Tzafrir Cohen wrote:
> The simplest test case here is running 'lh_build' in an empty directory
> (no configuration).
note that this has never been meant to be working; you are always
supposed to run lh_config before lh_build. i'll add a check so that any
helper will fail immediately if there's no config directory available.
> This will fail at lh_chroot_install-packages because
> the list of packages to install includes the two:
>
> ${LH_UNION_FILESYSTEM}-modules-2.6-amd64
> auto
>
> To which both Daniel Bauman and Chris Lamb claimed that an empty
> configuration is invalid, because it has not been blessed by running
> lh_config on it.
that is not true. i only said that in the case of LH_INITRAMFS, this is
a cosmetic bug that it enforces having run lh_config before build. this
is not trivial to solve (as far as i can see, it's not fixable at all
given that we want the auto value to be expanded at runtime, hence the
check for config in each helper soon). however, LH_LINUX_PACKAGES
handling is still correct as it is.
> I consider this a bug. In order to resolve this I would suggest to set
> some variable in lh_config before calling Set_defaults . That would
> allow Set_defaults to make more informed decisions and specifically
> avoid those two issues.
i don't think this is a sane solution.
--
Address: Daniel Baumann, Burgunderstrasse 3, CH-4562 Biberist
Email: [EMAIL PROTECTED]
Internet: http://people.panthera-systems.net/~daniel-baumann/
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
