BusyBox Strikes Again?
On Thursday 03 April 2008 22:08:34 Daniel Baumann wrote:
> Kel Modderman wrote:
> > loop mounting is just one example to highlight that klibc-utils/mount
> > is -ne to busybox/mount.
> > Another example is readlink, where klibc don't support -f option.
> > This may not even be the cause of this _specific_ bug, but I mention it anyway.
> what do you suggest, should we basically just call busybox everywhere to
> get rid of the failures?
No, thats probably overkill. A careful audit is required though, of existing
code and future imports from casper, to ensure busybox specific features are
used correctly. I think only in very compelling cases should a klibc-utils
binary be overridden like in above example.
I finally think that the initramfs shell environment is consistent for both
when busybox is present or not in initramfs. It makes life difficult for
live-initramfs, however, because of the overall size of the scripts...