Dirk Eddelbuettel

Just to chime back in: the tests ran, they had no run in a bit so no gain from `ccache` and it took even longer after which I had to...

Perfect. I will let the current test run to its end (think "days" not hours, sadly) and then add whereever we're at then. As the pull request template suggests, a...

@andrjohns I seem to now see rstan failing to install. Can you corrobate locally?

No rush. This second run has to finish anyway but given it was `rstan` I thought I might as well holler...

Thanks for the fix. The second run, started after your first addition to the PR, had four packages with issues and all of them are fine with the updated PR....

No new issues in third and final reverse depends run so ready to merge. Big thank you once, this is a helpful and nice PR!

I am actually not sure I can reproduce these with `clang++-17`, at least not from building Rcpp. `uword` hints at RcppArmadillo but it could also be any one of 1100+...

I looked into this some more. One can tickle a fair amount by for example testing our embedded package `testRcppClass` via the test file `test_module_client_package.R`. And while I can subsitute...

~Hah! Following #1303 this is now different and (for the one test I looked at) much quieter. Will keep looking.~ Not so. I forgot the required setters for expanded tests....

New branch [bugfix/size_t_casts](https://github.com/RcppCore/Rcpp/tree/bugfix/size_t_casts) pushed. It behaves here. I think the key difference to what gave me minor trouble yesterday is the one-line change in Module.h.