Looks good. What matters most I think is that the common case
where there is no IPv6 address included, is efficient.
I think a test case should be possible for this also.
- Michael.
On 09/06/14 06:40, Chris Hegarty wrote:
This looks good to me Pavel.
-Chris.
On 6 Jun 2014, at 18:42, Pavel Rappo <pavel.ra...@oracle.com> wrote:
Hi everyone,
Could you please review my change for JDK-8027308?
http://cr.openjdk.java.net/~prappo/8027308/webrev.00/
The issue in question actually consists of 2 issues. The first one is about supporting
zone id syntax in URIs. Which is described in [1]. The second one is about handling
link-local addresses in HTTP headers. The current status of [1] is "Proposed
Standard". I wasn't able to use the described syntax even with the system utils
locally (OS X 10.9.3). So I think the work on the first issue can be postponed until [1]
is widely implemented.
The use case in the second issue can be easily addressed. The only thing is
that given the complexity of interaction between HttpURLConnection and
different authentication schemes the change won't fix the scenario of proxied
http connection. This is something that should be done after the refactoring of
sun.net.www.protocol.http.
-----------------------------------------
[1] http://tools.ietf.org/html/rfc6874
Thanks
-Pavel
