RFR: JDK-8246010: AdditionalLaunchersTest is not enabled, and fails.

[Alexander Matveev]

Hi Andy,

Looks good now.

Thanks,
Alexander

On 6/1/20 2:43 PM, Andy Herrick wrote:
> yes - I will fix that comment in place.
>
> /Andy
>
> On 6/1/2020 5:18 PM, alexander.matveev at oracle.com wrote:
>> Hi Andy,
>>
>> http://cr.openjdk.java.net/~herrick/8246010/webrev.03/test/jdk/tools/jpackage/share/AddLauncherTest.java.html 
>>
>>> * additionallauncherstest*.* installer. The output installer should 
>>> provide the
>> Should it be addlaunchertest*.*?
>>
>> Otherwise looks fine.
>>
>> Thanks,
>> Alexander
>>
>>
>> On 6/1/20 1:40 PM, Andy Herrick wrote:
>>> please review revised fix at [4] to issue [2].
>>>
>>> After some discussion we think existing Error messages are 
>>> sufficient and will just add the missing resource and modify test 
>>> name after enabling.
>>>
>>> [4] - http://cr.openjdk.java.net/~herrick/8246010/webrev.03/
>>>
>>> /Andy
>>>
>>> On 5/31/2020 12:13 PM, Andy Herrick wrote:
>>>> Please revice revised webrev at [3]
>>>>
>>>> I have revised to only throw FileServiceException right before 
>>>> operations that will fail.
>>>>
>>>> [3] webrev: http://cr.openjdk.java.net/~herrick/8246010/webrev.02
>>>>
>>>> /Andy
>>>>
>>>> On 5/29/2020 7:38 PM, Andy Herrick wrote:
>>>>>
>>>>> On 5/29/20 7:16 PM, Alexey Semenyuk wrote:
>>>>>> Andy,
>>>>>>
>>>>>> The check is added to WindowsAppImageBuilder.java which means it 
>>>>>> will be done for app image bundling too. However the problem is 
>>>>>> with msi bundling only as far as I understood from your explanation.
>>>>>> Did you try fix suggested at [1]?
>>>>>
>>>>> Yes - should add more comments in the bug report.
>>>>>
>>>>> The problem exists for app-image bundle as well, from within 
>>>>> ExecutableRebrander.
>>>>>
>>>>> /Z
>>>>>
>>>>>>
>>>>>> [1] 
>>>>>> https://stackoverflow.com/questions/15431210/wix-relative-path-ends-up-being-too-long
>>>>>>
>>>>>> - Alexey
>>>>>>
>>>>>> On 5/29/2020 7:12 PM, Andy Herrick wrote:
>>>>>>>
>>>>>>> On 5/29/20 7:00 PM, alexander.matveev at oracle.com wrote:
>>>>>>>> Hi Andy,
>>>>>>>>
>>>>>>>> Why we need error.lock-resource?
>>>>>>> because resource was used , but never defined.
>>>>>>>>
>>>>>>>> Also, I do not like 194 limit obtained by testing, since once 
>>>>>>>> new files are introduced in runtime this value might not work 
>>>>>>>> anymore. Can we check on actual file paths to make sure they 
>>>>>>>> have valid length?
>>>>>>>
>>>>>>> I don't like it either, but otherwise we need to test length 
>>>>>>> every time we create a path, and the check needs to be platform 
>>>>>>> dependent.
>>>>>>>
>>>>>>> /Andy
>>>>>>>
>>>>>>>>
>>>>>>>> Thanks,
>>>>>>>> Alexander
>>>>>>>>
>>>>>>>> On 5/29/20 7:36 AM, Andy Herrick wrote:
>>>>>>>>> Please review jpackage fix at [1] to issue [2]
>>>>>>>>>
>>>>>>>>> /Andy
>>>>>>>>>
>>>>>>>>> [1] http://cr.openjdk.java.net/~herrick/8246010/webrev.01/
>>>>>>>>>
>>>>>>>>> [2] https://bugs.openjdk.java.net/browse/JDK-8246010
>>>>>>>>>
>>>>>>>>
>>>>>>
>>