Re: Ports & Packages infrastructure & feature import. (fwd)

["Martin S. Weber" <Ephaeton%gmx.net@localhost>]

On Wed, Aug 31, 2011 at 06:41:04PM +0100, Matthias Scheler wrote:
> On Wed, Aug 31, 2011 at 07:24:26PM +0200, Marc Balmer wrote:
> > If only more developers would develop a sense of quality and at least
> > test-compile their stuff, or run a relase build, before committing these
> > figures would already look a lot nicer...
> 
> Yes, I agree. I'm afraid that we are relying too much on our users and
> automated tools to report such problems quickly instead of preventing
> them in the first place.
> 
> I'm just trying to say that it doesn't make NetBSD-current completely
> useless, only less useful than it could be.

Mostly because it scares people away from trying / running -current,
either by bad past experiences ("ok, this was one crash too many!",
or "my userland is b0rked AGAIN! That's IT!") or by lore ("my friend
used to run -current and encouraged me to do so but he was facing
problems with something and got angry about -current. Now he's only
ever running the release branch. He knows way more about NetBSD than
I do, so I follow his lead"). I myself think (maybe I'm mistaken)
that in the days where there still were alphabetic symbols on -current
(remember 1.6ZA?) the quality of -current was higher, although I
don't have evidence backing that up. On the other hand, after some
crashy time on -current, I, too, stopped using it. And, with me, I
guess another ~dozen people I've converted to NetBSD (see lore above).

So yeah, what I'm trying to say is: you're right. People should eat
more of their own dogfood and compile, regression test and test-run
it. Not more, but always.

Regards,

-Martin