Re: [Gnu-arch-users] Re: Making microbranches popular

[Andrew Suffield]

On Mon, Jan 26, 2004 at 10:44:29PM -0500, Mirian Crzig Lennox wrote:
> address@hidden (Tom Lord) writes:
> >     > > Is the inevitable resulting user confusion really worth it?
> >     > I think when you start munging names like this, user confusion is
> >     > inevitable :(
> >
> > That's the #1 reason why I resist changes like this.   We have a
> > perfectly servicable namespace right now with a clear structure.
> > It's all downhill from here.   And, for what?  An imperfect support
> > for "arbitrary names"?
> 
> Well, here's the thing: I would dearly love to sell my workplace on
> the idea of ditching CVS for Arch.  However, that won't happen unless
> Arch can support our version naming scheme, which is typically
> [productname]-x.y-rel-z (for natural numbers x,y and z).  For example,
> they want to be able to tag something as "2.1-alpha-4", "2.1-beta-3",
> "2.1-rc5", and so on.  They most definitely are NOT interested in
> saying "alpha-2.1.4", "beta-2.1.3", "rc-2.1.5", etc. because those
> names have entirely different meanings to them, and to our customers
> and beta sites.
> 
> And for what it's worth, I firmly agree.  Our version naming scheme
> isn't uncommon, complicated or irregular; it's perfectly reasonable to
> expect Arch to cope.  If it can't, we ought to fix it so it can.

This is a fairly good example of why it wasn't a particularly good
idea to embed a versioning system into arch in the first place.

Because it's there, people think they have to use it (while most of
the time, the right thing to do is to stick a 0 in the version field
and use the branch name).

-- 
  .''`.  ** Debian GNU/Linux ** | Andrew Suffield
 : :' :  http://www.debian.org/ |
 `. `'                          |
   `-             -><-          |

signature.asc
Description: Digital signature