axis-java-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Steve Loughran" <stev...@iseran.com>
Subject Re: checkstyle on axis
Date Sat, 31 Aug 2002 04:17:33 GMT

----- Original Message -----
From: "Tom Jordahl" <tomj@macromedia.com>
To: <axis-dev@xml.apache.org>
Sent: Friday, August 30, 2002 14:15
Subject: RE: checkstyle on axis


>
> A few observations/questions about checkstyle:
>
> - It complains about public member elements in classes, for instance there
are a ton of complains about the Holder types, which have a public "value"
member.  And we like it that way.

me too

>
> - It complains about redundant 'public' and 'abstract' modifiers in
interfaces, which confuses me a bit.  The interfaces it complains about are
a bunch of Sun JAX-RPC classes, and you would think that *they* would know
if these belonged there or not.

they are redundant on interfaces; abstract is implicit, and I think the
scope is that of the interface declaration, not methods inside it.
>
> - I like the unused import errors, along with the brace-style complains,
we should fix those (with the tool Steve mentioned if possible).

fix yes, with jalopy, maybe not. Jedit's javastyle can do that and add stub
javadocs.


> - It complains about empty catch blocks, which I might be OK but I kind of
think that in most instances this is as intended.  See
encoding.ser.BaseSeriaizerFactory.java for lots of examples.

> - The error "'&&' should be on a new line." is complaining about the form
that I consider the right thing, the && at the end of the line instead of
the beginning.  I guess I would be willing to change if that Java style
guide says different.

that's tough. at the end of line says 'there is the continuation', at the
beginning of the next is says 'here is a continuation'. Either way is better
than the _ hack of VB.

> So, my recommended plan would be to :
> - turn off the public member variable errors
> - turn off the redundant modifiers errors (at least for the javax.* files)

or fix them and tell Sun/JCP?

> - Fix the braces across the code base
> - Fix the imports across the code base
> - Possibly fix the conditionals (&& || at the beginning of the line)

> This should bring the number of errors down considerably, to something
that we can then manage.
>
> Opinions?

I dunno if we have enough control to turn off some of the reporting, but,
well, that is what XSLT is for.


Mime
View raw message