Hi everyone, below you will find the updates from the Desktop team from the last week. If you’re interested in discussing a topic please start a thread in the Desktop area of Discourse.
We also have our weekly meeting on IRC. We meet on Tuesday at 13:30 UTC in #ubuntu-desktop on Freenode. There will be an “Any Other Business” section at the end where you are welcome to raise topics. These topics might be discussed during the meeting, or afterwards depending on the time, depth of conversation, topic and so on.
Last week’s notes are here: Desktop Team Updates - Monday 23rd November 2020
Slightly shorter week - Took some time off on Friday.
- Subiquity Client:
- Discussion with
@mwhudson on client-server use cases.
- Drafted a couple more complex tests from ^.
- Automatic Theme Snap Installation:
- Began laying a foundation for some automated testing.
- Reviewed and uploaded 7.0.3 to Groovy.
Was a short week because of Thanksgiving holiday.
I was off last week, trying to remember what I did before…
- All of our non-x86 autopkgtest hardware was offline for maintenance, so I babysat taking that down and playing with IS when they brought it back up - testing that the cloud was functioning properly.
- Looked into why the autopkgtest for
mysql-connector-c++ was always being killed. That turned out to be a funny bug. We have a script to look for stale cloud instances (VMs) where autopkgtest died but without being able to clean up its resources. It does this by:
- Getting a list of all the running cloud instances
- Matching those back to currently running autopkgtests
- If there’s no match, the instance is stray and is deleted
That last matching is done by using
pgrep on the process list (not the cleanest, but it is what it is).
pgrep takes a regular expression as the pattern to grep for.
+ is a special character in regular expressions. So we thought there was never a matching process, since the
+ wasn’t being treated literally, so we failed to find the actual running process and ended up killing the VM while the test was still running. It takes close to an hour for this test to run, and we run this cleanup script at
:00 every hour, so it was very difficult for the test to finish. Fixed by escaping the instance name. Fun!
- Investigated why my USB devices were now powering off when they shouldn’t be, turned out to be a bug in libfprint(!), and Marco is already fixing it - yay
- Submitted a bunch more MRs to britney upstream, got feedback and iterated based on that.
- Got access to the InfluxDB for Ubuntu metrics (coming soon), we can start to try to write them now, going to write a simple one soon to get a feel for how it works.
- No further feedback on x-d-p PR #549 (OpenURI fixes for doc portal URIs) after initial review from hadess.
- I also created x-d-p PR #550, that improves the document portal to not proxy some files we know a snap will be able to access directly. With this patch applied, a snap connecting the
home interface should receive real file paths from a portal file chooser if the path is within the user’s home directory.
Ubuntu Core GDM test
- With the dbus activation feature merged to snapd, I did some experimentation trying to use it in the Ubuntu Core GDM test image, with mixed results. Despite having a gadget snap that enables the experimental features, actually making use of them seems to prevent seeding from completing. I’ll need to follow this up with the snapd devs.
Other snapd work: