[09:18] <alan_g> racarr: I think I know what I missed. Working on a quick fix. (But it opens up a whole load of simplifications to qtmir that I'm trying to resist for now.)
=== chihchun_afk is now known as chihchun
=== pete-woods is now known as pete-woods|afk
[09:48] <alan_g> racarr: I've not had time to test and have a hospital appointment - but I've pushed what I think is a fix to port-to-msh-Shell
=== alan_g is now known as alan_g|afk
[10:10] <alf_> racarr: How is the 0.11 release going? Asking because I need 0.11 to be released so I can release some Mesa fixes/changes.
[10:10] <alf_> racarr: (Good morning!)
=== pete-woods|afk is now known as pete-woods
=== sturmflut_ is now known as sturmflut
[14:01] <alf_> racarr: duflu: @mir abi broken, we knew about this and since mir_connection_platform_operation was not (and couldn't have been) in use, we chose to ignore it. What changed?
[14:03] <duflu> alf_: I lost track... ask racarr
[14:18] <racarr> alf_: It was a series of confusions because we had a crash moving a client from .10 to .11 so it was like
[14:18] <racarr> oh ABI must be broken
[14:18] <racarr> so then we ran compliance checker and its like
[14:18] <racarr> ok yeah ABI broken
[14:18] <racarr> then at the end of the day we realized it wasn't in use ( I was mistakingly imaginging the mesa EGL platform using it forgetting the fds were preauthed)
[14:19] <alf_> racarr: yes, all external code (i.e. XMir) still uses the mesa_connection_drm_* functions
[14:20] <alf_> racarr: and mesa doesn't use mir public client functions at all directly