Ben

Inside of `fmatch()`, the integer `ncol*omf` is referenced as "`n.mc`". In the nodes table that it builds and returns, `-n.mc` is the supply associated with the node labelled "`(_Sink_)`" (usually...

Good idea!

I have an example where I passed `scores()` a svyglm w/ no missings and that had been created as follows: ```r object$call svyglm(formula = form0, design = surv.d1, family =...

Good point, I didn't think to consider the `weights = NULL` possibility. Perhaps current condition w/ something like the following tacked on: `|| is.null(object$call$weights)`. >

Wickham's R Packages recommends against use of Enhances, so I'm skipping that. An issue with my machine prevented me from testing it [issue194-survey 2fc6f52], which is all that remains to...

The issue94-survey branch is ready for merging into the master branch, I believe. @josherrickson would you like to give it a once-over and either kick it back or merge it...

I hadn't thought about the check warnings problem. How about putting survey into Enhances after all? Calling it a dependency seems less accurate; plus I'm not persuaded of the Wickham...

(I went ahead and moved the survey entry to Enhances, in master 6f526d3.)

My sense of the problem agrees with yours, Josh. Perhaps it would be better to list survey, biglm, brglm and arm under Enhances, even if some of them also have...

Ah, thanks for pointing this out. The improved communication to users that we'd get from listing those packages a second time under Enhances doesn't warrant the cost of a Note....