Busybox #101
Busybox #101
Conversation
sixtyfourktec
force-pushed
the
busybox
branch
2 times, most recently
from
c9c7fa7 to
b2e352c
Compare
Don't try to make modifications to base files if the user didn't expressed its indent to do so. This removes the defaults of those options.
This creates a basic Linux directory/files setup using the merged /usr approach.
Add more options on how the config is gathered. Bring this in line with e.g. the linux kernel, where you could specify a custom package with a defconfig or a busybox defconfig or the standard defconfig. Note, the creation of debugging symbols now depends on the BASEMENT_DEBUG variable. If set, the CONFIG_DEBUG option is set automatically. Also only regenerate the .config if the source config has changed.
Also switch to explicit "if" statements, because otherwise we may miss some errors.
If udhcpc is enabled add a default script which does a basic config.
If the inittab feature is enabled, add a default one.
If enabled put some default init scripts into /etc/init.d.
Add a global config for the init system in use. Here we start with sysvinit compatible init scripts for busybox.
Also add a minimal /etc/network/interfaces file, so that ifupdown is happy when started.
There was a problem hiding this comment.
I was wondering what the full enumeration of the allowed CONFIG_INIT values looks like. Obviously, it is at least sysvinit and systemd. I vaguely remember that we discussed that there might be differences between busybox and "full" sysv-init. Would it make sense to allow something like sysvinit/sysv and sysvinit/busybox and match in the recipes to sysvinit/*?
There was a problem hiding this comment.
This file is apparently from buildroot which is licensed under GPLv2+. I think we should properly attribute this by adding a line or two to state where the file is coming from and under which license it is.
There was a problem hiding this comment.
Again, this file is from buildroot. Please add some copyright header.
No description provided.