[Dapper] Who decided to hose wpasupplicant?

Daniel Robitaille robitaille at gmail.com
Sat Mar 25 23:58:17 UTC 2006


> b) we'd like more ordinary users to be involved in testing
>    Dapper, if at all possible.
>
> Which suggests that, while Dapper will be less stable than
> Breezy, it really ought to avoid containing a large class of
> errors. There are instabilities, and then there's
> carelessness. The former should be acceptable, the latter
> not. I believe that this wpa_supplicant mess is the latter.
>
> I worked for a (Windows-based) software company, where every
> single build went through rigorous regression, performance,
> and bug testing. Every time someone discovered a new bug,
> they added a test to the test harness so that that bug
> wouldn't appear in the next release. I realize that for
> various structural reasons this is less likely in the
> open-source world. But I assert that it is vital.

Did your company was supporting 15,000 applications?  Did it had more
developpers than Ubuntu?  (Ubuntu has some thing like 12-15 paid
developpers)?


Also:

$ apt-cache policy wpasupplicant
wpasupplicant:
  Installed: (none)
  Candidate: 0.4.8-0ubuntu1
  Version table:
     0.4.8-0ubuntu1 0
        500 http://archive.ubuntu.com dapper/universe Packages

>From the ubuntu web site it is really clear that the  universe
repository  comes with no support of any kind.  Technically the Ubuntu
developpers are not even officially looking at these packages.

I know that people install packages from universe as needed (we all
do, including me), but there is small risk in doing so.  Maybe it is
less so now with the large MOTU group (compared to the time around
Warty in 2004), but still these extra packages comes with no official
kind of support of any kind.  It's something to keep in mind when we
all enable these extra repositories in sources.list: buyers beware,
and there is is always a risk of something breaking, especially on the
very unstable Dapper.


--
Daniel Robitaille




More information about the ubuntu-users mailing list