PING! - RFR 7032154: Performance tuning of sun.misc.FloatingDecimal/FormattedFloatingDecimal

[Iris Clark]

Hi, Brian.

This looks fantastic.

The changes in Formatter look fine to me.  

I also took a quick look at FormattedFloatingDecimal and am thrilled.  I think you changes have also addressed this bug:

5057835-  FormattedFloatingDecimal: massive code duplication in J2SE
http://bugs.sun.com/view_bug.do?bug_id=5057835

I think you're good to go once you verify that you successfully ran the regression tests in test/java/util/Formatter.

Thanks!
iris

-----Original Message-----
From: Martin Buchholz [mailto:martinrb at google.com] 
Sent: Wednesday, May 29, 2013 4:58 PM
To: Brian Burkhalter
Cc: core-libs-dev
Subject: Re: PING! - RFR 7032154: Performance tuning of sun.misc.FloatingDecimal/FormattedFloatingDecimal

I would like to see the real world-class experts on this scary math stuff (Alan Eliasen, Tim Buktu) be made honorary jdk reviewers, just for their area of specialization, if that was bureaucratically possible.  Shouldn't the original authors be cc'ed?

The code is awesome.  My thorough review found only these defects:

Use third person in the first sentence of a javadoc - e.g.
s/Retrieve/Retrieves/.

---

Don't use the denigrated C style
char zero[] =
Instead use
char[] zero =

---
This looks odd:

normalized as a//binary*


The original comment looked more normal:

normalized as a *binary*

---

Otherwise, approved!



On Fri, May 24, 2013 at 10:09 AM, Brian Burkhalter < brian.burkhalter at oracle.com> wrote:

> Originally posted one month ago today.
>
> http://mail.openjdk.java.net/pipermail/core-libs-dev/2013-April/016355
> .html
>
> https://www.youtube.com/watch?v=D9kv_V5lhiE
>
> Thanks,
>
> Brian
>