Skip to Content.
Sympa Menu

sm-commit - Re: [SM-Commit] GIT changes to master grimoire by Peng Chang (Charles) (321ae4a9af4a214510d1a4f22d5f1f7a09cae17c)

sm-commit AT lists.ibiblio.org

Subject: Source Mage code commit list

List archive

Chronological Thread  
  • From: flux <flux AT sourcemage.org>
  • To: sm-commit AT lists.ibiblio.org
  • Subject: Re: [SM-Commit] GIT changes to master grimoire by Peng Chang (Charles) (321ae4a9af4a214510d1a4f22d5f1f7a09cae17c)
  • Date: Tue, 20 Jul 2010 15:02:10 +0900

Andraž 'ruskie' Levstik (ruskie AT codemages.net) wrote [10.07.20 14:25]:
> > With regards the statement "dependencies should be explicit not
> > implicit", to an extent, yes. However, there is a point where explicit
> > dependencies are unneeded and, in some cases, even a bad idea.
>
> In case of texinfo it's enough that one thing depends on install-info
> and it is necessary.
>
> With libs it's needed so that any up triggers work.


My point wasn't specific to either texinfo or ncurses. I was making a
general point that we should generalize a document specifying when
things can be implicitly depended upon and when they can't. This is
more clear in the case of gcc and glibc, but can be true for others as
well. I'm not saying that texinfo and ncurses are not explicit
dependencies of texlive. Just making sure we are both understanding each
other on this. :)

> > In the case of depending on a compiler (like gcc), I don't think that
> > makes sense on a source-based distro unless a spell really depends on a
> > specific compiler (like it works only with gcc due to gcc extensions or
> > something similar).
>
> Not everything but texinfo is optional to basesystem. And there's quite
> a few spells that have a ncurses dependency and there's probably a lot
> more that don't have it.

Although the spells that seem to be implicitly depended on (allowably?)
come from basesystem, I don't think that should be the primary factor
determining whether it can be implicit, though it is a factor. I'm just
wondering if can describe an applicable rhubric to determine whether a
dependency can be implicit rather than explicit, and then whether it
should be implicit.

For the current case of compilers, I think implicit is better, at least
until we have a PROVIDEs for COMPILER. That way, if a true alternative
for gcc becomes available, we can allow any compiler to do the job
(pending that other things have been set up properly). The same goes for
glibc (like swapping for eglibc or some other libc in case someone
wanted to try to adapt SMGL for embedded systems). If everything
dependend explicitly on gcc/glibc/etc. now, then attempting to adapt the
system to use an alternative would require re-coding all those
dependencies for pretty much every single spell. That would be a mess.
Thus, this is one argument (though I'm welcome to someone showing me
this is wrong) for where having an implicit dependency should be done.
The case is less clear for ncurses with some spells (not all), since
some will use either ncurses or slang, or perhaps neither but will end
up depending on it if only it's installed (which it will be on all SMGL
systems, so it's hard to know otherwise). Maybe this isn't really an
issue, but I think it can at least sometimes be confusing to people who
aren't aware of it.

--
Justin "flux_control" Boffemmyer
Cauldron wizard and general mage
Source Mage GNU/Linux
http://www.sourcemage.org

Attachment: pgp5rWuMtliJN.pgp
Description: PGP signature




Archive powered by MHonArc 2.6.24.

Top of Page