If this is the case, I would encourage the use of a fluent API, e.g.
*AssertJ's* assertions, to replace *Spring* core's capable Assert
<http://docs.spring.io/spring/docs/current/javadoc-api/org/springframework/util/Assert.html>
[1]
class.
Resorting back to the ugly and often brittle conditional blocks at the
beginning of methods, as in...
public <ReturnType> someMethod(Object someArgument) {
if (*isNotValid*(argument)) {
throw new IllegalArgumentException(String.format("[%s] argument is not
valid... ", argument));
}
...
}
Would be quite disappointing and make the codebase more difficult to
read/maintain. Often times, developer's "validation logic" is not even as
elegant as what I have demonstrated above, but even this small snippet is
cumbersome to repeat (defying DRY).
I also think duplicating the *Spring* Assert class in the Geode codebase
would not be a good option either.
My $0.02,
-John
[1] http://docs.spring.io/spring/docs/current/javadoc-
api/org/springframework/util/Assert.html
On Wed, Oct 5, 2016 at 4:45 PM, Jinmei Liao <[email protected]> wrote:
> Is there a initiative to get spring-core dependency out of geode-core? The
> only places we are using spring-core classes in geode-core are those Assert
> statements like: Assert.isNull, Assert.notNull etc. Should we try to get
> rid of those?
>
> --
> Cheers
>
> Jinmei
>
--
-John
503-504-8657
john.blum10101 (skype)
