[Rcpp-devel] [ANN] Rcpp Micro Releases on Rcpp Drat Repo

Dirk Eddelbuettel edd at debian.org
Wed Jan 20 15:35:32 CET 2021


Hi all,

As for the last few years, we are placing minor increments marking Rcpp
version snapshots into the drat repo at

   http://rcppcore.github.io/drat/

Following the 1.0.6 CRAN release of Rcpp, I first folded in a PR from a
branch created right before the 1.0.6 allowing packages with dots in their
names to have per-package type headers: that became Rcpp 1.0.6.1.

Then Inaki very graciously and nicely revived and completed a branch of mine
from last May finally fixing an issue Luke Tierney had raised and which
plagued users aggregating large sets of objects (for which I always used STL
containers avoiding this, but hey a bug is bug). This got fixed in Rcpp
1.0.6.2 but we did some (inadvertent) changes to the published API so Rcpp
1.0.6.3 restores this.  If you follow the GitHub repo you see when new
version are marked in DESCRIPTION; I generally upload new tarballs to the drat.

Please help Rcpp in testing these. It really helps minimize surprises once we
go to CRAN, for which we need to use some care given the now well over
two-thousand direct reverse dependencies.  The Rcpp 1.0.6.2 and 1.0.6.3
versions seem to require a rebuild of packages involving rstan or StanHeaders
and/or maybe a few others.  In reverse depedency checking I initially ran
into about 170 or so failures almost all from run-time behaviour of tests and
odd segfaults.  But having just re-confirmed in the 1.0.6 release what we C++
programmers "know" that some 'weird' run-time issues do go away sometimes
under rebuilds, I opted to reinstall several hundred packages (which I had
not done in May, thinking my branch was doomed when in fact in wasn't) and
that addressed the concernt. (Summaries of rev.dep checks are always public
too in the rcpp-logs repo in the RcppCore org.) So that really is a reminder
that we should test as widely as we can.  Please help if you can---these
microreleases are stable (they are my default uses) but broader coverage helps.

Once again, big thanks for everybody who helped and this round especially Inaki!

Cheers, Dirk

-- 
https://dirk.eddelbuettel.com | @eddelbuettel | edd at debian.org


More information about the Rcpp-devel mailing list