tphhec01 at degas.ceu.hu
Sat Jun 21 08:32:27 PDT 2003
In article <20030619155339.GB17564 at urusai.localnet.rcl>, richard at nezumi.plus.com wrote:
> If you set some program to scan configure configure scripts you will
> have to find beer in a refrigerator, and sacrifice a vegetarian voodoo
> chicken. There is a whole pile of still that configure scripts look for
> just to see if anyone reads the output.
Couldn't be this dependency tracking done based on header files? Of course,
even if it worked, this would find *all* utilizable other packages, and not
just the ones needed for a sane basic functionality.
But can we *ever* hope for better? The "sane basic functionality" is
somewhere between the minimal stuff for get the package built and the set of
all utilizable things, and it's of a quite subjective nature. I don't see
how could it be find algorithmically.
Anyway, an algorithmic way of finding a correct overestimation of the
dependencies wouldn't still be helpful? Trimming down such a superset
shouldn't be so hard.
At various Linux forums you can pick out the experts by checking who say /
imply "Real men use Debian". How stupid. No doubt, real men use LFS :)
But the toughest guys are the ones who, besides using LFS, have such an own
package management system which copes with Debian's.
Unsubscribe: send email to listar at linuxfromscratch.org
and put 'unsubscribe lfs-dev' in the subject header of the message
More information about the lfs-dev