apr-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "William A. Rowe, Jr." <wr...@rowe-clan.net>
Subject Re: APR-Iconv 1.0.2 Released
Date Thu, 17 Mar 2005 21:14:18 GMT
At 03:03 PM 3/17/2005, Paul Querna wrote:
>Jim Jagielski wrote:
>>Cliff Woolley wrote:
>>
>>>>I still think this is an over-reaction as no one operated in bad faith
>>>>here. I maintain that any effort would be better placed at fixing the
>>>>problems and rolling a new apr-iconv 1.1.0 that fixes whatever problem you
>>>>seem to think is present today.  That could likely be done in less time
>>>>with less animosity and less emails.  -- justin
>>>
>>>For what it's worth, I agree with Justin here.
>>
>>FWIW, it does appear to me that the releases did go out
>>quite quickly... Why the rush? 
>
>Tuesday: Tagged, and sent to dev@apr.
>Thursday Night: Took my count of the votes, latter pushed the files to the mirrors.
>Friday Morning: Sent out the announcements.
>
>I didn't feel I did it in a rush. I don't feel it was an excessively fast release cycle.
 Sure, it didn't take months, but I don't think APR 1.0.0 is the best example of a release
cycle.

Three days for considering the opinions of every member of the 
project is not unreasonable, was always the yardstick at httpd,
and was generally the yardstick here.

Security patches are a bit different.  The 'majority' of the new
candidate is considered for three days, but the security fix is
often injected at the very end of the release process, after testing
by the security team.  (There too, a few days, depending on the
attitude of the reporter, and 3 +1's.)

Most members who have been here more than 2 years would consider
the speed at which this was released 'rushed'.  Many newcomers
might not object at all.

>I was partially motivated to complete it before this weekends infrathon. I did not want
to have to deal with this next week.  

Well of course we will be constrained this weekend.  Constrained
to get a release out, but also constrained from fixing urgent
issues that might be introduced by this release.

Right before major changes is not the time to announce a release.

>Since many of the ASF resources would not be available during the weekend, my personal
target when I started the cycle was to release on Friday.  If there weren't enough votes,
or there were other issues, I would of held it up.

There were not enough votes for apr-iconv.  Voters were explicit.

There were other issues identified with apr-iconv.

>I sent an email Thursday night, declaring my count of the votes.

For apr.  No mention of apr-util, apr-iconv.

There are enough votes for apr-util to be released.

>If there was a problem, I would of hoped it would of been brought up then.

That note didn't name apr-iconv.
That note didn't mention 1.0.2

So you expect anyone to be a mind-reader?

>When I counted the votes, I interpreted all of the +1s to be for the entire group (apr,
apr-util, apr-iconv).  I am sorry if I misinterpreted any of the votes.  This was not my intention.

Accepted.

Still 1.0.2 is not released by anyone but pquerna.  Not by the ASF.

>I believe the best course of action is to fix apr-iconv, and then release 1.1.0.

Agreed, after we pull down 1.0.2 until it has sufficient votes.



Mime
View raw message