Closed krlmlr closed 6 years ago
@krlmlr thanks for the issue- I'll try to clean this up ASAP. The issues leading up to that is I think I remember older dplyr
doesn't always allow functions of the form dplyr::n()
inside mutates and summarizes on databases (file as https://github.com/tidyverse/dplyr/issues/3500 ). And CRAN check doesn't like seemingly unbound function names such as n()
.
@krlmlr , please let me know near your CRAN release time so I can push this version to CRAN. I want to wait in case other changes are needed.
Thanks for the very quick fix. I'd like to push to CRAN as soon as possible. Usually, CRAN accepts the submission as long as we've notified all maintainers of packages we break, and it takes a few days/weeks until the update is accepted.
replyr
0.9.4
is now up on CRAN
and contains the above tested fixes.
The release candidate of dplyr 0.7.5 will use the correct implementation of
n()
only if this function isn't overridden. Most likely because of this, CRAN checks of replyr fail now: https://github.com/tidyverse/dplyr/blob/master/revdep/new-problems.md#replyrA solution seems to be to assign
dplyr::n
to the local variant ofn
, or to avoid overridingn
.Same problem as in https://github.com/randomchars42/bioset/issues/1.