[DRE-maint] Re: Ruby-full

[Rudi Cilibrasi]

I have given Vincent's ruby-full package a try.  I tested with three
machines: two i686 and one powerpc.  In all three cases using apt-get
with his repository, the ruby-full package worked fine and indeed
pulled in many ruby libraries and binary packages.  It seems to be
quite a timesaver actually.  So it looks like it really would be more
convenient for a fair number of people.

The one problem I have found in the package is that I can build it with
svn-buildpackage --svn-ignore
but not without the
--svn-ignore
flag because of the control-file regeneration I think.

That is the only problem I have found.  I can't think of any potential
negative reprecussions from this package, can anybody else?  Cheers,

Rudi

On 12/23/06, Rudi Cilibrasi <cilibrar at gmail.com> wrote:
> I think it would be good to have one large package with a bunch of
> Depends: instead
> of Suggests:.  The package would pretty much be redundant for all of us old ruby
> experts who know which packages we need, but it would be great for
> beginners and intro to ruby articles.  That is the main benefit to me;
> it's just too much to type to get ruby and irb and something else for
> somebody new to ruby and so many books are already written assuming
> all these things.  It would help clear things up.  I use apt-get and
> it also seems to require more than 1 package to get a
> "full-featured"ish ruby.
>
> It does seem like it is a core ruby team issue though to me too.  I
> wonder if the best way to do it would be to just procede with
> developing it here and Cc the core ruby mailing list so they can
> decide if they want to support it or veto it?
>
> Cheers
>
> Rudi
>
> On 12/23/06, Vincent Fourmond <vincent.fourmond at 9online.fr> wrote:
> > Paul van Tilburg wrote:
> > > On Fri, Dec 22, 2006 at 12:23:31PM +0000, vincent.fourmond at 9online.fr wrote:
> > >>> I don't see how: no packages depend on ri, rdoc and irb to provide a full
> > >> ruby install. Or at least, nothing that aptitude shows me. So, unless I am very
> > >> mistaken, this problem is still on. Moreover, people I wrote to this morning
> > >> were using relatively recent debian versions, and they didn't seem to notice
> > >> this.
> > >
> > > When using apt or aptitude, installing ruby gives you ruby1.8 which
> > > gives you rdoc1.8 and ri1.8.  IRB can be seen as a separate app, thouhg
> > > it's a usefull one, so maybe it can be added.
> >
> >   It merely suggests, which is far from enough. Don't forget this
> > package is aimed at beginners. Moreover, it is far from instinctive to
> > use irb1.8, ri1.8 instead of the simple ri, irb...
> >
> > >> Namely, what has been required is a package ruby-full that would depend on:
> > >>
> > >> ruby, irb, rdoc, ri, libgdbm-ruby, libdbm-ruby, libruby-extra and maybe the tk
> > >> library.
> > >
> > > The gdbm, dbm, openssl and tk libraries were split of of libruby1.8 for
> > > a reason.  Maybe Akira can enlighten (again) us about this. I am fine
> > > with it, more so because it was a careful compromise at the time.
> >
> >   I don't ask to put them back in one package, I just want to provide
> > users with a simple way to get a full ruby install.
> >
> > > These libraries are almost never used and drag in extra dependancies.
> > > Also, IMHO, libruby-extras has nothing to do with it and this possible
> > > meta-pkg also lies within the realm of ruby/ruby1.8 core package team.
> >
> >   Agreed, but I believe there is some possibility to get this into Etch,
> > but we need to move fast. The package is ready (or nearly so). I'm ready
> > to give over maintainorship if that is necessary.
> >
> >   I just find it a pity to read 'debian sucks' on the lists when there's
> > such a simple solution ;-)...
> >
> >   Cheers,
> >
> >         Vincent
> >
> > --
> > Vincent Fourmond, PhD student
> > http://vincent.fourmond.neuf.fr/
> >
> >
> > --
> > To UNSUBSCRIBE, email to debian-ruby-REQUEST at lists.debian.org
> > with a subject of "unsubscribe". Trouble? Contact listmaster at lists.debian.org
> >
> >
>
>
> --
> Experiment with Artificial Intelligence  at http://complearn.org/
>


-- 
Experiment with Artificial Intelligence  at http://complearn.org/