I am sending this to the devel and spins lists. Feel free to forward to
other places people who might be affected by it should see it.
We setup the spin-kickstart project on fedorahosted a long time ago.
At the start it was just the various spins and it's trac instance was
used in the new spins workflow. Then, it was a handy place to put the
dvd iso kickstart that releng made also. Then over time we added cloud
and docker and everything else that used a kickstart to it. In the
start we added some bugzilla components for the various spins for end
user bugs, and used the trac for workflow things. Over time due to
all the images there lots of people have been given commit access to
the repo. Then the spins sig went quiet and we have sort of been limping
along since then with trac tickets not getting to anyone who cares,
I'd like to propose the following plan of action:
* Setup a new project at pagure.io called just kickstarts or
releng-kickstarts or something. (Bikeshed alert).
* Setup tags for all the various groups that have kickstarts. ie,
'xfce' 'docker' 'cloud' 'atomic' 'workstation'
etc. And get someone
from each of those groups to actually watch the tags or someone to CC
on who will actually look at those tagged issues.
* Reduce the number of commiters a bunch and ask people to submit PR's
when they want a change. This will allow releng/qa to control changes
when in freeze. ie, we can require a freeze break and point to the
PR/bug with the exact change and merge it when it's approved.
* Copy the git repo over to it from fedorahosted. Close that repo.
* Mass close all the fedorahosted trac tickets and close trac to new
ones with a note to file new issues in pagure. ( 21 tickets
* Close the "LiveCD*" components in bugzilla to new bugs and close any
existing ones with a note to refile at pagure. (18 bugs currently):
* Adjust koji config to pull from pagure.io instead of fedorahosted.
Thoughts? Additional steps? Better plans?