Lucid Packaging and QA
Hey everyone!
First off, great job on all of your hard work for what is adding up to be an amazing release cycle. We have began a little work on the timelord project and that has started a bit of buzz. It would be nice if we could find some time in the next week or so to work out the kinks in the project and get it rolling 110%. I know there has been some buzz from people on the Internet who would like to get involved with Project Timelord, but they are saying it is just vaporware at this time. Lets change that! On to packaging. I was attempting to work on some packages for Lucid and was going through our exhausting list of branches in ~kubuntu-members. Maybe something we can think about is trying to figure out how to clean this up and make it easier to view. It would be great if we had just the core KDE packages listed somewhere else, possibly ~kubuntu-dev, because under ~kubuntu-members, there are more than just the core packages, which helps with the confusion a bit. With the packaging I was working on, one thing I come across was branches marked as UNRELEASED. I think this is fine if you just do a few changes or some minor changes, but if you are doing a large amount of changes, or changes that are non-trivial, I think you should release it. KOffice is one. I was looking at the package and the one thing I noticed is the large difference between the package that is in the repositories and the files in the bzr branch. There are changes in the repo packages that aren't in the bzr branch. With the next release of 2.1.1 happening within the next couple of days by upstream, there are more changes that need to be made as well, and right now, unless you have a bit of time on your hand and a lot of patience, it is a fairly large task. QA. Everyone's favorite topic, but a topic that hasn't been carried through with in some time. I was going through some bug reports messing around with a script or 2 from Brian Murray and friends, and something I noticed was prior uploads/releases should have closed out reports. I would greatly appreciate it, and I am sure everyone else would as well, if you could go through the bugs for the package you are uploading, and see if it fixes any of the open issues. If it does, close the bug via the Debian changelog. Granted this will take a little more time, but I don't see why we are constantly rushing to get packages uploaded. Speaking of bug reports, we have a ton, and then another ton that seem stale/outdated/no longer an issue. Seeing as we do have a small developer team, maybe it is time we start doing some Hug Days and trying to get people involved. It is working quite well with KDE, and typically the group that shows up isn't always the largest, but they are all working together and cleaning up the bug lists. We need this badly, and I would be wicked happy to help out however I can. With that said, I know I need to put up or shut up, and would be happy helping lead a Hug Day and what not. With out current development configuration with bzr and all of that, I find it a bit of a pain to really do work that is as efficient as possible. I am trying to get a good workflow going, and that will help me with working on bug reports and fixing stuff at the same time. Another thing I am finding difficult is trying to work on a bug, checking out the source, only for it to change while working on fixes. This kind of goes hand-in-hand with trying to push packages a bit to fast and without much of any QA going on in that process. I found myself rushing a package recently and uploaded it without even looking at the copyright file, which means the copyright file was never touched after doing a dh_make. One thing I have noticed is it seems we have gotten better of not uploading packages that are trying to overwrite another one, but on some things I worked on recently, there was an exhaustive list of list-missing items that should have either a) been included or b) documented in a not-installed file. So, with that said, maybe we should schedule a meeting where we can start hashing out some ideas and plans. Meetings talking about is this going to be default or is that going to be the wallpaper is great, but we really need to start concentrating on some QA work, especially since 4.4 is closing in on its final release. If you have any more ideas, questions, or whatever you feel like replying to, go ahead. If you aren't currently contributing to Kubuntu, but thought about it, now is your chance to follow along and see where you might be able to fit in. Thanks! -- Name| Richard JOHNSON Title| Developer WWW| http://www.ubuntu.com Email| nixternal@ubuntu.com GnuPG| 3578 0981 A21D D662 2A96 7623 F4C1 838C D8C4 4738 -- kubuntu-devel mailing list kubuntu-devel@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/kubuntu-devel |
Lucid Packaging and QA
Am Montag, 11. Januar 2010 00:26:39 schrieb Richard JOHNSON:
> QA. Everyone's favorite topic, but a topic that hasn't been carried through > with in some time. I was going through some bug reports messing around with > a script or 2 from Brian Murray and friends, and something I noticed was > prior uploads/releases should have closed out reports. I would greatly > appreciate it, and I am sure everyone else would as well, if you could go > through the bugs for the package you are uploading, and see if it fixes any > of the open issues. If it does, close the bug via the Debian changelog. > Granted this will take a little more time, but I don't see why we are > constantly rushing to get packages uploaded. Most of the time you do not know which bugs might be fixed by just looking at their subject, so you need to digg in and check (which in case of say kdebase- workspace will take at least a day). So I think in order to carry that out first strong bug triage must be established so that the triagers have time to care about stuff like making the subject reflect a bug's actual root. > Speaking of bug reports, we have a ton, and then another ton that seem > stale/outdated/no longer an issue. Seeing as we do have a small developer > team, maybe it is time we start doing some Hug Days and trying to get > people involved. It is working quite well with KDE, and typically the group > that shows up isn't always the largest, but they are all working together > and cleaning up the bug lists. We need this badly, and I would be wicked > happy to help out however I can. With that said, I know I need to put up or > shut up, and would be happy helping lead a Hug Day and what not. With out > current development configuration with bzr and all of that, I find it a bit > of a pain to really do work that is as efficient as possible. I am trying > to get a good workflow going, and that will help me with working on bug > reports and fixing stuff at the same time. We could always start off with triage-only I suppose? > Another thing I am finding difficult is trying to work on a bug, checking > out the source, only for it to change while working on fixes. This kind of > goes hand-in-hand with trying to push packages a bit to fast and without > much of any QA going on in that process. I found myself rushing a package > recently and uploaded it without even looking at the copyright file, which > means the copyright file was never touched after doing a dh_make. One thing > I have noticed is it seems we have gotten better of not uploading packages > that are trying to overwrite another one, but on some things I worked on > recently, there was an exhaustive list of list-missing items that should > have either a) been included or b) documented in a not-installed file. If I had server resources and what not I would have dropped a QA suite months ago, epsecially the list-missing stuff is pretty easy. -- Harald Sitter Kubuntu Core Developer http://www.kubuntu.org -- kubuntu-devel mailing list kubuntu-devel@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/kubuntu-devel |
| All times are GMT. The time now is 01:26 PM. |
VBulletin, Copyright ©2000 - 2013, Jelsoft Enterprises Ltd.
Content Relevant URLs by vBSEO ©2007, Crawlability, Inc.