R-devel warnings

[josherrickson]

R 4.3.2 is coming out on Halloween. Against R-devel we see:

* checking S3 generic/method consistency ... NOTE
Mismatches for apparent methods not registered:
update:
  function(object, ...)
update.NodeInfo:
  function(old, new, ...)

predict:
  function(object, ...)
predict.CBPS:
  function(x, newdata, type, ...)
See section 'Registering S3 methods' in the 'Writing R Extensions'
manual.

[josherrickson]

@benthestatistician For the update.NodeInfo, it looks like while update.formula takes old and new arguments, stats::update takes in only the object named argument and this is becoming a NOTE in the development version.

It looks like update.NodeInfo isn't exported, nor can I see it being called from anywhere. Any objection to changing the first argument to object instead of new so it passes check? All tests & check pass with this modification.

[benthestatistician]

I have no objections, either to the proposed update.NodeInfo() modification or to the analogous modification to predict.CBPS(). Thanks for a nice tidy solution, @josherrickson.