RFR: DIO-23: [I2C] UnavailableDeviceException when Zulu JRE is used

[Sergey Nazarkin]

Alexander, I’m waiting for patch from you. And no, this function is not for testing handle validity

Sergey Nazarkin




> On 02 Feb 2017, at 15:18, Alexander Mironenko <alexander.mironenko at oracle.com> wrote:
> 
> So it is necessary to change return type from boolean to int and make a several return codes:
> 
> {
> ...
> if(device == INVALID_DEVICE_REFERENCE) return DEVICE_CLOSED;
> if(!lock) return DEVICE_UNAVAILABLE;
> 
> return DEVICE_HAS_BEEN_LOCKED;
> }
> 
> So upper Java code can handle return code properly. 
> 
>> On Feb 2, 2017, at 3:00 PM, Sergey Nazarkin <snazarkin at azul.com> wrote:
>> 
>> Hi,
>> 
>> if obtained device is invalid, FALSE must not be returned at all since it may confuse user with improper exception (UnavailableDeviceException instead of DeviceClosedException)
>> 
>> Sergey Nazarkin
>> 
>> 
>> 
>> 
>>> On 02 Feb 2017, at 13:56, Alexander Mironenko <alexander.mironenko at oracle.com> wrote:
>>> 
>>> Hi,
>>> 
>>> Seems strange if we return true even if device_reverence is INVALID_DEVICE_REFERENCE… It is clear that this function does not lock anything, but anyway if obtained device reference is invalid, it could be locked, right?
>>> 
>>> Regards,
>>> Alex
>>> 
>>>> On Feb 2, 2017, at 1:45 PM, Sergey Nazarkin <snazarkin at azul.com> wrote:
>>>> 
>>>> Hi 
>>>> 
>>>> Please review small fix for the issue reproducible on non-Oracle JRE (Azul Zulu Embedded particularly). 
>>>> 
>>>> Issue:	https://bugs.openjdk.java.net/browse/DIO-23
>>>> Webrev: 	http://cr.openjdk.java.net/~snazarki/dio-23/
>>>> 
>>>> Sergey Nazarkin
>>>> 
>>>> 
>>>> 
>>>> 
>>