Request for Review: Warnings cleanup in java.lang., java.util.* (original) (raw)

Stuart Marks stuart.marks at oracle.com
Sun Dec 4 16:17:23 PST 2011

On 12/2/11 9:00 AM, Omair Majid wrote:

On 12/02/2011 07:18 AM, Alan Bateman wrote:

cc'ing core-libs-dev as that is the place to discuss these changes. I see on the sign-up sheet [1] that omajid has signed up for java.lang, maybe you are working together? Unfortunately, David and I were not working together :( Anyway, here is my webrev for java.lang.* http://cr.openjdk.java.net/~omajid/webrevs/warnings-day-2011/01/ Sorry about the delay, building and testing this took longer than I had expected it to. Also, this patch only addresses the warnings pointed out by javac. The classes java.lang.ThreadLocal and java.lang.ClassValue are not included in the webrev. The fixes for ThreadLocal were basically sprinkling @SuppressWarnings("unchecked") throughout and the fixes for ClassValue looked slightly more complex when I tried to reduce the use of rawtypes. I will post a separate webrev for those later.

I've been mulling over what to do with these two patches for the past couple days. Initially I was thinking of merging the patches and generating a new one combining the best of both. But after I looked over both of them, I felt that we should go with Omair's patch, for the following reasons. First, he had signed up to do this area already, and second, his patch stayed will within the guideline of "simple, minimal, and risk-free warnings fixes" that I had posted in kicking off the WCD [1].

The bugid for this change is 7117612. I'll kick off a separate review thread for this on core-libs-dev shortly.

Omair, regarding other changes you mentioned you might work on, please be advised that John Rose is doing some other work on ClassValue [2], and he has indicated that he's doing some warnings cleanup in there as well. You might take a look at his webrev. On ThreadLocal, it looks like there are a bunch of rawtypes warnings there too, so some judicious application of generic type parameters seems called for in those cases. Let us know what you intend to do in these areas.

See you on the other thread.

s'marks

[1] http://mail.openjdk.java.net/pipermail/jdk8-dev/2011-December/000380.html

[2] http://mail.openjdk.java.net/pipermail/core-libs-dev/2011-December/008541.html

More information about the jdk8-dev mailing list