Has Red Hat helped or hurt?

[Christian Tanzer]

Sean Reifschneider <jafo at tummy.com> wrote:

> >Do I understand you correctly in that it was perfectly reasonable for
> >Red Hat to first ship a broken gcc and then change to a newer one
> >during the 7.x series?
>
> It apparently fit into their upgrade policy...  My understanding, from
> listening to C++ programmers who were really following the issue, is that
> the push to that "broken" compiler (which compiled without problem all
> the extras we were shipping, BTW) was really necessary because the previous
> compiler was also broken for normal use...  Dealing with things like
> templates and the like...

The compiler they shipped generated a lot of question why Python was
"broken".

In such a case, it would seem reasonable to ship a standard gcc for
normal use and an experimental gcc for the C++ programmers interested
in templates and stuff.

> >I doubt anybody would care which Python version Red Hat uses for their
> >tools if they didn't make it unreasonably difficult for their users to
> >use a current Python version for non-Red Hat purposes.
>
> It doesn't seem unreasonably difficult to use Python 2.x on Red Hat 7.x
> boxes.  You pick up the RPMs from python.org and install them...

IMHO, the distribution should not force a specific Python version to
be the default. I.e., having /usr/bin/python point to 2.x should not
break any number of Red Hat utilities. Forcing the user to specify a
Python version in the #! line or on the command line sucks.

> Or (shameless plug), you buy KRUD which installs Python 2.2.1 as part
> of a normal Red Hat-based install...

I'd rather continue to use Debian, thank you.

-- 
Christian Tanzer                                         tanzer at swing.co.at
Glasauergasse 32                                       Tel: +43 1 876 62 36
A-1130 Vienna, Austria                                 Fax: +43 1 877 66 92