Fixset around SHIFT-TAB documentation

[Jan Lahoda]

Hi Shinya,

Thanks for the report and patches. Comments inline.

On 17.4.2015 11:00, ShinyaYoshida wrote:
> Hi,
> I found some bugs(?) around SHIFT-TAB documentation.
> Please review my fixset for the bugs.
>
> Here is the fixset and tests:
> http://cr.openjdk.java.net/~shinyafox/kulla/201504171/webrev.00/repl/src/impl/SourceCodeAnalysisImpl.java.udiff.html
>
> Description of issues
> 1. REPL doesn't print method/constructor document which don't have any
> parameter:
>
> -> "".length([SHIFT-TAB]
> [no output]
>
> We can't get String#length documentation.
>
> -> "".geｔBytes([SHIFT-TAB]
> java.lang.String.getBytes(int arg0, int arg1, byte[] arg2, int arg3)
> java.lang.String.getBytes(java.lang.String arg0)
> java.lang.String.getBytes(java.nio.charset.Charset arg0)
>
> There is String#getBytes() but it is not printed
> https://docs.oracle.com/javase/8/docs/api/java/lang/String.html#getBytes

I agree, the methods with no parameters should also be included.

>
> 2. REPL doesn't treat the last argument as argument:
> -> "".getBytes(""[SHIFT-TAB]
> java.lang.String.getBytes(int arg0, int arg1, byte[] arg2, int arg3)
> java.lang.String.getBytes(java.lang.String arg0)
> java.lang.String.getBytes(java.nio.charset.Charset arg0)
>
> REPL prints all getBytes methods.
> But I think it should print only the method which the first argument is
> String in this case.

I am not sure about this one. Considering:
"".getBytes(""
there is still viable continuation of this that would lead to the use of 
the first method:
"".getBytes("".length(), ...)

So it may be more appropriate to print also the other method overloads 
that we cannot rule out reliably.

Thanks,
    Jan

>
> Regards,
> shinyafox(Shinya Yoshida)
>