dplyr 1.0.8

[romainfrancois]

We're in the process of releasing dplyr 1.0.8 (which depends on dev rlang which we'll release soon as well) and timetk appears to fail against them, because of this example:

#' library(tidyverse)
#' library(timetk)
#'
#' m4_monthly %>%
#'     group_by(id) %>%
#'     tk_augment_differences(value, .lags = 1:20)

We see:

── After ─────────────────────────────────────────────────────────────────────────────────────────────────────
> checking examples ... ERROR
  Running examples in ‘timetk-Ex.R’ failed
  The error most likely occurred in:
  
  > ### Name: tk_augment_differences
  > ### Title: Add many differenced columns to the data
  > ### Aliases: tk_augment_differences
  > 
  > ### ** Examples
  > 
  > library(tidyverse)
  ── Attaching packages ─────────────────────────────────────── tidyverse 1.3.1 ──
  ✔ ggplot2 3.3.5          ✔ purrr   0.3.4     
  ✔ tibble  3.1.6          ✔ dplyr   1.0.7.9000
  ✔ tidyr   1.1.4          ✔ stringr 1.4.0     
  ✔ readr   2.1.0          ✔ forcats 0.5.1     
  ── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
  ✖ dplyr::filter() masks stats::filter()
  ✖ dplyr::lag()    masks stats::lag()
  > library(timetk)
  > 
  > m4_monthly %>%
  +     group_by(id) %>%
  +     tk_augment_differences(value, .lags = 1:20)
  Error: tk_augment_differences(.differences) is missing.
  Execution halted

> checking data for non-ASCII characters ... NOTE
    Note: found 2750 marked UTF-8 strings

1 error x | 0 warnings ✓ | 1 note x

This pull request hints in the direction of the issue. I'm not sure this error should remain, given that .differences has a default value. In that case, perhaps the example should be updated.

This might be related to this change in rlang::is_missing() r-lib/rlang#1106 . cc @lionel-

[lionel-]

In the next version of rlang, is_missing() treats default arguments as missing, consistently with base::missing().

[mdancho84]

OK, I will likely need to update more than just the example because there are many situations where this could occur. Let me take a look and update. I'll approve the change for the time being.

[romainfrancois]

Thanks @mdancho84 the pr is mostly to point in the direction :-)

[mdancho84]

@romainfrancois I appreciate it! 🙌