[15:35] <oSoMoN> hey all
[15:35] <oSoMoN> I’m looking at this weird chromium/libreoffice interaction issue, where symbol mir_display_config_release somehow comes into play
[15:36] <oSoMoN> I could use some help in understanding it
[15:36] <oSoMoN> https://bugs.launchpad.net/ubuntu/+source/chromium-browser/+bug/1710789
[15:38] <greyback> alan_g: ^
[15:38] <alan_g> looking...
[15:39] <oSoMoN> from my tests in VMs, the issue happens only on amd64, not i386
[15:39] <oSoMoN> and zesty and artful are affected
[15:39] <oSoMoN> trusty and xenial are fine
[15:44] <alan_g> mir_display_config_release was introduced in Mir 0.21 - xenial originally had Mir 0.17 - so I suspect there's an old libmirclient on the system
[15:46] <alan_g> oSoMoN: could you $ apt show -a libmirclient9 on an affected system?
[15:48] <oSoMoN> alan_g, http://paste.ubuntu.com/25326372/
[15:51] <alan_g> Oh! Look what I find:
[15:51] <alan_g> /usr/lib/chromium-browser/libmirclient.so.9
[15:51] <alan_g> /usr/lib/x86_64-linux-gnu/libmirclient.so.9
[15:51] <alan_g> Is chromium-browser shipping its own libmirclient?!
[15:52] <oSoMoN> alan_g, yes, see https://cs.chromium.org/chromium/src/third_party/protobuf/README.chromium?q=libmirclie&sq=package:chromium&l=50
[15:56] <alan_g> I guess the stub needs syncing with more the recent symbols.
[16:00]  * alan_g thinks there must be a better way
[16:14] <alan_g> oSoMoN: it's a guess, but I suspect that soffice.bin (or its dependencies) needs the real libmirclient, but gets the stub
[16:21] <oSoMoN> alan_g, ack, I’ll dig further into this direction
[16:31] <Saviq> https://developer.ubuntu.com/core/examples/snaps-on-mir is finally back up
[16:34] <alan_g> \o/
[16:45] <alan_g> bschaefer: AlbertA anyone for a design discussion? https://code.launchpad.net/~alan-griffiths/miral/pre_init-CommandLineOption/+merge/329098