gossip-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Josh Elser <els...@apache.org>
Subject Re: [DISCUSS] Leave apache and the incubator
Date Thu, 05 Jul 2018 16:50:59 GMT
Hey Ed,

I think taking Gossip back to Github is the right decision. The 
Incubator is designed to support, long-running, low-activity projects 
like you outline. The regular reports on progress become a time-sink 
because there isn't the expected level of volume or participants.

The incubator works well when there is an established community of folks 
who are active. Like you say, this otherwise becomes a burden on you 
(the sole contributor, best as I can see), taking your time away from 
actually producing software.

+1 from me

On 7/5/18 9:18 AM, Edward Capriolo wrote:
> All,
> It was a sizable effort moving everything Gossip into the incubator. Thank
> you all for your help along the way. Recently activity has slowed down,
> much of this falls on me.
> My opinion is gossip would be much better off moving back to github and be
> independently hosted (by me or someone else) I believe for the following
> reasons:
> Infrastructure:
> I forgot my password and when I clicked the "FORGOT MY PASSWORD" link I got
> back a PGP email. No instructions in the email, lol city. I had to google
> around ASFs site to try to figure out what to do. Try googling 'pgp lost
> password' on apache's docs and figuring out what you need to do.
> Releases:
> The release process for most Java projects using maven 'mvn release:clean
> && mvn release:prepare && mvn release:perform". Hosting is free and sign
> takes less then a day. Apache incubator wants to see releases as a sign of
> health, yet the release process is involved. We have to do the maven steps,
> generate an email with the checksums of all critical files, post a vote on
> the incubator list, release the artifacts to central, and copy them to an
> svn directory.
> That is a vote across 2 mailing lists. and all the maven steps, and other
> manual steps, and that does not even count getting the website changed. For
> me what is a 5 minute thing turns into a days long process. The net result
> is we have features in trunk not in the release because doing a release is
> just a drag. No one is even half interested in taking on this process and I
> only did it because it is the only way.
> Community:
> Apache incubator is about building communities.
> Mailing list
> The mailing list is fairly opaque to me. I am sure there is some way to
> figure out what the subscriber base is but I don't know it.
> Jira
> Jira is great tool but the implementation slows people down. New users have
> to sign up, and they are unable to assign themselves tickets until I
> navigate into JIRA and add them to a group. With open source and attracting
> contributors it helps to be able to strike while the iron is hot. Having
> users confused as to weather they can start on a task does not help with
> that.
> Apache has git but not github is only a mirror. When I have to merge
> peoples stuff I have to do it by hand with git commands. (No squash and
> merge button)
> Updating the site:
> Another series of obscure svn commands, making simple things hard, (much
> like the release process)
> Reports
> I get an 4-5 emails at different rates titled "Incubator Report Due" at
> different rates. Only one of them is for this project. We were never great
> with reports. Almost all the info in the report could be automatically
> generated. We missed a report, we got placed onto a report now do every
> month category.
> The last one was due yesterday, sorry I was enjoying a hotdog at a bbq, I
> went to fill it out today and, saw yet another email chain on the incubator
> list about how Gossip should leave apache. The report is just another huge
> time suck, the time I spend doing it I could be doing an hour of code
> review.
> Even though some things are in the incubator 7 years, and some apache TLPs
> have no activity Gossip out of the incubator seems to be a constant thing
> for some....
> So lets get out of here.

View raw message