Dplyr: Compile error in 0.7.3

[Jafet]

The 0.7.3 CRAN package fails to compile. The error is the same as reported in #3078:

> install.packages("dplyr")
[...]
join_exports.cpp: In function ‘Rcpp::DataFrame semi_join_impl(Rcpp::DataFrame, Rcpp::DataFrame, Rcpp::CharacterVector, Rcpp::CharacterVector, bool)’:
join_exports.cpp:174:21: error: ‘Rcpp::DataFrame {aka class Rcpp::DataFrame_Impl<Rcpp::PreserveStorage>}’ has no member named ‘nrow’; did you mean ‘nrows’?
   indices.reserve(x.nrow());
                     ^~~~
                     nrows
/usr/lib/R/etc/Makeconf:141: recipe for target 'join_exports.o' failed

This is just a typo, isn't it? Changing to x.nrows() makes the install succeed. Line: join_exports.cpp:174

This happened when installing dplyr over Debian's R packages. Environment:

% lsb_release -d
Description:    Debian GNU/Linux testing (buster)

% R --version
R version 3.3.3 (2017-03-06) -- "Another Canoe"
...

% g++ --version
g++ (Debian 7.2.0-4) 7.2.0
...

[krlmlr]

Thanks. What Rcpp version do you have? This should disappear when you upgrade, but we might just change it too.

[Jafet]

It is Debian's r-cran-rcpp 0.12.9-1. I see that DataFrame::nrow wasn't available until 0.12.10.

[krlmlr]

:man_facepalming: :man_facepalming:

I'll try to push a fix soon.

[Jafet]

✓

[calvinwy]

I am having the same error, seems like it haven't been fixed yet? At least not from the CRAN repository?

[krlmlr]

The GitHub version should work, the CRAN submission is underway.

[podalv]

GitHub version has still the same problem. Does nobody test stuff before release?

[lionel-]

install_github("tidyverse/dplyr#3098")

[calvinwy]

@podalv , I download it from GitHub and fixed the problem manually, and then install it locally.

[podalv]

Yeah, had to do the same, I was just being a bit frustrated that 3 days ago this issue was marked as fixed in GitHub. Oh well :)

[krlmlr]

@lionel-: Thanks, there seems to be a downside to not merging release candidates.

[calvinwy]

@krlmlr : No worries, as long as everything works. Thanks for your effort fixing the problem!