AshesITR
AshesITR
I think what we should do is silently ignore "invalid" glue calls. An object_usage_linter wouldn't be expected to produce "invalid glue expression" as a lint, would it? The only other...
Seems like a better fit for `unreachable_code_linter()` to me. But yes, should be linted.
The empty expression should be `{}`. NB `NULL` is equivalent and clearer in intent I think. We should allow `{}` if `allow_single_line == TRUE`.
Well, no disagreement on what we should do then :) Although `while (might_block_forever()) {}` is a code smell in and of itself 😉
(Sorry I accidentally hit edit instead of comment. Is there a way to revert?) > > We should allow {} if allow_single_line == TRUE. > Is that an "if and...
Somewhat of an extension to not only check for validity, but also for sane values.
Currently this is not possible. Note that distinguishing between the three can be hard without running the code, cf. library(R6) build_fun
In my own packages, I use CamelCase for R6 classes as well, but they aren't so many, so using dev lintr's new feature to only exclude `object_name_linter` works for me....
Crossref #1104
IINM, `touch` creates a 0-byte empty file, not a 1-byte file containing `\n`.