On Mon, 10 Jan 2005 00:04:36 +0100, Danilo Piazzalunga <[EMAIL PROTECTED]>
said:
> Alle 19:56, domenica 09 gennaio 2005, hai scritto:
>> This is not quite good enough, since the two packages still
>> contain the same file, and hence must conflict with each
>> other. �The �solution is not to have the kernel-headers package
>> create that link, �or to have it delete the link when the proper
>> kernel-image package is �to be loaded.
> I'm sorry, but there is no conflict: kernel-package does its job
> perfectly :-) and doesn't build two packages stepping on each
> other. Our link only belongs to one package:
No, you are wrong. Here is a listing from my machine:
__> egrep 'build$' /var/lib/dpkg/info/kernel-image-2.6.*.list
/var/lib/dpkg/info/kernel-image-2.6.8.list:/lib/modules/2.6.8/build
/var/lib/dpkg/info/kernel-image-2.6.9-2-skas3-v7.list:/lib/modules/2.6.9-2-skas3-v7/build
/var/lib/dpkg/info/kernel-image-2.6.9-skas3-v6.list:/lib/modules/2.6.9-skas3-v6/build
/var/lib/dpkg/info/kernel-image-2.6.10-skas3-v7.list:/lib/modules/2.6.10-skas3-v7/build
> Therefore, implementing this would not hide a real conflict, but
> only suppress a somewhat spurious warning: when we say that the
> directory should be empty, we mean it shouldn't contain any modules
> (or modules-related file), but I don't see why the presence of a
> convenience link shouldn't be allowed.
kernel-package does produce image packages with a real build
symlink, so any header package should not have one of those.
manoj
--
No matter how cynical you get, it's impossible to keep up.
Manoj Srivastava <[EMAIL PROTECTED]> <http://www.debian.org/%7Esrivasta/>
1024D/BF24424C print 4966 F272 D093 B493 410B 924B 21BA DABB BF24 424C
