Request for Review of 7116890 (Warning Cleanup java.io)

[Stuart Marks]

On 12/1/11 2:13 PM, Stuart Marks wrote:
> On 12/1/11 12:38 PM, Alan Bateman wrote:
>> On 01/12/2011 18:35, Sebastian Sickelmann wrote:
>>> :
>>> Thanks Alan,
>>> L67-68 was a backporting (from a more complex solution to a small warning
>>> cleanup) issue. I missed the type parameters here.
>>> I changed L119-120 also.
>>>
>>> New webrev is here:
>>> http://dl.dropbox.com/u/43692695/oss-patches/openjdk8/Warning_Cleanup_Java_io/CR7116890_1/index.html
>>>
>>>
>> Looks fine to me.
>
> Hi Sebastian!
>
> The change looks fine to me too. I'll apply the patch, build it, and push it in.
>
> Thanks for your contribution!

Well, turns out this patch had an error in it. :-( In ExpiringCache.java the 
constructor that takes a long parameter has code that looks like this:

     map = new LinkedHashMap<>() { ... }

It turns out that it's illegal to use diamond in the construction of an 
anonymous class! Didn't you compile the code, Sebastian? :-) To be fair, 
neither of us reviewers caught this either. The compiler is the ultimate 
reviewer in this case.

I've taken the liberty of expanding the type parameters to

     map = new LinkedHashMap<String,Entry>() { ... }

and I've gone ahead and pushed the changeset.

Again, thanks for your contribution.

s'marks