Message: previous - next
Month: November 2011

Re: [trinity-devel] November: Post Release Meeting 11/15/2011

From: Darrell Anderson <humanreadable@...>
Date: Wed, 2 Nov 2011 10:55:34 -0700 (PDT)
> Hello all,
> Good work team, and thank you to everyone who helped with
> the release! 
> Now that the release is out the door, we are looking ahead
> to the next release.  At this meeting we will discuss our
> goals and our ideas for 3.5.14. If everyone could start to
> thinking about what they would like to do, or see done for
> 3.5.14 that would be great. Please bring your thoughts to
> the meeting.
> Please checkout the following links and contribute as
> such:
> Roadmap for 3.5.14:
> Release Goals for 3.5.14:
> Etherpad for November Meeting:

I still need to create 3.5.13 packages for Slackware, which means I'm not even out of the starting gate yet. :( I believe there is another person who has created packages for Slackware, so I am not anticipating significant problems. My schedule still has not opened as I want. If I am unable to create binary packages real soon now, perhaps that other person can upload Slackware binary packages?

The Trinity Paper Cut project never got officially launched after 3.5.12. I would like to see that project receive official attention. Many bugs were quashed for 3.5.13, but I can't test any bug reports I filed until I build and install packages. I have received only a few notices that any of the bug reports I filed received attention.

I would like to see a relatively dependable 6 month release cycle. Maybe even one or two 3 month cycles until the Paper Cut project significantly reduces the bug list. That means biting only what can be chewed.

The cmake port was unexpected and caused a long delay in the 3.5.13 release. That kind of regular delay will be bad for public relations. If 3.5.14 and 3.5.15 focused on bug quashing and each was released in a 3 month cycle, that keeps Trinity in the news and shows potential users that developers are serious.

I am not asking for a rapid release cycle. Only to consider a couple of quick releases that focus entirely on serious bug quashing.

The 3.5.14 road map implies about 30 days of bug quashing before migrating to GIT. That's good.

I imagine the shift from SVN to GIT will require a one-time massive download to create and initialize a local GIT tree. I imagine the new GIT tree will be about the same size as the SVN tree and will be 4 to 5 GB in size. That translates into a long download session for many people. Is there a way to convert a local SVN tree to GIT without the long download?

Build scripts will need to be modified to support building from the new GIT tree. Therefore there is some work involved for any packager or developer to migrate to GIT. The road map indicates the code repository will remain frozen until the transition to GIT is complete and that is a good idea. The code base should remain frozen until a few of the more talented developers have modified and tested their build scripts for the new GIT tree so they can help others do the same.

The road map indicates "code hacking" will begin after the migration to GIT. Does that mean focusing on enhancement requests in the bugzilla?

Possibly for public relations and keeping Trinity in the news, 3.5.14 could focus only on bugs and GIT and release in 3 months. Postpone enhancement requests and "hacking" until 3.5.15?