[
https://issues.apache.org/jira/browse/CRUNCH-276?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13793132#comment-13793132
]
Sean Owen commented on CRUNCH-276:
----------------------------------
OK, here's the very big patch. Spot-check some of the changes and comment on
things that look good and things that raise eyebrows. It is not hard to undo
classes of changes. Tests pass, etc. of course. There is a findbugs config now
too.
> Apply static analysis fixes, improvements
> -----------------------------------------
>
> Key: CRUNCH-276
> URL: https://issues.apache.org/jira/browse/CRUNCH-276
> Project: Crunch
> Issue Type: Task
> Affects Versions: 0.7.0
> Reporter: Sean Owen
> Priority: Minor
> Attachments: CRUNCH-276.patch
>
>
> Browsing through the Crunch code with a static analyzer, I see a number of
> minor issues that can be cleaned up, even automatically, by the same tools.
> Most are likely completely non-controversial since they do not affect the
> functionality of the code:
> - Remove unused imports
> - Fix bad javadoc
> - Typos
> - Unnecessary casts
> - Redundant modifiers on interface methods
> - Access modifiers that have no effect (public constructor in private class)
> - Missing @Override, @ Deprecated
> - Bad literals like 0d
> Others are also likely uncontroversial although might be termed a matter of
> style, although the changes would be towards standard Sun style; for example:
> - Braces around all blocks
> - No unnecessarily final on locals
> - Unnecessarily inverted conditions
> Some are signature changes, although only for entirely private elements:
> - Making methods static
> - Type weakening
> - Remove use of old classes like Hashtable
> - Declare collections by interface
> - Raw use of generic types
> And finally, there are a few which could conceivably break a caller, but only
> if doing something unintended. I want to avoid these unless the likelihood of
> an issue is very remote:
> - Making utility classes final and non-instantiable -- should not be
> subclassed or instantiated
> Before I get way into this -- it will likely touch 200+ files -- thoughts? I
> believe Josh was broadly supportive but here are the specifics.
--
This message was sent by Atlassian JIRA
(v6.1#6144)
