All,
Two things have come up with respect to today's Rcpp 1.0.4 release:
1) On macOS, the most recent PR in the 1.0.4 series (to refactor exception
handling) has a side effect that appears to bite in some (but not all)
builds depending on included order. So if builds of your package act up
on macOS only, keep this in mind.
2) For the somewhat outdated R 3.3.* versions, the earliest PR in the the
Rcpp 1.0.4 series overlooked one missing definition. Builds fail there.
Both issues have already been fixed in, respectively, Rcpp 1.0.4.1 and
1.0.4.2, which are available via the Rcpp drat. But as CRAN builds may be
affected widely enough by 1) we may need to do a bug-fix follow-up.
David Vaughn, who noticed 1), ran some reverse depends test on macOS now and
may have a guestimate as to the number of affected packages.
We do a fair amount of testing, but due to technical, organisational, and,
(frankly) time limits can only cover reverse depends for one release and
architecture making 1) a little tricky to catch. For these cases, as well as
by now rarer build setups (such as the three years old R 3.3.*) we can only
rely on the community so if, come the next release, there is a chance to test
a pre-release please consider doing so.
Cheers, Dirk
--
http://dirk.eddelbuettel.com | @eddelbuettel | e...@debian.org
_______________________________________________
Rcpp-devel mailing list
Rcpp-devel@lists.r-forge.r-project.org
https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/rcpp-devel
