[aerogear-dev] checkstyle on aerogear-parent

Bolesław Dawidowicz bdawidow at redhat.com
Fri May 23 04:35:09 EDT 2014


On 05/22/2014 01:50 PM, Matthias Wessendorf wrote:
>
>
> On Thursday, May 22, 2014, Karel Piwko <kpiwko at redhat.com
> <mailto:kpiwko at redhat.com>> wrote:
>
>     This is the same issue as with unit tests. If they tend to take too
>     much time, people are skipping them.
>
>     What about activating checkstyle in CI and or Git push hook, for
>     both PRs and
>     direct commits upstream. This way a developer can develop without
>     constraints
>     *until* he wants to push the code upstream. Would it help?
>
>
> -1
>
> I think we should work on proper templates for our IDEs instead


Major issue with IDE templates is that you cannot get both IDEA and 
Eclipse to reformat code in whole project into exactly same outcome. 
Unless you force people to use Eclipse formatter plugin in IDEA.

Also... checkstyle maven plugin is buggy... :)

>
>
>     Thanks,
>
>     Karel
>
>     On Wed, 21 May 2014 11:38:02 -0300
>     Bruno Oliveira <bruno at abstractj.org> wrote:
>
>      > I don't think you need to prove anything, to me is just the matter of
>      > reach a consensus and think about what works for the whole team.
>      >
>      > Is still possible skip checkstyle and leave the mess to the
>     person doing the
>      > release, even worse, the person doing the release can also skip this.
>      >
>      > The only thing able to guarantee the quality of the code is the
>     developer's
>      > conscience. I don't think that enforcing checkstyle default rules
>     will make
>      > us more conscious tbh.
>      >
>      > On 2014-05-21, Karel Piwko wrote:
>      > > That's not true, it was discussed during team meeting at
>      > > the times we've been using G+ as well (I can't prove that) and
>     on ML as
>      > > well:
>      > >
>      > >
>     http://aerogear-dev.1069024.n5.nabble.com/aerogear-dev-JavaScript-formatter-td5200.html
>      > >
>      > > It just took 7 month for the PR to be merged. The point of initial
>      > > checkstyle rules set was chosen because:
>      > >
>      > > * It was identified as the least annoying setup on ML, there
>     were a lot of
>      > > +1s
>      > > * There was a bug in released library caused by using unused
>     import that
>      > > was not available on classpath. I might be able to dig it out
>     but it would
>      > > take me a lot of time as I don't recall what project it was.
>      > >
>      > > +9001 for abstract's proposal on having same setup for IDE as
>     well. There is
>      > > nothing more annoying then something working in IDE but not in
>     Maven. -1
>      > > for the specific profile, as it shifts responsibility to clean
>     up the mess
>      > > to the person doing the release.
>      > >
>      > > Karel
>      > >
>      > > On Tue, 20 May 2014 16:20:52 -0300
>      > > Bruno Oliveira <bruno at abstractj.org> wrote:
>      > >
>      > > > The only reason to have checkstyle enabled by default is: if
>     we agree on
>      > > > which rules should be active or not and provide an specific
>     IDE setup.
>      > > >
>      > > > Other than that, people like me will skip it. Why? Simple,
>     I'm trying to
>      > > > solve critical problems and also strugging to figure out why
>     checkstyle
>      > > > is do care about method lenght.
>      > > >
>      > > > So to me if you're guys really want to introduce it we need:
>      > > >
>      > > > - Definition of which rules were supposed to be active
>      > > > - IDE profiles for Eclipse/IntelliJ
>      > > > - Make the error messages something clear
>      > > >
>      > > > Otherwise, I'm -∞. It was never discussed here and if it
>     exists on
>      > > > aerogear-parent is all our fault.
>      > > >
>      > > > On 2014-05-20, Daniel Bevenius wrote:
>      > > > > -1 I'd prefer to have checkstyle enabled by default, and
>     integrate the
>      > > > > checkstyle into the IDE to avoid having to discover issue
>     later when
>      > > > > building with maven.
>      > > > >
>      > > > >
>      > > > > On 20 May 2014 20:56, Bruno Oliveira <bruno at abstractj.org>
>     wrote:
>      > > > >
>      > > > > > +1 on having an specific profile for checkstyle
>      > > > > >
>      > > > > > On 2014-05-20, Matthias Wessendorf wrote:
>      > > > > > > Hello,
>      > > > > > >
>      > > > > > > w/ the advent of the 0.2.0 parent, we have checkstyle
>     enabled;
>      > > > > > >
>      > > > > > > However, I am not that happy, as the default rules are
>     IMO a bit odd
>      > > > > > (e.g.
>      > > > > > > the unused imports is pretty nasty when developing)
>      > > > > > >
>      > > > > > > We could:
>      > > > > > > a) get rid of it (perhaps not)
>      > > > > > > b) disable it on normal execution and only execute it
>     on a release
>      > > > > > profile
>      > > > > > > or like that
>      > > > > > >
>      > > > > > > right now I am running w/ skip - but that's a bit nasty...
>      > > > > > >
>      > > > > > > Thoughts?
>      > > > > > > Matthias
>      > > > > > >
>      > > >
>
>
>
> --
> Sent from Gmail Mobile
>
>
> _______________________________________________
> aerogear-dev mailing list
> aerogear-dev at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/aerogear-dev
>



-- 
Bolesław Dawidowicz
JBoss Portal Platform Architect | GateIn Portal Project Lead


More information about the aerogear-dev mailing list