requests for nALFS2

Kevin P. Fleming kpfleming at linuxfromscratch.org
Wed Aug 11 07:52:13 PDT 2004


Jeremy Huntwork wrote:

> I totally understand digest checking items that nALFS may download as
> part of its process, however, if a user has gone through the trouble of
> gathering all the packages beforehand, to me, digest checking is just an
> annoyance and another point of failure.  In that instance I feel that it
> should be optional.

How can digest checking be a point of failure? If the digest check 
fails, you don't have the file that the profile is expecting, so you 
can't expect the profile to continue executing properly. If you removed 
the digest check, it still might not execute properly, because it's not 
the right file.

If the real issue is not the digest checking, but the 
compression/packaging format of the file, then see the other message I 
posted this morning with an alternative solution.

Saying that you want it to act this way is all well and good, but if 
there is no way to express this behavior _in the profile_ then it 
doesn't advance the discussion at all <G> I have strongly resisted 
putting any logic/conditions/execution behavior in nALFS itself, for a 
couple of reasons:

- when someone is reading/editing a profile, the XML they are looking at 
should completely control the final build
- nALFS is not the only ALFS tool around, and there is no documentation 
that tells other ALFS tool maintainers "oh, well, the syntax has this 
construct, but don't implement it that way"

Can you describe some means of informing nALFS that the digest checks 
listed in the profile should not be done, on a package-by-package basis? 
There is no information available to nALFS that it could use to infer 
this, somehow the user is going to have to tell nALFS what is needed.



More information about the alfs-discuss mailing list