On 30.06.2016 00:55, Sandro Knauß wrote:
Hi Sandro,
But maybe, next time we screw, drop a short note please. Danimo has
enabled the updater for 2.2.2 - happy updating :-)
And keep in mind, there are also distribution that build the ooc for their
users - If you think that version are screwed up, than this is a very
interesting information for distribution. Normally you guys complain ditros to
be "slow" in making the release ready for the users. But yourself do not trust
the released version and wait also some time before giving it to everybody.
It is not a question of "trust" and intention and all these big words.
We just saw that 2.2.0 and unfortunately 2.2.1 were troublesome, and
held them back. It was not a super critical problem, nor something we
could not have talked about. Most people did not even realize.
Needless to say that this was a huge pitty.
With the information in this thread I see that the attitude to try to build
the new version as fast as possible will maybe result in bad user experience -
So I really should change the packaging process for debian and wait some weeks
after release date before even think about starting packaging to be sure, that
the version is not screwed up. Is that what you want?
No, absolutely not. We are working hard to make every release very good
from day one. The incident I was talking about before was an exception
and not something we calculate or even accept.
Okay the text is polemic I know - but I really think, that hiding the
information that a version is screwed up is not a good idea, because this will
trigger problems downstream. F.ex. I have packaged 2.2.0 and 2.2.1 for debian
already, but not 2.2.2 because form the changelog it sounds like "okay some
bugs are closed" -> will do the packaging, when I have time for. With the
information at this thread I now know, okay I should package 2.2.2 faster,
because 2.2.0 and 2.2.1 are screwed up. ( Or better wait for 2.2.3 ?)...
It was not that the information was hid. There was no intention behind
not telling. It was a fail of certain people (lets say me), happening in
the hardest time of the project so far. And please remember that we
talk about enabling of the auto updater here, which is disabled in
Debian anyway, right?
The version was not so bad that we had to pull it or such. Please do not
overdo it.
But what does all this tell us: We need to communicate all this. Ok, I
apologize again for not doing it.
But honestly, our biggest problem that is the root of this is that the
pre-releases (betas and RCs) are only rarely tested in the community. I
would be happy to learn why not. If that remains that way, that will
probably force us to a different release procedure, which I would find
confusing.
regards,
Klaas
_______________________________________________
User mailing list
[email protected]
http://mailman.owncloud.org/mailman/listinfo/user
