Re: [gentoo-dev] USE & KEYWORDS relation

[Francesco Riosa <francesco@pnpitalia.it>]

Chris Gianelloni ha scritto:

>On Wed, 2005-03-02 at 20:47 +0100, Francesco Riosa wrote:
>  
>
>>bob already know how to downgrade but he can't because his boss has 
>>crashed its chair on his head and now he his at the hospital.
>>bad boss, bad world
>>    
>>
>
>In your scenario, Bob is an Admin.  If Bob were *my* admin and had not
>taken the time to read the ChangeLog for the package and not made
>changes on the test network before making them to the production
>network, Bob would be looking for a new job after that coffee.  Also, if
>Bob were not smart enough to not stay with the stable portion of Gentoo,
>which is designed to not have feature loss and other such problems, but
>was instead using the "testing" branch on production servers, then Bob
>would be looking for a new job and I would be looking for a smarter
>administrator.
>
>The whole point is that if you are blindly upgrading *anything* that you
>rely on for day-to-day work, that really there is only one person to
>blame, and that is yourself.
>  
>
Chris,
  Really we are not speaking about me, I blindly make all kind of crappy 
stuff but _**never**_ on production system. Seldom on backup of 
production systems. I like the risk but not when it compromise other 
person than me.

*The* whole points are
that there is no way of choose  a  set  of  features  (via USE flag) and 
expect that emerge keep updated my system obeying to my needs.
the stable/unstable stuff it's *not* a point what happen if upstream of 
say "l7" goes away for six month? your latest stable version will *stop* 
for six month.
Maybe iptables-1.3 must become stable before, and *ONLY* the people who 
want "l7" must stop at 1.2.9
* this * is * the * point *

Now you can say that it's easy or not, possible or impossible, you can 
ever say do it yourself (but expect that I do it).
What you can*not* say is that is a stable/unstable problem, or that the 
user must check before the user *will* check before he/she should not be 
costrained to mess with its /etc/portage/package.mask for *this* kind of 
stuff.

my apologies if that was not clear before, hoping that now it's better 
explained


--
gentoo-dev@gentoo.org mailing list