On Wed, 14 May 2025 22:19:39 GMT, Stuart Marks <sma...@openjdk.org> wrote:

>> I meant that the implementation of `subSequence` *might or might not* 
>> perform copy (so we need to tell the javaDoc reader), while `getChar` 
>> *clearly* performs a copy as part of it design. So `subSequence` can be 
>> surprising as this discussion proofs, while `getChars` will not. That was 
>> Stuart's point with `read`, and that was your point with `CharSequence`.
>
> @mkarg Please don't invoke my name to try to buttress your arguments. The 
> calls the Reader.of() instance makes on its backing CharSequence is a 
> different kind of issue from what promises or guarantees the concrete methods 
> of the Reader class makes to its subclasses.

@stuart-marks Sorry, didn't want to pull you in here, that's why I said, I am 
just *paraphrasing*.

That is correct, it is a different case, but the *final effect* for the caller 
is the same: If he needs to now whether a copy is made *or not*, he needs to 
have JavaDocs telling him *how* the new method works inside (whether it calls 
`subSequence` or whether it calls `getChars`). Otherwise he might make false 
assumptions about the result.

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

PR Review Comment: https://git.openjdk.org/jdk/pull/24728#discussion_r2090314165

Reply via email to