On 24 Feb 2015, at 19:27, Seán Coffey sean.cof...@oracle.com wrote:
Thanks for suggested edits Chris. Taken them in. Shouldn't we also keep check
for possibility of (p == null) ? After all, we could be dealing with bad
proxy selector. I added checks for null and also updated test case to
This issue relates to another bug that I own : JDK-8062305
It seems to be an area that causes alot of issue (for plugin mainly) and
the suggestion from Chris in 7178362 bug report to return a direct
connection type for bad ProxySelector implementations seems appropriate.
webrev :
Sean,
Thanks for moving this forward. I think it might be possible to simplify this
by flipping the logic, like:
diff --git a/src/java.base/share/classes/java/net/SocksSocketImpl.java
b/src/java.base/share/classes/java/net/SocksSocketImpl.java
---
Thanks for suggested edits Chris. Taken them in. Shouldn't we also keep
check for possibility of (p == null) ? After all, we could be dealing
with bad proxy selector. I added checks for null and also updated test
case to test for same. It should help preserve behaviour in that area.
updated