Gábor Csárdi

> Is there a way to document methods added dynamically with a $set()? Nope. > Is there a way to tell roxygen2 to be quieter about missing documentation for methods?...

@mikefc Please open issues for `$set()` and non-documented methods. Edit: I guess this one is for `$set()`. :)

Looking at your use of `$set()` at https://github.com/coolbutuseless/minihtml/blob/46f1cf6c9c3243fff4fb50948161625a817fc5d8/R/htag.R#L264 I guess you don't want to add documentation for each method....

Maybe a single roxygen warning would be a good compromise? Edit: I mean, eventually we'll add support for `set()`, but for your use case, a single warning seems clearly better....

Re `$set()` support, I can already see some challenges. One is that currently the R6 doc generator works at the block level, and it should really generate the output later...

> I don’t think $set() returning null should be a problem - you can just inspect components of the call to find the right class. For the simple cases, yes....

FYI, this will not make it into the next roxygen release, but we still want to implement it later.

`@rdname` is a file name, not a function name, and file names that start with a dot are special. So yeah, and `@rdname` that starts with a dot is not...

FWIW the documentation says this: > It overrides the default file name generated by roxygen and merges documentation for multiple objects into one file. But PRs to improve this are...

`R CMD check` does not care if your R6 methods are documented or not, and also does not care about the warnings that roxygen throws, they are not running roxygen....