Lately, I've felt a lot of frustration and unhappiness, and that has affected me in a lot of different ways. I'm trying to take steps to turn that around, and one of them is to try to write more about what I'm working on or thinking about.
== Fedora 19 ==
Now that the Alpha is done, I'm in the process of updating my laptop to Fedora 19. Wish me luck!
== Flock ==
I've been working closely with Ruth and Robyn to try to put together a new conference for Fedora Contributors, called Flock. We were waiting on the Red Hat budget to finalize before we started, then when that happened, we found out that there was a good chance that we'd have a very nice conference venue donated to us, so we started doing some very initial planning around that. Unfortunately, that venue wasn't applicable over a weekend, so we couldn't use it. We evaluated several sites as possibilities, one of which was Charleston, SC, which was especially compelling because they were also willing to let us use them as a venue space at no cost.
I spent a lot of time looking at how other Linux distributions run their events (Debian, OpenSUSE, Ubuntu), and trying to base Flock in that model.
One of the main goals of Flock is to bring together as many of our contributors as we can to learn, hack, and improve Fedora. As part of this goal, we want to use as much of the budget as we can to cover travel costs for contributors. We're going to need sponsors to help us do this (Red Hat gave us a generous budget allocation, but it will only go so far). I've been working with Ryan Lerch to make a prospectus PDF, and he finished it up last night. We might still tweak it a bit, but the meat should be fine. Here is a link to it: http://spot.fedorapeople.org/flock/flock-2013-sponsor-prospectus.pdf
We had a pretty productive IRC meeting yesterday to discuss Flock, and we setup two mailing lists (flockinfo
, for low-traffic announcements about flock, and flock-planning
, a high-traffic mailing list for planning the event).
I talked a bit about the format of the conference, where things are mostly scheduled in advance. Luke Macken has been making an improved registration app that can also handle submissions for talks (or other scheduled things like sprints), based on the work that Ian Weller did for FUDCon Lawrence. I also talked about how we wanted all the talks to be live-streamed, with volunteers coordinating questions from IRC and running a clock.
I also brought up the idea of having a "Social Outing" day at Day 3 (the middle of the schedule driven conference), where the Flock attendees could go do something fun together and build new friendships. People really seemed to like that idea a lot. I can't take credit for it, I stole it from DebConf. :)
Several suggestions came out of that IRC meeting that were very useful (and that I like a lot):
* tatica suggested that we have a scheduled split between talks (45min-1hr) and workshops/sprints/mini-hackfests (2 hrs).
I liked that idea, and proposed that we use the morning for talks, then lunch, then after lunch, scheduled workshops/sprints/mini-hackfests. We already wanted to block off two days at the end for unscheduled hacking, this is in addition to that.
* j_dulaney pointed out that we need good quality audio for streaming talks, something that didn't happen very often in Lawrence.
* inode0 suggested that it would be better if the scheduled part of Flock ended on the weekend, so we are strongly considering
August 7 - 13, 2013 for the dates (August 12 and 13 would be optional "hackfest days" with no assigned schedule in advance).
Because we wanted to get something moving on Flock quickly, I wasn't as open or transparent as I could (and should) have been on what was going on. I'll definitely try to do better on that going forward. From here on out, everything is happening on the mailing lists and IRC meetings. :)
== LibreCAD ==
I packaged up LibreCAD
(dependency) for review in Fedora. libdxfrw got reviewed this morning, and it has some items I need to fix up.
== R ==
Over the last several days, I've been working on updating R
to 3.0.0 in Fedora. Normally, my position on updating R is that I push updates for new stable releases to all stable branches of Fedora (and EPEL 5 & 6), but with 3.0.0, because of core changes, it requires that all R modules be rebuilt against R 3.0.0. Initially I was going to just do this for rawhide, but then I thought it might be good to do it for Fedora 19 before it releases. I did EL6 too (there are not very many R modules packaged for EPEL, so it wasn't too hard).
In parallel, there is also a bug in the "kmeans" function in R that wasn't fixed until 3.0.0.https://bugzilla.redhat.com/show_bug.cgi?id=948083
As a result, I'm going to push R 3.0.0 to Fedora 18 too, which will also require all R modules in Fedora to be rebuilt. This is trickier than it sounds, since most of these modules are dependent upon each other. In doing the initial round of rebuilds, Pierre and I only found one dependency loop (R-GenomicRanges and R-IRanges), which I fixed with a conditional bootstrapping. R is very good about not having these sorts of circular build dependencies, but when you run the R check routines, they often check for things that are not otherwise dependent for the R module to build. This just means we need to conditionalize some BuildRequires as being only needed for the %check section.
Also, since we cannot chain-build (automatically tag builds into the buildroot, then kickoff dependent builds) in any branch except Rawhide, I have to manually kick off a build, then tag it, and then kick off another build. Takes a while. :)
I also want to rebuild things for EL5, but I can't get R to build in the EPEL-5 buildroot, seems like something might be broken with the EL-5 builders, I need to check into that.
== Enlightenment ==
I'm helping Rahul update some of the core E components to their current releases in Fedora 19 and Rawhide.
== Sponsorship ==
I sponsored a new packager into Fedora this week, Lokesh Mandevar. He packaged up spectrwm
== OpenCASCADE ==
I reached out to the OpenCASCADE upstream and asked them to consider:
* Dropping the wording in their existing license which makes it non-free
* Relicensing to a standard license to make OpenCASCADE GPL-compatible (LGPL is the obvious one, but there are plenty of others).
I got a good response from them, they're going to look into it. I hope this results in a positive outcome, because that codebase is pretty interesting and useful for CAD and 3d printing tools.
== Alien Arena ==
I contacted the Alien Arena upstream and asked them if they would be willing to drop the commercial distribution restriction on their content, and they were willing. They couldn't drop the requirement that the alienarena content not be distributed independently, but it isn't separated in Fedora, and that requirement shouldn't be a problem for us. (The reason they can't drop that requirement is because they purchased licenses for some of their content.) It hasn't happened yet, as far as I can see, but it is a good sign of intent.