On Sun, 2 Oct 2022 17:57:35 GMT, Markus KARG <d...@openjdk.org> wrote:
> I think the leak is actually not a problem. BAIS is a fixed size. This is different to BIS where it wraps an input stream that may be changing, e.g. write an input stream to a file that is growing and call transferTo several times then you'll see what I mean. We need to mark sure we look at this very closely, that is all I'm saying. > Regarding the `-1` check: I did that to align it with _all other_ checks of > `markpos` in the existing source code of BIS. I can undo that change, but > following your warning, shouldn't we fix _all other_ checks (independent of > this PR) from `markpos < 0` to `markpos == -1` then? Probably yes. ------------- PR: https://git.openjdk.org/jdk/pull/10525
