Hi all! This update comes a little bit late, because I’ve been busy with XDC 2018 last week. I also wanted to publish at least one non-status-update blog post before this one, but… Erm, well, let’s talk about XDC!

XDC is a conference organized by X.Org. It’s not really about the X11 server, the scope of the event is broader (in fact most people there were into graphics drivers and lower-level stuff).

I met there Drew, Markus, Scott and Guido who are all members of the wlroots crew. It was good meeting them, apart from Drew it was the first time I’ve seen them! I discussed a lot with other XDC attendees too, leading to me realizing we need to improve our DRM backend. Some discussions were very helpful, because people took the time to explain how to handle faulty DisplayPort cables or if our current damage tracking implementation was correct. We also talked about Multi-Stream Transport, hotplugging graphics cards and FreeSync which would be cool features to have in wlroots (we should already support MST, but I’ve been told it needs a few fixes to work properly).

Another discussion was about presentation-time, a Wayland protocol to provide feedback about the exact time at which a frame has been displayed to the user. This is useful for games, video players and medical software, all of which require very precise presentation timestamps. Talking with actual people needing this protocol helped to make me implement it on the plane back to France, because it suffered from a chicken-and-egg problem: no one was using it because it wasn’t available outside Weston.

There were also some interesting talks. One was about VKMS which would allow us to test our DRM implementation more easily by emulating a graphics driver. Others included the Intel CI for Mesa, a rewrite of the Intel driver (to use Gallium) and an introduction to Intel assembly language for graphics cards. Cool stuff!

Apart from XDC (and earlier in the month), I’ve been trying to make wlroots work well in case more outputs are connected than supported by the graphics card (ie. there are more enabled connectors than CRTCs). Starting from Not Working At All™, I finally got everything working on the wlroots side, though some more work is needed on the sway side to make everything smooth. I also started to work again on kanshi, which I use to configure my outputs. We also managed to merge Laaas’ pointer-constraints pull request. With the work-in-progress relative-pointer pull request, this should allow us to have a way better gaming experience on sway!

We also have good progress for mrsh. Drew and other contributors helped a lot this month, thanks! I almost completed the parser, I just need to add in arithmetic expressions support. The AST is now annotated with more precise positions and the highlighter example has been improved. A lot of builtins are now supported, including cd, ., eval, shift, export, pwd, true. The shell supports subshells and for/while loops. We also have some harness tests (and the beginning of conformance tests). We’re pretty close to having a usable shell here! There’s still a lot to be done, so if you’re interested, ping me on IRC or by email.

Damn, this status update is getting pretty long. Let’s mention go-openpgp-hkp, a Go library I’ve created to create OpenPGP HTTP Keyserver Protocol clients and servers. I’ve been pushing fixes to go-dkim, its behavior is now in line with the existing DKIM tools (even if it’s not strictly in line with the RFC).

All right, that’s enough! See you in a month (or less if I somehow manage to publish another article by then), and if you have questions/comments send everything to my public inbox.