On 6/1/11 12:43 PM, Scott Garman wrote: > On 06/01/2011 02:47 AM, Phil Blundell wrote: >> On Tue, 2011-05-31 at 12:53 -0700, Scott Garman wrote: >>> This adds a -native recipe for the shadow utilities. >>> >>> The custom --root option allows the the following utilities to be >>> run within a chroot when invoked under pseudo: >> >> Rather than patching the code for all these utilities, can't you just >> wrap them in a call to chroot(8)? That is, make useradd.bbclass do: >> >> eval $PSEUDO chroot ${STAGING_DIR_TARGET} useradd ... >> >> rather than the existing >> >> eval $PSEUDO useradd --root ${STAGING_DIR_TARGET} ... > > That's a reasonable suggestion. I haven't tried it yet, but I have found > that pseudo's chroot(2) implementation is not complete. One of the cases > where it does not work is when forking child processes, which breaks the > jail and the child processes are no longer chroot'ed. > > My guess is that chroot(8) is going to call chroot(2) and then fork a > child process to run its additional arguments.
chroot should be complete in pseudo, if you can reproduce any failures we should pass them upstream. The reason I suggested the --root option was primarily for the ease of people who are NOT using the automated scripting, i.e. someone manually adding a preinst (or similar) to their recipes. The --root option is easier (to me at least) to understand that having to if-def stuff around a chroot.. (but that is personal preference...) --Mark > cc'ing Mark directly in case he has additional comments or needs to > correct me. > > Scott > _______________________________________________ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core