Public SourceMage Discussion List

[Jeremy Blosser <jblosser-smgl AT firinn.org>]

On Jan 27, David Brown [dmlb2000 AT gmail.com] wrote:
> > Thanks, but why is this in the linux spell instead of an 
> > smgl/smgl-mkinitrd
> > or something?
> 
> Well, eventually I was thinking on enabling building of the initrd by
> default in the linux spell,

Why would we make this the default?  Is the kernel planning to start
requiring an initrd?

> which means that it would have to be in the linux spell (since the linux
> spell isn't really supposed to depend or anything depend on the linux
> spell).  At least this is how its been in the past.

Having the linux spell not depend on anything make some kind of sense
but it obviously does depend on some things now whether it names them
or not (gcc, for one).  Regardless, I don't think the point there is to
make any potential dependencies part of the linux spell itself.

> I'd have to do more testing to make sure mkinitrd doesn't cause the spell
> to fail and you could choose to use the initrd or not.
> 
> Also linux-initramfs has some scripts that you could put into your
> kernel config to use the in-kernel methods of generating initrds and
> building into the kernel. The mkinitrd script generates a separate
> file and they don't really have any connection.
> 
> Depending on how far we want to go we could move all the scripts to a
> separate spell, smgl/linux-devel or something but linux would have to
> depend on that, I don't really see a problem with this its just not
> they way its been done in the past.

I would rather see us add it as an optional_depends than have it in the
spell itself... if the xen patches ever become a fully integrated option in
the linux spell we'd need to do that kind of thing anyway.

Attachment: pgpntrLBCPyIL.pgp
Description: PGP signature