FESCo wants to ban direct stable pushes in Bodhi (urgent call for feedback)

[Kevin Kofler]

James Antill wrote:

> On Wed, 2010-03-03 at 01:54 +0100, Kevin Kofler wrote:
>> Well, I see where you're getting to now, but this is really not what
>> updates-testing is for! Updates-testing is for TESTING updates, not for
>> being used as production by some users, even those who want more updates.
> 
>  It would be used as production by users like _you_, not normal users.

The implication being that I'm not normal? I find that insulting.

>>  I
>> want many updates, but I don't want to be the guinea pig for updates
>> which just hit testing,
> 
>  And nobody else wants to be the guinea pig for _you_.

People who use updates-testing under the current system are signing up to 
doing testing. Under your proposal, they'd be forced to sign up to get any 
current updates. Sure, it may be a way to get more testers, but this kind of 
coercion makes me (and I'm sure I'm not alone there!) want to use another 
distro rather than giving in to your racket. Of course, the fact that there 
isn't any other distro offering the kind of policy we currently do could 
work to your advantage, but that won't make it any less coercive.

>>  and I also don't want to have to selectively update
>> because it's a mess.
> 
>  Why's that? Maybe because of the sheer volume of updates ... because of
> packagers like...

No, because updates may depend on previous updates to work properly. We 
can't possibly test or support all possible combinations of updates.

>>  Do you want users of stable to suffer through KDE
>> bugs or be forced to use testing?
> 
>  Again with the "suffering users" because they don't get 6GB of updates
> a month? I think not, and I'm not alone.

They're suffering because their bugs are not getting fixed.

>>  You're effectively forcing everybody to
>> use testing
> 
>  No, that's your proposal ...

Nonsense. My proposal forces nobody to use testing, they get tested updates 
which are pushed to stable after sufficient time in testing to perform 
actual testing (which may well be zero for trivial updates, but is 
definitely nonzero for stuff which has a significant probability of breaking 
something), but not some arbitrary exponential time made up to arbitrarily 
penalize updates.

>  mine allows people who want to get lots of packages to do so

Only by using updates-testing => you're effectively forcing those people to 
use updates-testing.

>  and those who have a working system to not be forced to test your package
>  of the week.

They're not testing it, they're receiving it already tested.

>  Yes, that means that people who want insane amount of updates (like
> you) will have to do more testing than the rest ... but this is fine,
> nobody owes you months of testing.

Everything longer than 3 weeks stops being testing and is more like 
arbitrary withholding.

>  You keep saying that 7 days is "enough" but I haven't seen you provide
> _any_ evidence to support it. Noting that it will often take 3-4 days
> before a package in testing can be seen by all users.

The 7 days are counted from the time of the update push. (That's why there's 
also a distinction between "may be pushed directly to stable" and "0 days 
minimum in testing". The latter means the package should be allowed to reach 
testing and should only get queued for stable once it's in testing. (And all 
this is "should" because I consider such minimum requirements to be 
indicative only. Forcing them only makes sense if one is trying to turn 
updates-testing into something different than a place for testing, like what 
you're trying to do.))

> So maybe you are under the impression that all the users who would test
> your package are anxiously waiting for your packages to be available?

For those packages where regressions actually matter to people, they 
definitely are. People keep asking us: when will KDE x.y.z finally be 
available? They ask it even before upstream officially announces the 
release!

If nobody cares about the package, then frankly I'm not going to worry all 
that much about potentially breaking it. And waiting longer isn't going to 
magically turn up testers for a package nobody uses, or at least no updates-
testing user.

>  More likely you just don't want any real testing done, because hey ...
> that's what all the users in stable are for, right?

See above. The times we're using in KDE SIG have been working well for us, 
they're not random, but based on experience.

>> whether it's the first, the second or the 10000000th update.
> 
>  While it wouldn't surprise me if you did 10 million updates for KDE,
> and assumed that was a good idea, I would bet a lot that users wouldn't
> think that was a good idea.

This was obviously an arbitrary number. :-/ No package is going to get 
updated 10000000 times in a single release. ;-)

        Kevin Kofler