[09:06] <alan_g> tjaalton: yes, it looks the same. A guy we both know has offered to help (but he's not up yet). That may be quicker than trying to learn it ourselves.
[09:44] <alan_g> tjaalton: I've just spotted your name on the SRU Wiki. I'm trying to update Mir in Xenial so that IoT can dispense with the overlay PPA (bug 1685186). Is this something you can help me with?
[10:27] <tjaalton> alan_g: I can ack stuff on the queue, yes
[10:33] <alan_g> tjaalton: that sounds like a step forward. Thanks.
[10:35]  * alan_g is suffering from xml (in the mesa code)
[10:44] <alan_g> tjaalton: I think src/intel/vulkan/anv_entrypoints_gen.py should be generating the function declaration, but I need to figure out how it is supposed to work.
[11:13] <alan_g> tjaalton: I'm still not set up to try a full build, but you need 'VK_KHR_mir_surface' in SUPPORTED_EXTENSIONS in src/intel/vulkan/anv_entrypoints_gen.py
[11:16] <tjaalton> alan_g: oh, I'll try that out
[11:17] <tjaalton> thanks
[11:17] <alan_g> np, someone needs to grok this stuff. I have to start somewhere.
[11:21]  * alan_g is wondering which (if any) laptop deserves Artful at this time
[11:24] <tjaalton> alan_g: new errors this time :)
[11:25] <alan_g> Nice!
[11:25] <tjaalton> vulkan/anv_entrypoints.c:712:5: error: unknown field ‘CreateMirSurfaceKHR’ specified in initializer
[11:25] <tjaalton> etc
[11:27] <alan_g> Well, that's protected by #ifdef VK_USE_PLATFORM_MIR_KHR - so let me see where that should come from...
[11:28] <alan_g> Actually, no. it's a void* otherwise
[11:30] <tjaalton> -> afk, will check back in a bit
[11:33] <alan_g> tjaalton: does struct anv_dispatch_table in your anv_entrypoints.h not have a CreateMirSurfaceKHR member? Running that python script here generates it OK, maybe you just need to clean it?
[11:52] <tjaalton> perhaps I shouldn't be running sbuild but a chroot
[11:59]  * alan_g doesn't need to cross build often enough to learn the tools.
[12:03] <tjaalton> uh, chroot build passes fine
[12:05] <tjaalton> anv_entrypoints.h looks fine
[12:05] <tjaalton> which is hardly surprising at this point
[12:05] <tjaalton> anyway, guess it's on me to figure out why the package build fails
[12:05] <tjaalton> using the upstream tarball & autoreconf
[12:23] <tjaalton> yeah, the tarball comes with anv_entrypoints.c/h so they're not regenerated
[13:06] <alan_g> tjaalton: I'll leave you to deal with that then
[13:07] <tjaalton> removing just those two didn't help :(
[13:07] <tjaalton> oh well
[13:22] <tjaalton> ha, typoed the removal, works after all
[13:23] <alan_g> \o/
[14:11] <alan_g> tjaalton: abut the SRU bug above. Who (that's left) can get it targeted to xenial?
[14:22] <tjaalton> alan_g: anyone on ubuntu-bugsquad or what was it called.. if you mean the technical bit of modifying the bug to show that
[14:24] <tjaalton> as for allowing the update to enter proposed, maybe get some buy-in from #ubuntu-release first
[14:24] <tjaalton> then anyone on the sru team can be more confident in letting it in
[14:37] <xnox> alan_g, what's the bug number?
[14:38] <alan_g> xnox: bug 1685186
[14:39] <xnox> alan_g, the package should land into unapproved queue.
[14:39] <xnox> alan_g, meaning somebody needs to publish the silo.
[14:39] <xnox> then ubuntu-sru team will review the unapproved queue
[14:40] <sil2100> I can potentially take a look at if if needed
[14:40] <sil2100> *it
[14:40] <sil2100> This will need an archive admin for final approval anyway
[14:42] <ogra_> how do you make sure the mir-kiosk based snaps keep working ? (snapcraft will have used what was in the archive at the time of the snap build to link against libmirclient)
[14:43] <ogra_> i.e. is 0.26 fully backwards compatible so it doesnt break existing deployments out there ?
[14:48] <sil2100> Seeing the number of ABI bumps, that would be 'no'
[14:48] <alan_g> ogra_: 0.26 is what is in the stable phone overlay
[14:48] <alan_g> and the "mir snaps" are built with that
[14:49] <sil2100> Anyway, this needs to first go through an archive-admin anyway, I guess it's best if the review is done by someone with the proper powers
[14:49] <ogra_> alan_g, even the client apps ?
[14:49] <alan_g> Getting 0.26 into xenial means we can stop using the overlay
[14:50] <alan_g> ogra_: client apps should be using mir-libs
[14:50] <alan_g> And that uses overlay
[14:51] <ogra_> alan_g, ok
[14:52] <alan_g> xnox: @"somebody needs to publish the silo." - how do I make that happen?
[14:57] <xnox> alan_g, get an Archive Admin to review it on e.g. #ubuntu-ci-eng due to ABI bumps. For example, I would recommend pinging slangasek but I'm not sure if he is available today.