xxs: RFR 8081780: JShell tool: Completion contains $REPL00DOESNOTMATTER

[Jan Lahoda]

The changes to completion seem fine to me. I'd suggest to reuse the 
constant also in tests (no need for another round of review just for that).

Thanks!

Jan

On 3.6.2015 20:18, ShinyaYoshida wrote:
> Hi Robert,
> Thank you for your review.
>
> I think the use of "$REPL00DOESNOTMATTER" in Eval#trialCompile is in the
> different context between SourceCodeAnalysisImpl.
> Appearing $REPL00DOESNOTMATTER in completion is not caused by
> Eval#trialCompile but SourceCodeAnalysisImpl#wrapInClass.
>
> Should I replace "$REPL00DOESNOTMATTER" in Eval#trialCompile by
> DOESNOTMATTER even if it is in the different context?
>
> If I replace, should DOESNOTMATTER be in Util?
>
> Regards,
> shinyafox(Shinya Yoshida)
>
>
> 2015-06-04 2:08 GMT+09:00 Robert Field <robert.field at oracle.com>:
>
>> Hi Shinya,
>>
>> Not a full review, but, defining DOESNOTMATTER is good, but then it should
>> match the code that creates it, so, Eval.trialCompile() should use this
>> constant,
>>
>> Thanks,
>> Robert
>>
>>
>> On 06/03/15 09:58, ShinyaYoshida wrote:
>>
>>> Hi Jan,
>>> Please review this.
>>>
>>> webrev:
>>> http://cr.openjdk.java.net/~shinyafox/kulla/8081780/webrev.00/
>>>
>>> bugs:
>>> https://bugs.openjdk.java.net/browse/JDK-8081780
>>>
>>> Regards,
>>> shinyafox(Shinya Yoshida)
>>>
>>
>>