Piavlo wrote: [Sun Jul 26 2009, 05:06:29PM EDT]
Guido Günther wrote:
Hi Piavlo, On Mon, Jul 13, 2009 at 12:43:14AM +0300, Piavlo wrote: [..snip..]
So something must be wrong in the debootstrap code, and the following patch fixes it
# diff -Nru /usr/share/debootstrap/scripts/lenny.orig /usr/share/debootstrap/scripts/lenny --- /usr/share/debootstrap/scripts/lenny.orig 2009-07-13 00:15:25.000000000 +0300 +++ /usr/share/debootstrap/scripts/lenny 2009-07-13 00:15:16.000000000 +0300 @@ -80,7 +80,7 @@
second_stage_install () { x_core_install () { - smallyes '' | in_target dpkg --force-depends --install $(debfor "$@") + chroot $TARGET dpkg --force-depends --install $(debfor "$@") }
p () { #
could you check what the issue is here? Is it the smallyes pipe part or the in_target call? Both look harmless and I can't seem to reproduce this here. -- Guido
Hi Guido,
I fail to understand what's wrong in in_target_failmsg defined in /usr/share/debootstrap/functions which is called from in_target function:
in_target_failmsg () { local code="$1" local msg="$2" local arg="$3" shift; shift; shift if ! chroot "$TARGET" "$@"; then warning "$code" "$msg" "$arg" return 1 fi return 0 }
I don't understand why the if is entered while: chroot "$TARGET" "$@" finishes with 0 exit code?
I have not been following this conversation previously, so I apologize if this has already been mentioned. Guido's patch also removes the smallyes pipe, which could be the culprit if dpkg (or a subprocess called by dpkg) wants a tty on stdin. Does Guido's patch still "solve" the problem if it's written: - smallyes '' | in_target dpkg --force-depends --install $(debfor "$@") + smallyes '' | chroot $TARGET dpkg --force-depends --install $(debfor "$@") or alternatively, does this work as well as Guido's patch: - smallyes '' | in_target dpkg --force-depends --install $(debfor "$@") + in_target dpkg --force-depends --install $(debfor "$@") Aron