[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Bug#197049: ITP: conglomerate -- an XML editor for GNOME



On Thu, Jun 12, 2003 at 11:48:10AM -0500, Manoj Srivastava wrote:
> On Thu, 12 Jun 2003 18:12:03 +0200, Sven Luther <sven.luther@wanadoo.fr> said: 
> 
> > On Thu, Jun 12, 2003 at 09:49:23AM -0500, Manoj Srivastava wrote:
> >> On Thu, 12 Jun 2003 10:31:56 +0200 (CEST), Jérôme Marant
> >> <jerome.marant@free.fr> said:
> >>
> >> > En réponse à Arnaud Vandyck <arnaud.vandyck@ulg.ac.be>:
> >> >> Hi,
> >> >>
> >> >> IMHO, unstable is a good place for it.  Upload Conglomerate in
> >> >> unstable and fill an RC bug "not usable at the moment" to be
> >> >> sure it will not go to testing, even if it's buggy.
> >> >>
> >> >> Everyone knows that unstable is not stable ;-)
> >>
> >> > BZZZT. Wrong. "unstable" means unstable packaging, not unstable
> >> > software. unstable is not for alpha/beta software except when
> >> > there is a stable version around (shipped as -snapshot packages).
> >>
> >> Hmm, no. I would not upload an alpha version of a software package
> >> that is already in the distribution to unstable, since a simple
> >> upgrade may damage a users installation. However, a totally new
> >> package, unless it destroys unrelated data ort causes crashes,
> >> would go into unstable, not experimental.
> >>
> >> Also, packaging is under my control. I never upload packages
> >> without testing out the packaging -- and since I have packaged
> >> other packages, the packaging is never going to be the reason to
> >> upload into experimental.
> >>
> >> Experimental is, in my eyes, reserved for software flakey enough to
> >> destabilize the system, potentially destroy data or damage other
> >> packages, of a experimental version of a package already in the
> >> distribution.
> 
> > So it should be ok if i upload the package to unstable, even if the
> > package crashes often, and thus, since it is an XML editor, the user
> > can loose part of its work ?
> 
> 
> 	*Sigh*. Rules of thumb are no subst=itute for common sense or
>  thinking. 
> 
> 	Did I not say data loss is reason enough for experimental? 
> 
> > At least i should add a warning in the long description then or
> > something such ?
> 
> 	If it causes data loss, you should use experimental. 

Well, the difference is between data loss on the disk and loss of
currently editing stuff.

Friendly,

Sven Luther



Reply to: