[rfc][icedtea-web] Fix for PR1198, JSObject passed incorrectly to Javascript

[Jiri Vanek]

On 01/07/2013 04:54 PM, Adam Domurad wrote:
> On 01/07/2013 10:07 AM, Jiri Vanek wrote:
>> On 01/04/2013 08:36 PM, Adam Domurad wrote:
>>> On 01/04/2013 05:23 AM, Jiri Vanek wrote:
...snip...
>>>       @Bug(id = { "PR1198" })
>>>       public void testJSObjectSetMemberIsSet() throws Exception {
>>>           ProcessResult pr = server.executeBrowser("/JSObjectFromEval.html",
>>>
>>
>>
>> I'm now ok with the patch except missing unit tests. Also I'm missing the prommised :
>> " ( unboxPrimitives&&  (type == Float.class || type == Double.class) " fixes.
>
> Missing how ?
>  From patch:
>
> +        if ( type == Float.TYPE || type == Double.TYPE ||
> +                ( unboxPrimitives && (type == Float.class || type == Double.class) )) {
>
>
>>

Ok. Sorry, overlooked.

>> So lets push after those two are solved. Thank you!
>> J.
>
> Answer to PS:
>
>>
>> One forgotten  question to above chunk:
>>
>> You told that there was one important change in tests - well I have noted this one, it does not
>> look so dramatic:) Was there something else??
>>
>> J.
>
> Sorry, it was pushed as 'cosmetic change' because it would have been painful to separate :). It was
> removal of a test assert that explicitly checked that the return type was java.lang.Integer, but
> some JS engines actually returned java.lang.Double. The fix is in HEAD now.
>
> I have also posted new iteration of unit tests.

I see.. :-/ not only money is root of all evil!
>
> Thanks for the review!
> -Adam
>