RFR: JDK-8235252: Repair run_tests.sh by locating jtreg locally

[Alexander Matveev]

webrev.01 is incorrect, since it uses JTREG_ prefix for variables name.

Correct webrev is http://cr.openjdk.java.net/~almatvee/8235252/webrev.02/
It is identical to webrev.01, except prefix of variable names changed to 
JT_.

Thanks,
Alexander

On 12/10/19 3:54 PM, Alexander Matveev wrote:
> http://cr.openjdk.java.net/~almatvee/8235252/webrev.01/
> - Updated script to use JT_HOME first if set, then use JT_BUNDLE_URL 
> if set.
> - Also, renamed JTREG_ to JT_ in variables names to align with rest of 
> JDK.
>
> Not sure why we did not use jtreg used by all the other tests. 
> run_tests.sh used for convenience to generate and verify packages 
> generated by jpackage. This process done in 5 stages: generating 
> packages (jtreg test), installation (requires user input), verify 
> installation, uninstall (requires user input) and verify uninstall. 
> This is why we using helper script to simplify this process.
>
> Thanks,
> Alexander
>
> On 12/10/2019 3:32 PM, Alexey Semenyuk wrote:
>> http://cr.openjdk.java.net/~almatvee/8235252/webrev.01/ looks good.
>>
>> - Alexey
>>
>> On 12/10/2019 4:49 PM, Jonathan Gibbons wrote:
>>> This seems ... surprising.
>>>
>>> Why is the instance of jtreg used for all the other tests not good 
>>> enough here?
>>>
>>> -- Jon
>>>
>>> On 12/10/19 1:35 PM, Alexander Matveev wrote:
>>>> Please review simple script fix [2] for jpackage bug [1].
>>>>
>>>>  - Modified script to download jtreg bundle from URL provided by 
>>>> environment variable JTREG_BUNDLE_URL.
>>>>
>>>> Thanks,
>>>> Alexander
>>>>
>>>> [1] https://bugs.openjdk.java.net/browse/JDK-8235252
>>>>
>>>> [2] http://cr.openjdk.java.net/~almatvee/8235252/webrev.00/
>>
>