Thanks Rob.
Also, I am now looking at the previous changeset[1] where the regression
was introduced, once available() figures out
that a cached connection is no longer active, it should probably be
removed from the cache. (setting ret to null doesn't remove it
from the cache).
If there were other reasons for not removing it from the cache that I am
missing, that is fine - otherwise maybe we can file a separate bug for this.
Thanks,
Kurchi
[1]
http://hg.openjdk.java.net/jdk7u/jdk7u/jdk/diff/e6dc1d9bc70b/src/share/classes/sun/net/www/http/HttpClient.java
On 3/7/2013 1:19 PM, Rob McKenna wrote:
I've fleshed out the bug report a little to make that clearer, sorry
Kurchi!
Also, I'll add a testcase to this review soon.
-Rob
On 07/03/13 16:51, Kurchi Subhra Hazra wrote:
I am wondering why do you need two try-catch blocks here.
- Kurchi
On 3/7/13 8:18 AM, Rob McKenna wrote:
Hi folks,
This is a slight alteration of the fix contributed by Stuart
Douglas. This fix deals with a SocketException caused by
getSoTimeout() on a closed connection.
http://cr.openjdk.java.net/~robm/8009650/webrev.01/
-Rob
--
-Kurchi