Re: Developer Behavior
On Wed, Jan 10, 2001 at 02:34:39AM +1100, Russell Coker wrote:
> But I think that there is some merit to having discouragement towards running
> unstable on production machines. I've been getting flamed immensely recently
> about my lilo package that over-wrote lilo.conf incorrectly. Even though:
You're probably getting flamed more because it's a "this is just wrong"
sort of bug - there's no way it was ever going to be safe to install
that version of the package. While users should expect that unstable
may be broken some of the time there's also the expectation that
developers will try to minimse this breakage or (if it's unavoidable)
make an effort to warn people (as with the current INN package saying
"I'm about to hose your system - are you sure?").
> The people who flame the developers contribute nothing. When they report
> bugs that exist they invariably do so after more polite people have already
> reported them and the developer has started work. Then work has to be
> interrupted to spend time fighting off flames.
This I would agree with.
> I don't think that unstable should be limited to Debian developers, but I
> think that it should be restricted to discourage people who aren't reading
> debian-devel. What if we setup the servers to use a different random
> password every month that was only announced on debian-devel?
It would be nice if people actually used unstable. Besides, there's
still no guarantee that people are actually going to read the warnings
or even that they will be warned before whatever it is causes the
Mark Brown mailto:email@example.com (Trying to avoid grumpiness)