[01:36] <RAOF> Urgh. Mir now fails its testsuite on ppc64el with a segfault in lttng, but neither Mir nor lttng has changed 😒
[02:20] <soreau> RAOF: what is Mir for?
[02:21] <soreau> I mean what is its intended purpose?
[02:21] <Unit193> Wayland compositor for MATE?
[02:22] <sarnold> I think it's big in kiosk uses
[02:33] <RAOF> soreau: Obviously, it was originally the library for writing display server compositors that Unity8 used.
[02:34] <soreau> RAOF: not so obvious
[02:34] <RAOF> UBports still use it for Lomiri (what they've renamed Unity8 to); we use it for ubuntu-frame, our display-server-compositor for IoTish things.
[02:34] <soreau> so basically a compiz 0.9 replacement?
[02:34] <soreau> in a way
[02:34] <RAOF> soreau: It was originally developed primarily for Ubuntu Phone, with eyes towards desktop convergence.
[02:34] <soreau> since Unity used to run on compiz 0.9 as aplugin
[02:35] <RAOF> Kinda, yeah.
[02:35] <RAOF> Except with the display server folded in, as was the fashion of the time :)
[02:35] <soreau> RAOF: and Ubuntu Phone is a sailed ship, in that it will never see 'the light of day' (in a market sense)?
[02:35] <RAOF> No?
[02:35] <RAOF> We sold a bunch of them!
[02:35] <soreau> but now?
[02:36] <soreau> I mean 'sold' is past tense
[02:36] <RAOF> Yeah, not now. The UBports project has taken over that.
[02:37] <RAOF> Turns out it's hard to break into the phone OS market. Who knew? ;)
[02:37] <soreau> but now we can just run wayfire or anything on Pine64..
[02:37] <RAOF> I mean, yes. But also that's not a product.
[02:38] <Unit193> RAOF: I think Microsoft figured that out. :>
[02:38] <soreau> RAOF: it isn't?
[02:38] <RAOF> (UBports on PinePhone is a kinda a product)
[02:38] <soreau> what defines 'product' in this context?
[02:41] <RAOF> Maybe I mean something like mass-market product? Ubuntu Desktop is a product in that sense - there's a thing plus all the stuff you'd want to do with that thing. applicable to approximately whatever you want to run it on (and there are a bunch of supported platforms).
[02:41] <RAOF> Wayfire on PinePhone is not really that, both in that if you want something other than a PinePhone you're out of luck, and that there's not the same sort of application ecosystem for the generic Linux phone.
[02:42] <RAOF> The latter is changing, somewhat, but AFAICT quite slowly.
[04:12] <soreau> wayfire runs on many more platforms than just pinephone, for the record
[04:14] <soreau> probably could go as far to say that wayfire runs everywhere ubuntu does, if it has glesv2 accelration
[06:14] <RAOF> Oh, sure. But the set of phones that Ubuntu runs on is approximately the PinePhone :)
[06:15] <RAOF> (And probably the librem5?)
[06:15] <RAOF> But that's somewhat off-topic.
[07:43] <RAOF> Harumph.
[07:43] <RAOF> This is weird.
[07:44] <RAOF> There seems to be something broken about umockdev on ppc64el in impish.
[07:44] <RAOF> (As detected by all of MIr's tests which use umockdev dying with stack smashing early in startup)
[07:45] <RAOF> When using 0.15.1-1 from hirsute everything works.
[07:45] <RAOF> But if I rebuild 0.15.1-1 in a hirsute sbuild, the resultant packages exhibit the crash.
[07:46] <RAOF> So something appears to be broken in the toolchain, but I can't tell what.
[07:47] <RAOF> Rebuilding umockdev with gcc-9, or with gcc-11, or with the valac in the archive at the time that the working hirsute packages were built doesn't appear to fix anything.
[08:10] <soreau> RAOF: compare success to fail build logs line by line :)
[08:16] <soreau> maybe it's time to build a new sbuild system :P
[08:20] <soreau> I have an interesting problem myself.. everything works locally but when running wayfire from ppa, somehow, DISPLAY gets set to the value of WAYLAND_DISPLAY but nowhere in the code is this even readily possible
[08:21] <soreau> if I set DISPLAY to 'the correct value', xwayland apps work
[08:30] <soreau> I upload package to launchpad ppa, it says it will start in 6 minutes. Check back 15 minutes later, it says it will start in 33 minutes :P
[08:35] <seb128> RAOF, I just tried to rebuild bolt and it worked without issue on ppc64el which included successful integration tests that use umockdev, https://launchpadlibrarian.net/547442331/buildlog_ubuntu-impish-ppc64el.bolt_0.9.1-1build0.1_BUILDING.txt.gz , so it's not as simple as 'umockdev is buggy' there it seems...
[08:38] <RAOF> seb128: Maybe it's an interaction with lttng (that's what's in the backtrace)?
[19:56] <ricotz> sergiodj, hi, please look for running autopkgtests before requesting a re-try, https://autopkgtest.ubuntu.com/packages/libr/libreoffice/impish/armhf
[19:57] <sergiodj> ricotz: ah, sorry.
[19:59] <ricotz> sergiodj, it wouldn't hurt too much, but those libreoffice runs can take very long
[20:00] <sergiodj> yeah, absolutely.  I was retrying a bunch of stuff and didn't pay much attention to this one
[23:17] <DarkTrick> **Q** @GSettings/DConf management: If I remove the usage of gsettings from my package, will the information within the dconf database in Ubuntu be automatically removed? Or will it stay until an installation script specifically tells it to be removed?
[23:18] <sarnold> "installation script"?