02:12 lina: mairacanal: FYI, I started on the dma-fence/syncobj/scheduler abstractions yesterday though it's not pushed yet ^^ (in case it's useful)
06:08 DemiMarie: airlied: what about including the driver, but having it deliberately *not* work unless an environment variable is set at both compile-time and runtime? Or deliberately omitting it from release tags?
07:57 airlied: DemiMarie: omitting from release tags was one idea I had, but i still think some distro will enable it without teeling users
08:02 X512: Isn't is a problem of distros, not Mesa? Some distro may add large patch to Mesa that add new drivers or significantly modify functionality.
08:12 psykose: that is very different from everything already being there and just one toggle away
08:23 jannau: the question is then how large a patch needs to be categorized as different. the rebase pain could be avoided by including the uAPI code but explicitly disabling it
16:01 mairacanal: lina: are you working on the ww_mutex abstraction as well? or do you think it might be unnecessary?
18:21 alyssa: MR label maker tags any MR that adds a new source file as "meson"
18:21 alyssa: I'm not sure that's a super useful behaviour, tbh
18:21 alyssa: I get why it does that but
20:57 jenatali: Label it yourself? IMO the label maker is mainly useful for people who don't have labeling permissions