On Wed, 14 Jan 2026 22:04:15 GMT, Kevin Rushforth <[email protected]> wrote:
> I'm not sure this is equivalent. Also, I'd like to see a test for this. In
> general, adding null checks or other similar checks without understanding
> them is not something we want to do.
To add a little more to this, replacing code like this:
try {
// do something
} catch (NullPointerException npe) {
// log or ignore or ...
}
with something like this:
if (check for null) {
// log or ignore or ...
} else {
// do something
}
is generally fine and usually easy to prove equivalence.
Unless I'm missing something, this PR doesn't do that, rather it takes a state
-- startPos == -1 -- that currently causes an OOBE (which is silently ignored)
and clamps the value so it will actually do something. That's not equivalent,
at least not in all cases.
-------------
PR Comment: https://git.openjdk.org/jfx/pull/2037#issuecomment-3752024225
