Drop caller_env from DataMask
#6444
Conversation
|
I only skimmed through, but that looks alright, esp. if the tests agree. |
|
I'll finish this off after #6438 to avoid merge conflicts. |
| @@ -277,7 +277,7 @@ slice_rows <- function(.data, ..., caller_env, error_call = caller_env()) { | |||
| if (is_empty(dots)) { | |||
There was a problem hiding this comment.
Can caller_env be removed from slice_rows()?
R/summarise.R
Outdated
| @@ -205,7 +205,7 @@ summarise.rowwise_df <- function(.data, ..., .groups = NULL) { | |||
| summarise_cols <- function(.data, dots, caller_env, error_call = caller_env()) { | |||
There was a problem hiding this comment.
Can caller_env be removed from summarise_cols()?
|
Removing |
|
My revdep tracker issue also tells me that cubble is already broken from the |
| new_groups <- c(new_groups, compat_lazy_dots(.dots, env = caller_env)) | ||
| new_groups <- c(new_groups, compat_lazy_dots(.dots, env = caller_env(2))) |
There was a problem hiding this comment.
So I guess it still technically had meaning here, but we are ok with removing caller_env as an input arg because this is deprecated behavior?
There was a problem hiding this comment.
And the argument was particularly troublesome because (a) it lacked a leading . and (b) it had the same name as a function we called.
@romainfrancois @lionel- Does this look ok? As far as I can tell the only place this was used was when evaluating the glue spec for
across(), and I'm pretty sure using the environment of theacross()quosure is just as good (if not better).(This is in preparation for #6442)
I still need to remove the other various argument that this makes redundant, but I wanted to double check that I wasn't missing anything obvious before I started on that.