RFR: 8062389, 8029459, 8061950: Class.getMethod() is inconsistent with Class.getMethods() results + more

[Peter Levart]

Hi Mandy,


On 10/18/2016 04:14 PM, Peter Levart wrote:
> I have tried to capture the precise behavior in the changed javadocs 
> that I present here:
>
> http://cr.openjdk.java.net/~plevart/jdk9-dev/Class.getMethods.new/webrev.06/ 
>
>



On 11/02/2016 04:12 AM, Mandy Chung wrote:
> 1786      *        (clearly, it M's and N's declaring types are the same type, then
> 1957      *        (clearly, it M's and N's declaring types are the same type, then
>
> typo: s/it/if
>
> There are phrases such as "methods are kept” - what about “select” instead of “keep”? AFAICT, the algorithm is correct.  I feel that the spec wording could be improved a little but that can be done later.  I think this version is good for CCC submission.
>
> When you update the source, you use <OL>, <LI> capital letters - our convention uses lowercase <ol>, <li> etc.
>
> One leftover IDE-specific suppress warnings:
>   111         @SuppressWarnings("StringEquality”)
>
> The patch looks good to me.  I will do one more pass and reply to the open.
>
> Mandy
>


On 11/02/2016 04:43 AM, Mandy Chung wrote:
> A few other comments for consideration:
>    methods with same VM signature (return type, name, parameter types)
>
> Would it be better to say "methods with the signature (name and parameter types) and with same return type"?
>
>    declared public methods (including static)
>
> Alternatively, you could say:
>    declared public static and instance methods
>
>    • Include the results of invoking this algorithm recursively on all direct superinterfaces of C, excluding any static methods.
>
> An alternative:
>    All public instance methods of all direct superinterfaces of C
>
> Mandy
>

Here is a webrev incorporating your suggestions:

http://cr.openjdk.java.net/~plevart/jdk9-dev/Class.getMethods.new/webrev.07/


I think this could be used as a submission for CCC.

Regards, Peter