On 20/09/2025 17:40, Robert Engels wrote:
Curious, isn’t this the same issue with ANY native method call? Why
does this need special treatment - it is easily worked around by using
a worker pool in user level code if it’s a common operation.
The implementation of this method blocks in native code so pins the
carrier, it needs a re-write, too many other areas have been higher
priority to date. A thread pool would help with a small number of
virtual threads calling InetAddress::isReachable but isn't going to
scale when the typical timeout specified to the method is in seconds.
-Alan
