OK, thanks.
I don't have a problem with this changes to get this working, but we
should probably review any tests that are trying to connect to 0.0.0.0.
-Chris.
On 02/17/12 02:41 PM, Michael McMahon wrote:
Chris,
On 17/02/12 13:45, Chris Hegarty wrote:
Would it not make sense to fix the individual tests that are trying to
connect to 0.0.0.0, so they connect to the appropriate address/hostname?
We may end up doing that anyway. But the purpose of this change was to
deal with the situation where someone opens a socket to '0.0.0.0' and a
proxy
is enabled.
The tests you changed use 127.0.0.1 rather than the localhost. Does
InetAddress.getLocalHost().getHostAddress() not return an appropriate
address? Or is it being proxied?
It returns a "normal" IP address, which ends up being proxied by default
on Mac.
- Michael.
-Chris.
On 02/17/12 12:26 PM, Michael McMahon wrote:
Could I get the following webrev reviewed please for 7u4?
http://cr.openjdk.java.net/~michaelm/7146564/webrev.1/
Something I overlooked in the recent change to the DefaultProxySelector
is that some of our tests end up trying to connect to 0.0.0.0
and that needs to be added to the default proxy list.
There are a couple of other test failures related to this
which the change fixes.
Thanks
Michael
