Re: [gentoo-dev] package.mask vs ~arch

[Rich Freeman <rich0@gentoo.org>]

On Mon, Jun 30, 2014 at 3:11 PM, Tom Wijsman <TomWij@gentoo.org> wrote:
>
> A test of a package to determine whether it appears to be working OK or
> whether it destructs your system isn't too much asked for; if it works
> it can then be ~arch tested, if it breaks you have a bug # for p.mask.
>
> If someone can't test it at all, why was it added in the first place?

So that it can be tested?  Maybe the maintainer doesn't have the
ability to test the package (might require special hardware).  Maybe
the maintainer doesn't have the time to test it right away, but wants
to allow others to do so (especially if others show an interest).

In my example of mythtv, testing might require first updating all the
front-ends to be current and ensure that nothing breaks (it might only
be emerge --sync'ed monthly).  Then a window has to exist where
nothing will be recorded.  Then everything gets brought down and
backed up (not a big deal, but nobody is watching TV for a while).
Then you update everything and see if it works, perhaps having to
tweak things a bit.  Then you do the quick tests (record shows, play
things back, check the web front end).  Then you leave it alone for a
day and see if anybody screams - best not to do this if you'll be
really busy the next day.

If people are clamoring for an update, it may be more productive to
just let them have it with a disclaimer about quality, rather than
just putting them off for a week or two.

Sure, I can set up yet another overlay, which will be empty 99% of the
time.  But, what is the harm in just using a mask?  I've yet to leave
one sitting around for years (well, not for testing at least).

Rich