Skip to Content.
Sympa Menu

sm-users - Re: [SM-Users] menuconfig for casts

sm-users AT lists.ibiblio.org

Subject: Sourcemage Users List

List archive

Chronological Thread  
  • From: Andrew <afrayedknot AT thefrayedknot.armory.com>
  • To: sm-users AT lists.ibiblio.org
  • Subject: Re: [SM-Users] menuconfig for casts
  • Date: Sat, 17 Dec 2005 23:30:23 -0800

On Sat, Dec 17, 2005 at 03:58:56AM -0200, Arthur Nascimento wrote:
> static. And I got a question about that. Does the dependencies really
> have to be in bash scripts? What is wrong with a static listing?

Short answer: Yes
Longer answer: Most certainly yes, spells sometimes need that amount of
flexibility to accomplish what they need. Remember, we're building from
source here, so all sorts of doors open up. Restricting things would
limit flexibility very quickly.

> Everything else stays scripts, so sorcery remains turing-complete, if
> you guys insist.

Spell scripts, are turing complete, as they are bash scripts, which
was probably what you meant.

'sorcery' is not really a language, just as 'cat' isn't a language,
so it can't be turing complete. Languages can be turing complete, but
you cant really apply a property of a language to something that isnt
a language.


> I say this because after all, it is the way things
> are usually done. I can only guess since I have not looked into them
> (you are right, I know nothing of package
> configuring/building/installing - thank you for the previous
> explanation), but I imagine it is the way rpm and deb packages are
> made, as well as Portage, as Seth pointed. The particular dependencies
> can still be queried to give power, as Seth said, but there is no need
> for putting the actual queries along with the dependencies infos or
> checking their results right there. I am quite sure (from my very
> limited experience) that there is no dependency checking so complex
> that you actually need bash for it.

There is, and if there isn't, there will be someday. Without it, the
system is much less powerful, and that in the end leads to putting
un-necessary burden on the user.

> Most is as simple as 'require
> this', 'require this (optional)', 'require one of these' and 'require
> one of these (optional)'.

Key word: Most
Most is not all. Also, all this falls apart once we want to start
looking at CONFIGURE scripts, or the other scripts that are useful in
depends resolution but have no standard form, and would probably suffer
greatly if one was put on them. Of course, the grimoire team could decide
to speak up and tell me Im totally off base and they dont need all this
freedom, but I dont think that will happen.

> And some rare cases can be taken care of
> with just a little more thinking. To apply this idea to the configure
> step would be more tricky since configuration can be more complex.
Yes, you could, and in fact, when you were done you'd have either
a) a really complicated configuration language that tried to do
everything that everyone wanted but would failed miserably due to
non-extensibility; there will always be another situation that isn't
handled
b) a new and effectively pointless turing complete language, which is
what we already have

Just depends on when you want to give up trying...'a' would eventually
become 'b' if you never stopped trying to fix it.

> However, I think bash isn't really needed for that either and that all
> this is viable and good for the distro. A menuconfig-like interface
> would then be 'easy' to make as an optional way to handle the system.

I think the take home message is: we can do a lot more for the user with
scripts than with static configurations. Scripts are always going to be
ahead of the curve whereas a configuration format is always going to be
struggling to keep up. In the end, the script can do a lot more to make
the users life easier. But not knowing the alternative it can easily be
taken for granted.

>
> After all, it seems like not all of the ones on the mailing list are
> on the business of mathematics or computer science as I thought in the
> beginning. :) Thanks for the support.

At some point things transition between insulated user experience and
implementation in terms of real code. And at some point, real code has
to meet up with at least some theory. So I think its inevitable for that
to happen when people want justification for things...

-Andrew

--
_________________________________________________________________________
| Andrew D. Stitt | acedit at armory.com | astitt at sourcemage.org |
| irc: afrayedknot | Sorcery Team Lead | ftp://t.armory.com/ |
| 1024D/D39B096C | 76E4 728A 04EE 62B2 A09A 96D7 4D9E 239B D39B 096C |
-------------------------------------------------------------------------




Archive powered by MHonArc 2.6.24.

Top of Page