Romain François
Romain François
in https://github.com/tidyverse/dplyr/pull/6083 we've let `filter()` handle matrices of 1 column for the time being. At some point this should warn, and then error. Packages currently concerned: - egor: https://github.com/tilltnet/egor/issues/76 -...
closes #6100 ``` r library(dplyr, warn.conflicts = FALSE) library(purrr) attr(mtcars, "test") % attr("test") #> [1] "foo" mtcars_grouped %>% mutate(new_col = 1) %>% attr("test") #> [1] "foo" ``` Created on 2021-11-29...
related to https://github.com/tidyverse/dplyr/pull/6141 `dplyr` 1.0.8 which is about to get released will stop re-exporting `trunc_mat()` so this package will need to get it from `tibble` instead.
`add_n_miss` and `add_prop_miss` now have a `.parallel` argument, which was promoted there from the C++ side. Eventually, I'd like to put it back to the C++ side, but for now...
We're about to release dplyr 1.0.8 and this package currently fail against it, because one example uses matrices in `filter()` and we are trying to be stricter about what is...
I sometimes use `workout()` to disect several ways to do something, and would find an additional `total` row very useful. I hacked my way to this: but perhaps there's room...
Initially here: https://github.com/r-lib/testthat/pull/1544#discussion_r803969226 Maybe `style_hyperlink()` should handle url with prefix `rstudio:` when the `RSTUDIO_CLI_HYPERLINKS` env var is set, so that these links are specific to rstudio.
Related to https://github.com/rstudio/rstudio/issues/10337 it would be useful if the default value of `cpp_source(file=)` was the active file, or if we had a `cpp_source_active_file()`. This can take inspiration from `devtools::test_active_file()`: https://github.com/r-lib/devtools/blob/a999c84c03a8c955bf1bc81b2ae3f6d8e1c7110b/R/test.R#L59
Some packages (e.g. `arrow`) go back and forth between vendored and unvendored. An `cpp_unvendor()` would be nice: - unlink the `inst/include/cpp11` directory - add LinkingTo directive
closes #248