Agree on a single Checkstyle suppression mechanism

[fsteeg]

Came up in #60 (comment). I seem to remember that something was not possible with the original checkstyle-disable-line when adding support for @SuppressWarnings in d70d5d3. Maybe I didn't see a way to suppress MultipleStringLiterals for the entire class? See #60 (comment) for other pro/cons of both approaches.

[blackwinter]

Copying previous discussion for better overview:

---

Suggesting switching to @SuppressWarnings in ea54a0e, as it's the standard Java suppression mechanism.

Originally posted by @fsteeg in #60 (comment)

---

But it's also limited in where it can appear! Now the check is suppressed in the whole method org.metafacture.metafix.InterpreterTest.shouldInterpretNested(); should we - unintentionally - add another "magic number", we wouldn't notice.

Also:

1. At least in our case, check violations are errors, not warnings.
2. We can't limit the number of allowed exceptions (currently two with checkstyle-disable-line).
3. The SuppressWithNearbyCommentFilter setting should be removed if we're no longer using it.

Originally posted by @blackwinter in #60 (comment)

[fsteeg]

Discussed: switch to checkstyle-disable-line comments for consistency with metafacture-core, disable MultipleStringLiterals checks for tests.

No functional review (refactoring), code review: @blackwinter / @fsteeg