Am 02.11.2004 um 19:10 schrieb Michael Rueger:

> karl.ramberg at chello.se wrote:
>
>> I think we could/should agree on that a rollback of a cs in the 
>> unstable
>> stream should be the very last option to get out of a jam eg. a 
>> update that
>> screw up seriously. People comitting to be part of unstable should be
>> credited for their comitment and we should not pull the carpet under
>> their feet at every twist and turn. We need people working for 
>> enhancing
>
> But, please, lets not forget we are talking about an *unstable alpha* 
> stream here. alpha in my understanding already implies "all bets are 
> off, don't rely on it for serious work", and the word "unstable" has a 
> certain meaning as well.
> We should, however, keep unstable and stable closer together. The gap 
> we got in the 3.8 process was a result from too many things trying to 
> get in at once while we were still sorting out the m17n issues.
>

Yes, the original idea was to have unstable for testing new 
submissions, synched very soon with alpha. So we are talking about
just a handful of changesets and days, not hundrets of changesets and 
months.

This didn't work out in the 3.8 unstable, sorry for that and we will 
try hard to do it better next time.

     Marcus