RFR: 8214552: Resolve clash between 4947890 and 8130266

[Phil Race]

On 11/30/18 11:54 AM, Roger Riggs wrote:
> Hi Phil,
>
> The patch looks fine.

Ok, thanks.

>
> Where does it die?  SystemProps should skip null values in putIfAbsent().

You are right it checks and that explains why I didn't see it when I tested.
I saw the NPE only after I noticed the code and it seems was just from 
my debugging I
quickly added since System.out is likely not defined at this point ...

Anyway we still need this fix.

Thanks,

-phil.

>
> Sorry for the turbulence.
>
> Roger
>
>
>
> On 11/30/2018 02:43 PM, Phil Race wrote:
>> Roger,
>>
>> I just pushed 8130266 which had the CSR approved a couple of days ago 
>> only
>> to find that it clashed badly with the change you made in 4947890 to 
>> reduce
>> the cost of system property initialisation.
>> It appears that the new code there is not happy if there is no value 
>> for the
>> system property.
>>
>> I think java.awt.printerjob just needs to be removed from the new code :
>>
>> If this looks right to you I need to push ASAP :
>>
>> Bug: https://bugs.openjdk.java.net/browse/JDK-8214552
>> Webrev: http://cr.openjdk.java.net/~prr/8214552/
>>
>> Separately I think the SystemProps code ought to consider whether
>> it ought to error out the launcher if a property is not defined ..
>>
>> -phil.
>>
>>
>