We CAN NOT use lazy consensus for releases (and we should have made it
more clear that the project had not voted on the RCs when announcing
them on the users@ list).

ASF Release Policy requires three +1 votes from people who have
*installed and tested* the software.

Would everyone please read the ASF release policy (in particular the
"What is a release?" and "Release Management Questions" sections) before
cutting any more tarballs:

http://www.apache.org/dev/release.html

Regards,

Daryl


On 15/01/2010 4:26 PM, Kevin A. McGrail wrote:
> Fyi it's a holiday weekend in the US.  Using lazy concensus now might be 
> better done starting sunday or monday.
> 
> --- Original Message ---
> From:Justin Mason <[email protected]>
> Sent:Fri 1/15/10  3:45 pm
> To:Warren Togami <[email protected]>
> Cc:Justin Mason <[email protected]>,SpamAssassin Dev 
> <[email protected]>
> Subj:Re: PROPOSED 3.3.0-rc3
> 
> On Friday, January 15, 2010, Warren Togami <[email protected]> wrote:
>> On 01/15/2010 11:13 AM, Warren Togami wrote:
>>
>>
>> -1
>>
>> Even though my rc3 was unofficial, it is entirely uncool from packager
>> perspective to reuse names. Names are meaningless and cheap. This should
>> have been named rc4.
>>
>> Furthermore I am unconvinced that we should change the "use bytes" thing
>> at this last minute. It has been going on since early 3.2.x and we're
>> fixing it only at the last moment before 3.3.0? This seems unsafe.
>>
>> Could someone provide a sample message that takes an obscene amount of
>> time?
>>
>> Warren
>>
>>
>> Furthermore, what is releasing an official rc3 at this point going to gain 
>> us, especially if it will take 3 days to become an official release?
> 
> BTW I think I need to clarify.  ANY release will take 3 days from
> tarball cutting to become an official release -- 72 hours to provide
> time for -1 votes.  That is ASF protocol on releases.
> 


Reply via email to