Galago

I finally, after a long 8 months of development, put out Galago 0.5.0. The list of changes are huge, but I feel it’s a pretty solid release. I’ve been wanting to get this release out for a while now, but I kept finding one more thing to fix. I finally bit the bullet and prepared the release. Oh, and this one includes Python bindings!

I have a news post at the Galago site talking a bit about the release and linking to release notes.

I feel like a weight has been lifted from my shoulders. I’m going to sleep well tonight.

I discovered last night that galago.info had expired just a couple of days ago. I was a bit shocked by this and looked through my inbox for expiration notices, but could find none. They may have been caught by the spam filter :(. So, it’s in this state where it’s unavailable for everybody (myself included) for 35 days. After 35 days, it’s available for purchase again. I could pay $90 to get it back now, it turns out, but that just seems too much for the domain.

So, I’ve moved to galago-project.org. Please update your bookmarks and check out the new svn repository. This covers all galago work and all notification work. Sorry for the inconvenience. It sucks.

The release of Beagle 0.0.12 rocks. Among the many reasons is support for Galago, provided by Lukas Lipka. This just kicks ass. Simple, informative, useful integration that doesn’t get in your way. Now if some other IM clients will step up and provide at least optional support for Galago, non-Gaim users will be able to get this functionality as well.

So over the past few hours, I managed to put together a new Galago release. There was a lot done in this release, and I wouldn’t be surprised if I broke stuff, but at least now it should work with both D-BUS 0.23.x and D-BUS 0.3x. I have neat things planned for the next release, but as I’m basically going to be away from computers for a week, spending time with Jamie, I probably won’t get to any Galago hacking.

Releases of this magnitude take forever.. There’s just so many components to package up. I need to get a small release team together. Any packagers want to package things up for Fedora Core or some other distro?

I also need to find a sane way to announce releases for several independently versioned apps and libs at once. Shame I can’t really give them each individual release names without being too confusing. Hmm, and I need to get screenshots of the contact chooser dialog up. Lots to do still.

The Motley Fool had some great 401(K) advice. They say that if your 401(K) lets you invest in llamas, consider moving to a job where the owner is not insane, as llamas cannot possibly help your retirement. I think they understimate the power of the llama.

Galago with D-BUS 0.30

I’ve been hacking a bit on Galago’s D-BUS 0.30 support the past couple of days, and although it’s not yet complete, good progress has been made. Most things now work, with the exception of a few bugs here and there.

Thanks to D-BUS 0.30’s new struct container support, I can really start to clean up the protocol. However, for the time-being, not much protocol-wise has changed. I’ve been developing a small abstraction layer to minimize the number if D-BUS versioning checks in the code. You see, I’m not dropping D-BUS 0.23.x support just yet. Not until more desktops have 0.30+ installed.

I do all development nowadays on my laptop, which runs Ubuntu Hoary (containing D-BUS 0.23.x). I needed to be able to quickly and easily move back and forth between this and a distro with D-BUS 0.30.x support for testing, and I certainly didn’t want to set up a new distro on one of my other computers.

I ended up installing Fedora Core 4 test 3 inside a VM using VMware Workstation 5. It’s working beautifully, and just sits on one of my virtual desktops taking up a good portion of the screen. The theme and everything matches, so for the most part it doesn’t even feel like I’m using another virtual computer. I’ve been able to make a D-BUS 0.30 change, submit the code, switch virtual desktops, and make sure the 0.23.x support still works. And vice-versa. It truly makes this kind of development a lot easier to deal with.

Also, since I’ve been using my laptop and taking that to work with me, I’m able to just click Suspend on the VM when I need to power down to travel with it. Gotta love that. I wish I had suspend working on my laptop itself.

I’ve been hacking away at various pieces of Galago lately at an attempt to improve the API and Gtk+ widgets. This is leading to some very cool applications and utilities I’m developing. Hopefully if I can get some of this set up the way I want soon, I’ll get another release put together.

A wiki has been put up containing some installation documentation, a preliminary FAQ, and other documentation. It’s also hosting the wiki for Project Soylent, which we’re starting to develop some decent plans for (they’re not on the wiki yet, just in some discussions).

My latest cool Gtk-related Galago work is the Contact Chooser. I love this thing, seriously. It’s not finished yet, but is close, and is simple and easy to work with. The screenshots are fairly self-explanatory.

If you haven’t guessed, it’s a widget and dialog that resembles the Gtk+ file chooser, but it’s designed to select contacts. The gnome-presence-applet will be using this fairly soon. I need to work on some of the speed issues, but they’re not too major.

libgalago got some major code cleanups these past couple of weeks. A lot of the communication code was abstracted, and as a result, several hundred lines of code have been removed. This will dramatically ease the porting effort to D-BUS 0.33. I plan to keep compatibility with D-BUS 0.23.x as well. I’m not sure when exactly all this will be done, but definitely before the upcoming stable release.

I’ll post some screenshots of the design we’re considering for Project Soylent once we have something I feel is worth demoing.

I felt especially motivated today and started work on an API I came up with in the shower. I spent hours hacking on libgalago and, when I was done, I had reduced the codebase by over 800 lines, all of which consisted of D-BUS 0.23.x-specific communication code. There’s now a much cleaner API abstraction over D-BUS, which could easily in the future be expanded to not be D-BUS-specific, if that ended up becoming important (say, a Windows port of MacOS X port without D-BUS). Furthermore, it should now be much simpler to get Galago working cleanly with D-BUS 0.31+. I’m not going to be able to work on that right away, but it’ll happen sometime before libgalago 0.4.0.

The abstraction still has some work to go, but most of the codebase for libgalago has moved to it. I still don’t know what I’m going to do for galago-daemon’s D-BUS 0.31+ support, but it’s probably not as huge an effor as I’m thinking. Worst case, I do some more abstraction work, but it won’t be as bad as libgalago’s.

Also, I hacked up contact-lookup-applet a few days ago to check for libgalago at compile time and to use Galago for the Instant Messaging section of the Address Card dialog. If Galago’s dead due to some bug, it falls back on the default implementation of just showing a protocol icon. There is a patch available for now, and I’m hoping that perhaps support can be added to CVS at some point. Next stop, gnome-present-applet UI work and Evolution.