Dirk Eddelbuettel

> I believe this is a bug in RcppEigen. Yes/no/maybe/unsure/I don't care: It goes away if you do not include `R.h` and is ultimately caused by Eigen having a function...

I am very confused as to why I now need the `inst/include/$PACKAGE_types.h` when we did not before. Two of the patches I prepared above have it _and need it_. It's...

Thumbs up on more minimal diffs. That may affect more than one package. Needing an `inst/include/*h` is probably a tell. Will revisit before we file PRs. Agreed on that adding...

It's been a while. We needed it then for finetune wrappers and forwarding, it's been like this ever since. I think I should add a warning to RcppEigen. *Edit* Issue...

I was trying to look at the other packages that are neither a) a simple case of `compileAttributes()` running to update or b) a simple case of not including `R.h`....

Nice work. I clearly wasn' thinking straight in the case of `cpr` ("of course" it is a different error class now) and `VIC5` (nice idea lowercasing). I think I will...

So we have the set of affected packages ready (in #1406) along with a set of changes for each. I am still in the midth of a multi-months upgrade for...

Perfect (and no "chef's kiss" emoji here sadly) I actually had a similar thought of 'early merging' some non-conflicting bits. That is the way. And supporting the upcoming change with...

The font business can be tricky. In the [tinythemes](https://github.com/eddelbuettel/tinythemes) I avoid dealing with "Arial Narrow" explicitly as [hrbrthemes](https://github.com/hrbrmstr/hrbrthemes) has "machinery" for it: so if it fails in the tiny variant,...

Good catch! `git blame` points at @AdamSpannbauer in 2019. Adam, do you recall if you meant to fill another example in?