stplanr broken with CRAN dplyr

[rsbivand]

The dplyr release to CRAN has caused havoc - I guess you'd need to condition on dplyr version.

[Robinlovelace]

Yes - I believe this is the same as #201 I've switched some of them and to this 'tidyeval' thing and rlang - I think @richardellison is up for giving the other failing ones a go. Many thanks for the heads-up - we knew this was coming, just not this soon!

[rsbivand]

It is a weakness of tidyverse that they dictate breaking changes. It would be much less destructive if they permitted slower-moving reverse dependencies to stay on an earlier stable version, as much software does. This is cathedral, not bazaar ... and it is going to harm tidyverse, people will give up on a constantly moving target.

[Robinlovelace]

If it continues to move. Hadley has said sorry for breaking stuff and I suspect things will settle down now that tidyeval is allowing people to write code that is more concise and explicit (in the potentially unlikely event that they actually understand it!).

Best tidyeval resource I've seen: https://cran.r-project.org/web/packages/dplyr/vignettes/programming.html

But yeah longer time-scales and phased approach would be good, especially in summer now many people are on holiday (from Barcelona)!

[Robinlovelace]

I think this is fixed now - 2 tests on 0.1.9 so far are passing: