A week ago, I went to the GNOME Montreal summit. Much thanks to the GNOME Foundation for sponsoring me. Here's a brief summary of my trip:
- Friday: Got maybe an hour of sleep; left home at 3am for a 6am flight. I hate flying. Got to Montreal at around 1:30. Took a cab to the hotel. Took a nap. Finished drafting an email to discuss some AT-SPI refactoring to reduce the needd for synchronous D-Bus calls. Had dinner at a Thai place with Alex and Joanie. Went back to sleep.
- Saturday: Tried to get a Wayland test environment set up but wasn't successful--it seems that gnome-shell-wayland is exiting for some reason. Talked with Alex and Joanie some--we now have a page discussing the accessibility-related work that we need for Wayland. I need to figure out what to do about mouse events--AT-SPI has functionality to monitor them as well as synthesize them, but it requires X. Talked with Jasper a bit over beer; he suggested that mutter could possibly provide API for it, or otherwise I'd probably need to talk to the kernel directly.
- Sunday: Started looking at Mutter and Wayland. Also researched reading the mouse via the Linux kernel. Found /dev/input/mice, but reading it doesn't seem like a good solution--for one thing, it reports the amount that the mouse has moved since the counter was last reset, which presumably doesn't tell me anything about where the compositor thinks that the mouse cursor is on the screen. Need to look more at evdev, though. Met Miguel, who works with an organization that wants to distribute computers with Linux to people in the area who are blind, if I remember correctly; he thought that it would be a good fit because they often have limited incomes. He is interested in improving sbl to add drivers for recent Braille displays or improving brltty to add features present in sbl but missing from brltty, so going to keep in contact with him. He was having trouble getting Orca to read while installing OpenSUSE, which should work as far as I know as long as you're using the GNOME LiveCD, so I plan to follow up with him. Talked to Matthias about a totally unrelated gtk bug that we have a patch for in SLED and need a review for; he suggested emailing Marek and CCing him. Had some poutin(sp?) for dinner, then went to the beer summit.
- Monday: Looked at Wayland some more to try to understand the interfaces relating to input devices. Had lunch at a Greek restaurant. Left for the airport shortly after.
In general, I'd say that I have a somewhat better idea of what should be done for AT-SPI to support Wayland, although the details still need to be worked out. Matthias followed up by sending an email to the wayland-devel list to try to get feedback in terms of how best to port a11y functionality to Wayland. Also, see the accessibility team page discussing the summit.