Certainly some interesting statistics being gathered, but I think a few
rule refinements are in order:
- Illegal throws: don't throw Throwable
Probably generally good advice for end-user code and most libs, but less
so for java.lang.invoke :)
- Avoid throwing NullPointerException
While I've argued in the past that IllegalArgumentException may well be
semantically more appropriate that ship sailed long ago and it is a
firmly established convention in the core libraries that unexpected null
arguments cause NullPointerException to be thrown - it is in fact
mandated by the specs for the methods. (Of course it is better if you
get them for free by dereferencing the arg but that isn't always
applicable.)
Anyway just a brief observation.
Cheers,
David Holmes
On 22/11/2011 7:49 PM, Dalibor Topic wrote:
Since Evgeny's analysis is running on jdk7u/jdk7u, I think this will be
in particularly interesting for those of us working on JDK 7 Updates.
cheers,
dalibor topic
-------- Original Message --------
Subject: Sonar analysis of OpenJDK 7 available
Date: Tue, 22 Nov 2011 13:24:40 +0400
From: Evgeny Mandrikov<[email protected]>
To: [email protected]
Hi,
As per request of Dalibor Topic [1] I'm announcing that static analysis of
OpenJDK 7 [2] by Sonar [3] available at our public instance called Nemo [4].
Analysis is scheduled on a periodic basis once in a week.
Dedicated quality profile was not used, so there might be some
false-positive violations (like rule "Dont Import Sun"). However we are
open for collaborations and ready to create a dedicated quality profile and
I suppose that "Code Conventions" [5] might be used as a starting point.
[1] https://twitter.com/#!/robilad/status/138707382363635712
[2] http://hg.openjdk.java.net/jdk7u/jdk7u/jdk/
[3] http://www.sonarsource.org/
[4] http://nemo.sonarsource.org/dashboard/index/net.java.openjdk:jdk7
[5] http://openjdk.java.net/guide/codeConventions.html
