ALFS Status: Past and future [was Re: new guy, newbie questions.]

Felipe Contreras al593181 at mail.mty.itesm.mx
Mon Jan 21 21:07:43 PST 2002


On Mon, Jan 21, 2002 at 08:40:49PM -0800, Jesse Tie-Ten-Quee wrote:
> On Sat, Jan 19, 2002 at 04:48:31AM -0500, Rob Landley wrote:
> > This is only about half of a true package management system.  You also need 
> > to be able to uninstall packages, upgrade existing packages, and to query the 
> > system for what's installed and information about the configuration of what's 
> > installed.  Package management needs an inventory database (which could very 
> > well be XML based, sure).
> 
> Yes, but we aren't writting a package management system; at least not
> quite yet.
> 
> The initial goals are to make a system to compile and install LFS without
> user interaction, using XML "profiles" to describe the process.
> 
> Hopefully we will eventually add full package management support, but
> that is a *huge* ball of wax and not something we should implement
> lightly.

Why don't define this, ALFS is a project that aims to the creation of an
extendalbe system builder and package installer.


> > Are their data files compatable with your XML document type definition 
> > thingy?  (Do you have a standard they're implementating variants of?  If not, 
> > if you've defined and implemented a standard, would they be willing to change 
> > to become compatable with it?)
> 
> Not really, there's basiclly two camps.
> 
> Those writting ALFS implementations with shell scripts, and don't
> consider the use of XML or writting it in C/perl (or other languages)
> usefull.. so i doubt they will become that standard, i know of about at
> least 30 different ALFS shell script implementations. [it's either that,
> or they don't know about ALFS; there's been alot of ppl in #LFS like
> that recently or just like the idea of writting there own version]
> 
> Allthough someone could take the ALFS profiles and use XSLT to convert
> them to there own data format, i do believe someone is implementing
> this.  If not, they perhaps someone should consider approach :)

Totally true, it's much better to agree on a standard so all the
projects can share the information. I know pretty well the advantages of
having a format of a high level of complexity, it can be transformed
into simpler formats.

Could you please search for my post about "lots of ideas" and take a
look at how I tought an ALFS parser should work? Or is that irrelevant
right now?


> Then there are those that took the syntax i had been working on last summer
> and improved and added to it, nALFS and "ALFS" (Mark Ellis's Perl ALFS)
> are doing this, i'll need some time to look at all the changes they have
> made.. this is a messy subject though.
> 
> Syntax like a few people have mentioned is really the heart and true
> goal of ALFS, and it most likelly will never be quite "right" nor will
> everyone agree to it.
> 
> Hopefully in the next few weeks we can iron something out that we *all*
> agree on, if not...well, i'll just use what works and go from there.

I hope that too.

> ALFS isn't about educating linux, that's what LFS's long term goal is.

I think this should be the second step of ALFS, to be a "command
propagation system" so that at some kind having a good ALFS profile will
give you good ideas about how to install stuff.

-- 
Felipe Contreras
-- 
Unsubscribe: send email to listar at linuxfromscratch.org
and put 'unsubscribe alfs-discuss' in the subject header of the message



More information about the alfs-discuss mailing list