk/pull/29766#discussion_r3164486656
