possible segmentation fault

[spedygiorgio]

Hi, on my travis ci travis an error is raised after bindrcpp is installed...

Installing bindrcpp
 *** caught segfault ***
address 0xfffffff9, cause 'memory not mapped'
An irrecoverable exception occurred. R is aborting now ...
/home/travis/.travis/job_stages: line 54: 13342 Segmentation fault      (core dumped) Rscript -e 'deps <- devtools::dev_package_deps(dependencies = NA);devtools::install_deps(dependencies = TRUE);if (!all(deps$package %in% installed.packages())) { message("missing: ", paste(setdiff(deps$package, installed.packages()), collapse=", ")); q(status = 1, save = "no")}'
The command "Rscript -e 'deps <- devtools::dev_package_deps(dependencies = NA);devtools::install_deps(dependencies = TRUE);if (!all(deps$package %in% installed.packages())) { message("missing: ", paste(setdiff(deps$package, installed.packages()), collapse=", ")); q(status = 1, save = "no")}'" failed and exited with 139 during .
Your build has been stopped.

Is it possible a problem arise from your package?

[krlmlr]

Thank you very much for reporting. This is the first time I ever see a segfault during package installation. Something very strange is going on, but I don't think this is bindrcpp's fault. Can you please rerun the build?

[krlmlr]

Just seen it on one of my machines. Will observe.

[gaborcsardi]

What version of devtools?

[gaborcsardi]

And where is the Travis job, actually?

[krlmlr]

https://travis-ci.org/spedygiorgio/markovchain/builds/242031968 is one.

[gaborcsardi]

shot in the dark: cached incompatible build time dependency: bindr ?

[gaborcsardi]

Or Rcpp?

Anyway, I would try with a clean cache....

[bbolker]

Just a heads-up, still investigating. Was seeing some segfaults running tests for the broom.mixed package locally (GH repo here), that seemed to trace back to bindrcpp. devtools::test() or source("tests/testthat/test-lme4.R")

Now they've gone away (kinda heisenbuggy ...) but I'll report here if they recur.

[krlmlr]

bindrcpp is going away in dplyr 0.8.0 (already gone in the dev version), and is replaced by a reimplementation based on similar principles.