Package: r-base
Version: 3.5.1.20181215-1
Severity: important
Dear Maintainer,
Please note that I report this bug against r-base because this is where I got
the symptoms. However, I am almost certain that the problem is somewhere
in glibc (see below "Further note").
*** Reporter, please consider answering these questions, where appropriate ***
* What led up to the situation?
On a Debian testing system, routine update of systemwide hand-installed CRAN
packages :
- sudo -i
- R
* > setRepositories() ## I use CRAN and BioC software
* > update.packages(ask=FALSE, checkBuilt=TRUE, lib="/usr/lib/R/library/")
* > update.packages(ask=FALSE, checkBuilt=TRUE, lib="/usr/lib/R/site-
library/")
* > update.packages(ask=FALSE, checkBuilt=TRUE)"
This one fails, with a reproducible error :
I can reproduce this failure by removing the stale lock file and retrying
to install brms :
---------------------------------------------------------------------------
> install.packages("brms")
Installation du package dans ‘/usr/local/lib/R/site-library’
(car ‘lib’ n'est pas spécifié)
essai de l'URL 'https://cloud.r-project.org/src/contrib/brms_2.7.0.tar.gz'
Content type 'application/x-gzip' length 3818402 bytes (3.6 MB)
==================================================
downloaded 3.6 MB
* installing *source* package ‘brms’ ...
** package ‘brms’ correctement décompressé et sommes MD5 vérifiées
** R
terminate called after throwing an instance of 'std::runtime_error'
what(): Mutex creation failed
Aborted
Les packages source téléchargés sont dans
‘/tmp/RtmpRf29DO/downloaded_packages’
Warning message:
In install.packages("brms") :
l'installation du package ‘brms’ a eu un statut de sortie non nul
---------------------------------------------------------------------------
* What exactly did you do (or not do) that was effective (or
ineffective)?
- Retrying the installation
- Upgrading from r-base-3.5.1-2 (IIRC) to 3.5.1-20181215 from unstable
* What was the outcome of this action?
Both attempts were ineffective.
* What outcome did you expect instead?
A normal upgrade of brms.
Further note :
==============
Googling for "R what(): Mutex creation failed" gives a raft of bug
reports in various distributions and source Github repositories issues,
which have in common to point, after various discussions, to glibc 2.28...
which happens to be the current version in testing. Some of them also point
to C++ 11 functionality of gcc (but I am out of my depth here and do not really
understand this part of the issue).
Ouch.
I found no usable suggestion of workaround. Such a suggestion would be welcome
:
- brms is the bread and butter of Bayesian regression models. Missing it
is seriously annoying
- this failure seems to be much more general than R, and may impact other
vital
functionality.
-- System Information:
Debian Release: buster/sid
APT prefers testing
APT policy: (650, 'testing'), (60, 'unstable'), (50, 'stable')
Architecture: amd64 (x86_64)
Foreign Architectures: i386
Kernel: Linux 4.18.0-3-amd64 (SMP w/8 CPU cores)
Locale: LANG=fr_FR.UTF-8, LC_CTYPE=fr_FR.UTF-8 (charmap=UTF-8),
LANGUAGE=fr_FR.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash
Init: systemd (via /run/systemd/system)
LSM: AppArmor: enabled
Versions of packages r-base depends on:
ii r-base-core 3.5.1.20181215-1
ii r-recommended 3.5.1.20181215-1
Versions of packages r-base recommends:
ii r-base-html 3.5.1-2
ii r-doc-html 3.5.1.20181215-1
Versions of packages r-base suggests:
pn ess <none>
ii r-doc-info 3.5.1.20181215-1
-- no debconf information
