Hong Zhang wrote:
> Hi, Vince
>
>>> Thanks.
>>>
>>> Was this fix checked to be sure it covers all field types referenced
>>> in 2.4 correctly?
>>
>>
>>
>> It looks right. But that really isn't a valid answer for this
>> question. Since there isn't a devtest that caught this problem
>> earlier, and I did not see a devtest added with this code change...
>> the question of is this right is still fairly "open".
>
>
> I did some manual verification when I made the change. The "Everything
> else" in the Table 2-4 certainly doesn't make it easy to test all
> cases thoroughly ;-)
No, but testing the explicitly named items, plus one or two extra (for
the everything else category) would have caught this issue wouldn't it?
It is likely a problem that caused unlisted class A to be correctly
mapped, but unlisted class B to be incorrectly mapped would rather
esoteric, isn't it?
>
>> Without some kind of test, this issue could get reintroduced, too...
>>
>> I am surprised this wasn't caught by the CTS....
>
>
> The relevant cts tests are passing now.
Were they passing before? I can't imagine why not, but assuming they
were, that means CTS is insufficient to guarantee accuracy here. Not
surprising, but it means CTS cannot be relied on in this area. If CTS
did cover this and was not passing previously, that is more troubling.
-Peter
>
> Thanks,
>
> - Hong
>
>>
>>>
>>>
>>>> Hi, Peter
>>>> This problem is now fixed as part of the fix for issue 1121.
>>>> https://glassfish.dev.java.net/issues/show_bug.cgi?id=1121
>>>>
>>>> The fix will be available from tomorrow's nightly build.
>>>>
>>>> Thanks,
>>>>
>>>> - Hong
>>>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: dev-unsubscribe_at_glassfish.dev.java.net
>> For additional commands, e-mail: dev-help_at_glassfish.dev.java.net
>>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe_at_glassfish.dev.java.net
> For additional commands, e-mail: dev-help_at_glassfish.dev.java.net
>