When saving `.RData`, all data is initially saved in `.RDataTmp` which
is then renamed to `.RData`.

`.RDataTmp` should be added to `.gitignore` since it will only exist
while a save operation is in progress or something failed on save.

Relevant links:
- https://stackoverflow.com/questions/32098036/purpose-of-rdatatmp-temporary-file-r
- https://github.com/wch/r-source/blob/trunk/src/library/base/R/load.R#L145-L154

https://community.rstudio.com/t/should-i-include-the-rsconnect-folder-to-gitignore/98237

Common IME for `gettext` utilities like `tools::update_pkg_po('.')` to create these temp files (at least on Mac)

ignore knitr and R markdown default cache directories in subdirectories

Exclude secret deployment key in rsconnect/, see https://shiny.rstudio.com/articles/shinyapps.html for details.

These patterns handle cache diretories and temporary files created
by the widely used [knitr](http://yihui.name/knitr/) and [R markdown](http://rmarkdown.rstudio.com/)
packages.

N.B. this ignores only the exact file `.RData` generated by default from `save.image()` and `quit()`.  It is _not_ what is referred to in 1f84661643 and http://cran.r-project.org/doc/manuals/r-release/R-exts.html#Data-in-packages, which refers only to files with an `*.RData` _extension_.

Both rules are restricted to the top level only, so `*.tar.gz` files in data directories for instance will _not_ be ignored.

Created for example when using https://github.com/jennybc/googlesheets

Details: `RData` is a valid file format for bundling data with an R
package based on the "Writing R Extensions" manual:
http://cran.r-project.org/doc/manuals/r-release/R-exts.html#Data-in-packages

.Rproj is fine - it's .Rproj.user that needs to go.

RStudio - the only real R IDE, and an incredibly commonly used environment for R, particularly since it has git support - uses .Rproj files in projects, that follow the pattern projectname.Rproj. Adding *.Rproj to capture these and exclude them.