[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Wth "rolling"?



>> > 'rolling' is a statement by the project that we consider 'testing'
>> > (renamed to 'rolling')
>> Why the heck do we start by renaming testing? This will seriously
>> disrupt service for anyone for DAYS. There are just too many places
>> tools are using "testing" hardcoded. Too many users having that in
>> sources.list. Too many things assuming there is "stable, testing,
>> unstable". And all of them would suddenly, out of nothing, have broken
>> systems and need to fix them.

>> If somehow rules for testing get changed (to be whatever rolling wants
>> to be), fine. Thats one thing.

>> But for what reason change the name? That's worse PR than usually
>> done by politicians, and they generally do the things noone with a brain
>> ever does. So why?

> How much of that would apply if we renamed testing to rolling (because
> it reinforces the PR message), but kept a symlink from testing to
> rolling?

Most, only the user part goes away. OTOH if we just symlink rolling to
testing and keep the rest as is, then I don't care about rolling. And we
would have the same net effect. (And from FTP* discussions, the idea to
rename the suite for nothing (as this is) had, at best, a "braindead
stupid" idea as a response, so I guess this has to be a GR unless a
better argument than "it sounds nice in PR" comes up).


But you left out the other part I had, how to get stuff better for users
without all the useless PR games. What I see, we can have the stuff
behind this already. We can have regularly updated "releases" (call em
alpha or beta), including d-i and possibly updated packages, we can have
the users go over there. Without breaking the development chain we know
leads to another stable release.

It just needs people to not discuss but actually work on that.

And, double benefit, it would be in parallel to what we have now, not
forcefully killing one thing without knowing if the new thing has any
benefit[1]. And if it turns out better and helps us along, the worse thing
will die on its own. (I would *like* to see something like 2 to 4
"beta"-releases of testing a year go out, including a d-i and stuff). We
have the means to do it technically. We throw the manpower this would
need into discussions and GRs, hoping someone somewhere can be forced to
do it.


[1] Right now it is all blind guessing. "The users will love it". "The
    maintainers will support their packages better"(WTF? Due to a
    rename? Stop dreaming)

-- 
bye, Joerg
English side ruined! Must use French instructions! Le GRILL?? What the hell is that?


Reply to: