Closed RLumSK closed 4 years ago
Same here.
clang 7 R 3.6.2 macOS Mojave (10.14.6)
.R/Makevars
:
CXX_STD = CXX14
CC=ccache /usr/local/clang7/bin/clang
CC11=ccache /usr/local/clang7/bin/clang
CC14=ccache /usr/local/clang7/bin/clang
CXX=ccache /usr/local/clang7/bin/clang++
CXX11=ccache /usr/local/clang7/bin/clang++
CXX14=ccache /usr/local/clang7/bin/clang++
# clang: start
CFLAGS=-isysroot /Library/Developer/CommandLineTools/SDKs/MacOSX.sdk
CCFLAGS=-isysroot /Library/Developer/CommandLineTools/SDKs/MacOSX.sdk
CXXFLAGS=-isysroot /Library/Developer/CommandLineTools/SDKs/MacOSX.sdk
# clang: end
'me to' for several Bioconductor packages during R CMD build, e.g., http://bioconductor.org/checkResults/devel/bioc-LATEST/BEclear/celaya2-install.html presumably this is causing havoc for CRAN as well...
Also worth noting that the CRAN checks for bibtex 0.4.2.1 also show a segfault for read.bib()
on OS X but surprisingly that didn't stop bibtex 0.4.2.1 from being published:
https://www.r-project.org/nosvn/R.check/r-release-osx-x86_64/bibtex-00check.html
ping @romainfrancois
Seems unrelated to the compiler used, also crashes for me when using gcc
.
Edit: As a workaround, installing the "dev" version (which is not actually dev because it still is at v0.4.2) solves the "upgrading" issue when using ´remotes::update_packages()`.
I.e. remotes::install_github("romainfrancois/bibtex")
and the package will stay at v0.4.2 when you are calling remotes::update_packages()
.
A new version (0.4.2.2) of 'bibtex' was just released on CRAN. I tested my example from above again and now it works, thus thanks to @romainfrancois for the fix!
Issue
Description
Today, suddenly I got various fatal crashes of my R sessions (e.g., while using the package 'RefManageR' and 'vitae') which I tracked down to the update of 'bibtex' to version 0.4.2.1 (released on CRAN 2019-12-20), which leads to the following error message:
Remarks
Running Example
Example BIB-file
R call
Session info
Update (2019-12-24)
Now I see that also the CRAN tests for macOS fail, so I guess the problem is real and not related to a particular configuration. However, help is at hand #24 and will probably available once the CRAN is back from vacation.