Bugs in completion suggestion

[ShinyaYoshida]

Thank you for your fixing my patch & pushing, Jan!

Regards,
shinyafox(ShinyaYoshida)

2015-04-22 23:23 GMT+09:00 Jan Lahoda <jan.lahoda at oracle.com>:

> On 22.4.2015 15:06, andrei.eremeev wrote:
>
>> Great!
>>
>> Jan, could you look at the patch and push?
>>
>
> Done. Thanks a lot for the patch. I've tweaked the isNewClass a little
> bit, to avoid going through the new class path in cases like:
> new String(I.A|
>
> Jan
>
>
>> Andrei Eremeev
>>
>> On 04/22/2015 12:57 PM, ShinyaYoshida wrote:
>>
>>> Hi Andrei,
>>> I have the patch for 4:
>>> http://mail.openjdk.java.net/pipermail/kulla-dev/2015-April/000405.html
>>>
>>> If this patch is pushed, it will be resolved.
>>>
>>> Regards,
>>> shinyafox(Shinya Yoshida)
>>>
>>>
>>> 2015-04-22 18:48 GMT+09:00 andrei.eremeev <andrei.eremeev at oracle.com
>>> <mailto:andrei.eremeev at oracle.com>>:
>>>
>>>     Hi REPL team,
>>>
>>>     Bugs in completion suggestions:
>>>
>>>     Completion analysis does not work:
>>>     1. in method parameters,
>>>       - void f(Str|
>>>       - for (Str|
>>>
>>
> I'll take a look.
>
>
>>>     2. for primitive types:
>>>       - doub|
>>>
>>>     3. for Java key words (but this might be overkill):
>>>       - fo|
>>>       - whi|
>>>
>>
> Yes, completion for keywords is not implemented yet.
>
> Thanks,
>    Jan
>
>
>
>>>     4. "new" is not suggested when there is a non-static inner class
>>>     (see example).
>>>     class A {
>>>         class Inner {}
>>>     }
>>>     A a = new A();
>>>
>>>     a.| <- "new" is not suggested
>>>     a.new In| <- REPL suggests too much classes while Inner is only
>>>     applicable here.
>>>
>>>     Andrei Eremeev
>>>
>>>
>>>
>>