[aerogear-dev] To Groovy or not to Groovy in integration tests

[Summers Pittman]

On 07/16/2013 07:09 AM, Matthias Wessendorf wrote:
> Hi Karel,
>
> thanks for starting the thread and summarizing all the 
> facts/statements from the previous discussion!
>
>
> On Tue, Jul 16, 2013 at 1:03 PM, Karel Piwko <kpiwko at redhat.com 
> <mailto:kpiwko at redhat.com>> wrote:
>
>     Hi,
>
>     let me summarize the discussion from previous threads:
>
>     What were testing requirements?
>     * Do not mock
>
>     * Cover both backend and frontend testing at the same time
>     * Control test env from tests/Maven, so it runs on both CI and
>     local machine
>       without any setup required
>     => Those 3 requirements limited us to use Arquillian
>     * Cover unified push server specifications in readable way
>
>     Why Groovy instead of Java?
>     + Better support for JSON
>     + Spock provides very nice BDD support
>     + Still supports anything Java would do
>
>     What problems we faced with Groovy?
>     - Needs specific compiler - solved, configured for tests only
>     - Needs support in IDE - Intellij - ootb, Eclipse and NetBeans have
>       plugins
>
Not really, if you are using Maven as your project format then the IDEs 
really don't have to do any work.  The latest NetBeans has not 
craptacular Groovy out of the box as well.
>
>     - Needs to be deployed in test deployment - not addressed now,
>     prolongs test
>       execution by few seconds per deployment
>
>     What are currently raised concerns?
>     - Different language for development and testing
>
Why is this a concern?
>
>     - Raises bar for newcomers willing to write tests
>
Other than ending a FooTest in .groovy or .java how does it actually 
raise the bar?
>
> that's the 'concerns' I share as well: it a little burden on getting 
> back contributions, since the source of the server is java.
>
>
> Also, what would happen if others decide let's add Ruby and also Perl 
> for some sort of tests? That would mean a language nightmare, IMO :)
We reject the PR.  Groovy is a much "easier" language than Java. Sure it 
will allow people to write some particularly hellish code, but we just 
punt those PRs.

Punting the PR is also the answer for what happens if someone adds Ruby 
or Perl.
>
>
>     Thank you for additional advantages, concerns or proving some of
>     those are not
>     valid.
>
Pros of Groovy:
* Built in examples of how to call the code from a Groovy based 
application.
* It gives us a carrot to hang in front of community members who might 
not want to write Java or who want to stretch their polyglot legs.

Cons of Groovy:
* Language is slippery.  Lots of ways to do something which could be 
quite abstract or opaque.
* If we use indy then we are committing to Java 7+
* It does require us to make some annoying trade offs w.r.t simplicty of 
our build.
* One more point of failure in a build
>
>
>     Karel
>     _______________________________________________
>     aerogear-dev mailing list
>     aerogear-dev at lists.jboss.org <mailto:aerogear-dev at lists.jboss.org>
>     https://lists.jboss.org/mailman/listinfo/aerogear-dev
>
>
>
>
> -- 
> Matthias Wessendorf
>
> blog: http://matthiaswessendorf.wordpress.com/
> sessions: http://www.slideshare.net/mwessendorf
> twitter: http://twitter.com/mwessendorf
>
>
> _______________________________________________
> aerogear-dev mailing list
> aerogear-dev at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/aerogear-dev

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.jboss.org/pipermail/aerogear-dev/attachments/20130716/1fc54516/attachment-0001.html