Coverage decreased (-0.5%) to 96.849% when pulling 25d59b6 on Grafikart:feat-params into 6987497 on slimphp:3.x.

I don't get the change in coveralls % since my changes are covered by new tests :(

Coverage increased (+0.01%) to 97.321% when pulling dde1ba9 on Grafikart:feat-params into 6987497 on slimphp:3.x.

I'm on the fence about this one as I'm not sure it adds enough that it should be part of Slim or whether it should be in user-land.

Should be user-land IMHO.

This is not that easy to implement in user-land due to our current design and currently it won't break anything in core and its not doing any harm adding it in.

I would like to see this PR implement the 2nd possible change.

I can implement the 2nd possible change but do I keep the "first argument as array" ? or only the spread operator ?

This is interesting :)

If the key doesn't exist then it shouldn't be in the returned array as otherwise getParams() and getParams(['not_there']) will return different results which is inconsistent.

I disagree with @geggleto on point 2 as I think that getParams('foo', 'bar') isn't as clear that you're selecting a subset from the parameters array.

Finally, I should note that the user land implementation is a one liner:

$some = array_filter($request->getParams(), function ($k) use ($only) { return in_array($k, $only); }, ARRAY_FILTER_USE_KEY);

@akrabat that is not a one-liner ;)

I am torn on the 2nd point; I could go either way with it.

Merged after changing to not provide a default of null for '$only` keys that don't exist in the params list.

Thanks @Grafikart for your work on this.