RFR: JDK-7122887 - JDK ignores Gnome3 proxy settings

[Chris Hegarty]

On 10/23/2013 07:53 PM, Dan Xu wrote:
> Hi,
>
> I have moved initJavaClass function into
> Java_sun_net_spi_DefaultProxySelector_init to only initialize the java
> objects once. The updated webrev can be reviewed at
> http://cr.openjdk.java.net/~dxu/7122887/webrev.01/
> <http://cr.openjdk.java.net/%7Edxu/7122887/webrev.01/>. Thanks!

Looks good to me Dan.

-Chris.

>
> -Dan
>
> On 10/23/2013 11:28 AM, Dan Xu wrote:
>> You are right, Chris. The java class objects initialization should be
>> in the init method. I will update it now. Thanks for the review!
>>
>> -Dan
>>
>>
>> On 10/23/2013 11:21 AM, Chris Hegarty wrote:
>>> Hi Dan,
>>>
>>> Firstly, thank you for working on this. This is really useful, and
>>> I'm glad to see it going into JDK 8.
>>>
>>> I grabed the patch and built it. All looks fine, and the updates you
>>> added to the manual test are really useful to verify the implementation.
>>>
>>> All in all I'm happy with this change, just a small comment...
>>>
>>> Why did you move initJavaClass to
>>> Java_sun_net_spi_DefaultProxySelector_getSystemProxy. I think it
>>> should be in Java_sun_net_spi_DefaultProxySelector_init, so it is
>>> executed just once.
>>>
>>> -Chris.
>>>
>>> On 10/22/2013 08:30 PM, Dan Xu wrote:
>>>>
>>>> Hi All,
>>>>
>>>> Please review the fix towards gnome proxy settings. In the recent Gnome
>>>> versions, it has phased out GConf and swithed to use GSettings to
>>>> handle
>>>> all the system configurations. But in JDK, we are still use GConf-based
>>>> proxy resolution, which generates wrong results in the latest linux
>>>> distributions. In this change, I added a new way to resolve proxy
>>>> correctly, which makes use of GProxyResolver to handle newwork proxy
>>>> resolution. It is based on GSettings, so every time proxy changed in
>>>> the
>>>> GSettings, it will catch that  and give the updated results. In
>>>> addition, it can also handle proxy ignore pattern correctly. For
>>>> example, it can handle not only plain domain name pattern, like
>>>> domain.com, but it can also handle more complex patterns, like
>>>> *.domain.com, IPv4/CIDR (like 127.0.0.1/8), IPv6/CIDR, etc.
>>>>
>>>> The code changes have been tested manually. Here are the bug details
>>>> and
>>>> webrev changes.
>>>>
>>>> Bug: https://bugs.openjdk.java.net/browse/JDK-7122887
>>>> webrev: http://cr.openjdk.java.net/~dxu/7122887/webrev.00/
>>>>
>>>> Thanks,
>>>>
>>>> -Dan
>>
>