On 18 December 2018 at 15:17, Emmanuel Charpentier wrote: | Le mardi 18 décembre 2018 à 07:34 -0600, Dirk Eddelbuettel a écrit : | > On 18 December 2018 at 14:21, Emmanuel Charpentier wrote: | > > Package: r-base | > > Version: 3.5.1.20181215-1 | > > Severity: important | > > | | [ Snip... ] | | > There is nothing I can do for you here. If it is upstream in R, it is | > upstream in R. | > | > I didn't do anything R to prevent brms from working. | | This I'm aware of. But my filing of a bug was more along the lines of | "Against *wtah* should I file a bug ? ".
I know, and I _really_ appreciate it as we are before R 3.5.2. I was just about to head to work when I mailed and had to be brief. | A couple bugs against glibc *"look like"* this bug, but none is exactly | this one. Would you recommend filing a *new* bug againts the glibc | source ? | | In other words, how should I report this informatin in order to be | efficient ? Let's see if we can drill it down. You did the right thing by starting with install.packages. But based on personal experience I suspect that this will go away if we reinstall Rcpp and other build-dependencies of brms. Recall that brms also works at CRAN on Kurt's Debian system. So maybe reinstall Rcpp, rstan, rstantools, ... Maybe one by one? Maybe write down the old directory date so that we get an if it was glibc or the C++ library changing. You could well have packages built with an older g++ on your box too and those things sometimes bite... Dirk -- http://dirk.eddelbuettel.com | @eddelbuettel | e...@debian.org
