get dplyr backends working

[brshallo]

pairwise() does not work with most {dplyr} backends:

• dtplyr (not working pairwise does not work with dtplyr #7 )
• dbplyr (not working: pairwise will not work with dbplyr #6 )
• sparklyr (not checked, but I think dependant on dbplyr))
• multidplyr (tentative: Evaluate if it's possible to have each fun(col_x, col_y) run in parallel #5 )

@carlomedina you mentioned tidyverse/dbplyr@84ecd65#diff-dc5c5a9f543b08c775b46de4b0befea1cd1339144c4aac39e1fac2ca043af3ceR156 however I'm wondering if this is a more general issue.

When checking the traceback() it seems to have to do with the way the masking is being done:

I did notice that the most recent version of dplyr across() changes a few things in the set-up: https://github.com/tidyverse/dplyr/blob/master/R/across.R (e.g. I'm wondering if tidyverse/dplyr@04309fe#diff-a1a1a8d205afe860a05f3209274131f0db5e8cbd85326bc264b9ac6c8cedc5cd may have made a difference).

So don't know if updating to that or some other subtle change may fix (e.g. #1 ) or may be worth asking someone...