This is an infelicity in a recent glmnet release, I will follow up off
list in a minute.
Best,
Uwe Ligges
On 22.09.2011 13:16, Nick Sabbe wrote:
While building/checking a package of mine that depends on glmnet, I got the
following message:
* checking whether the name space can be loaded with stated dependencies
...Warn
ing: running command
'"C:/Users/nisabbe/Documents/R/R-2.13.1/bin/i386/Rterm.exe"
R_DEFAULT_PACKAGES=NULL --vanilla --slave -f
C:\Users\nisabbe\AppData\Local\Tem
p\RtmpZpjj07\Rin121ae93' had status 1
WARNING
Error: .onLoad failed in loadNamespace() for 'glmnet', details:
call: cat("Loaded glmnet", installed.packages()["glmnet", "Version"],
error: could not find function "installed.packages"
Execution halted
A namespace must be able to be loaded with just the base namespace
loaded: otherwise if the namespace gets loaded by a saved object, the
session will be unable to start.
Probably some imports need to be declared in the NAMESPACE file.
I have not changed anything to my dependencies (since my last successful
build the day before yesterday), and to my knowledge, I have not changed
anything relevant. I am using R 2.13.1 (unchanged since the last successful
build) on Windows 7 32bit. When starting a vanilla R terminal, the function
installed.packages is present and can be run. Apart from this, I do not get
any warnings or errors (e.g. the examples run fine). Also, as far as I can
tell, I've never had the warning wrt status 1 (up to today), but it is
unclear whether one is just the consequence of the other. Note: the sources
for this package can be found on R-Forge:
https://r-forge.r-project.org/R/?group_id=1130
The namespace itself contains as its import statement: import(glmnet, rgl,
Matrix, snowfall)
Can someone please help me avoid this warning, as I prefer 0 errors 0
warnings.
On a related note, while trying to alleviate the problem, I ran
update.packages, which led me to another problem: apparently, packages that
are loaded at the time of update.packages are not updated and may end up
broken (this was the case for lattice and Hmisc, for me). Luckily my
colleague had gone through the same troubles and pointed me to the solution.
I can imagine it to be hard to define exactly which packages should or
shouldn't be updated, but it would be nice if the update.packages at least
issued a warning against this.
Nick Sabbe
--
ping: nick.sa...@ugent.be
link:<http://biomath.ugent.be/> http://biomath.ugent.be
wink: A1.056, Coupure Links 653, 9000 Gent
ring: 09/264.59.36
-- Do Not Disapprove
[[alternative HTML version deleted]]
______________________________________________
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
______________________________________________
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
