A question about package development. Our package tarfile (spatstat_1.5-10.tar.gz) has already passed R CMD check. We install the package in a local directory, either by hand or by running the package checker. Starting R 2.0.1 and loading this package using library(spatstat, lib.loc="./spatstat.Rcheck") or whatever, we get the following error as soon as there is a data() command: > Error in lazyLoadDBfetch(key, datafile, compressed, envhook) : > file open failed The file permissions are all 644. This happens on at least two systems including pc-linux-gnu. Is this a bug in lazy loading ? oder? thanks Adrian Baddeley & Rolf Turner
lazy blighter
2 messages · Adrian Baddeley, Gabor Grothendieck
Adrian Baddeley <adrian <at> maths.uwa.edu.au> writes: : : A question about package development. : : Our package tarfile (spatstat_1.5-10.tar.gz) has already passed R CMD check. : We install the package in a local directory, either by hand or by : running the package checker. Starting R 2.0.1 and loading this package : using library(spatstat, lib.loc="./spatstat.Rcheck") or whatever, : we get the following error as soon as there is a data() command: : : > Error in lazyLoadDBfetch(key, datafile, compressed, envhook) : : > file open failed : : The file permissions are all 644. This happens on at least two : systems including pc-linux-gnu. : : Is this a bug in lazy loading ? oder? : You could try adding LazyLoad: no to your DESCRIPTION file although even if that works it may just get rid of the message and there may still be a real error in your package lurking. In the past I have found that often a lazyload error is not related to lazyloading at all and signifies some completely different error. Once I was able to get past it by using forward rather than backward slashes on a pathname in MS-Windows, for example. Another time I had to successively remove half my package until I pinpointed the real error. If you search the archives you can find comments similar to this by myself although I think I got the main points here.