On Wed, Jan 30, 2013 at 21:37:03 +0100, Andreas Tille wrote: > On Wed, Jan 30, 2013 at 09:32:47PM +0100, Julien Cristau wrote: > > > My analysis was wrong. It actually does access the network but *if* it > > > is able the build breaks (as well as the attempt to load the library in > > > R). I guess there are good chances that the build would have worked if > > > network access is not possible because there is a timeout defined after > > > that the version check (which causes the failure) is interupted. > > > > > IMO an attempt to use the network during the package build is a no-no... > > Do you honestly assume that all the x-1000 packages will fullfill this > criterion. I honestly think anything else is broken. Cheers, Julien
Attachment:
signature.asc
Description: Digital signature