Closed gdevenyi closed 6 years ago
gdevenyi
commented
6 years ago Commands that were running:
cat <<-EOF | Rscript --vanilla -
r = getOption("repos")
r["CRAN"] = 'http://cloud.r-project.org/'
options(repos = r)
rm(r)
library(devtools)
stopifnot((install_url("$RMINC", dependencies=TRUE, upgrade_dependencies=FALSE)))
stopifnot((install_url("$mni_cortical_statistics", dependencies=TRUE, upgrade_dependencies=FALSE)))
EOF
cfhammill
commented
6 years ago Hmm, the dependencies=TRUE changes the behaviour of install_url, it's trying to install all the dependencies after RMINC, including qvalue which is not available on CRAN. On my machine it can get it from bioconductor no problem, but about the VM. So I can't reproduce.
Can you try unpacking the stopifnot. Then cat or print the return value of install_url, that will hopefully shine some light.
gdevenyi
commented
6 years ago Well, it does install rgl, it can't find X11 since there's no X during the install, but it looks like it makes it past that.
Having said that, I told it not to upgrade dependencies and it did anyways? (RGL was already installed...)
gdevenyi
commented
6 years ago Followup: qvalue is in fact installed via the r-bioc-qvalue debian package
cfhammill
commented
6 years ago Did you try unpacking and checking what it's returning? Is it FALSE, or something else.
> stopifnot("hi")
Error: "hi" is not TRUE.... they renamed upgrade_dependencies in the new version they just released.
FFS.
gdevenyi
commented
6 years ago cfhammill
commented
6 years ago Alternatively, just assume it's returning some extra information, not TRUE, then do
res <- install....
if(isFALSE(res)) stop("Couldn't install RMINC")lol, so now it does what you wanted it to do before, failing with an error if it can't install. It silently returns the package name on success.
gdevenyi
commented
6 years ago Your isFalse wrapper seems to get things passing properly.
I guess stopifnot was too aggressive :)
Thanks.
cfhammill
commented
6 years ago I actually don't think it's needed anymore, if install_url throws an error on failure, there is no need to check the return code.
gdevenyi
commented
6 years ago Its not clear from the devtools NEWS that it does that...
cfhammill
commented
6 years ago it is reasonably clear from the code ;)
More issues with the VM.....
Everything looks fine, and logging in and testing, RMINC runs...
Any ideas?