Hi, I can’t imagine no one has noticed this yet but couldn’t find anything about it yet. Feel free to merge/move this to where it belongs

Ever since updating to 16.04/devel, once libertine started working again, it works differently from how it was in 15.04.

It used to open everything an app did in the same window. Even new apps started by an app, like when you launch gedit from terminal. This was cumbersome, but the new situation is the other extreme: everything, tooltips, drop down menus, right-click menus, dialogue boxes, everything opens in a new window. Some middle road must be able to be achieved?

See also the github issue I made for it. https://github.com/ubports/libertine/issues/10

Does anyone have this problem too, and is there any way to fix it? Is this an xmir/mir/Xorg thing?

Hi guys,

I'm trying to figure out how I can dualboot my Nexus 10 with an image from Ubports

I currently have a dualboot through Canonical's Dualboot Installer app in Android, running a very old an discontinued version of Ubuntu Touch (April 2015 i think) as described here: https://wiki.ubuntu.com/Touch/DualBootInstallation

learning about Ubports' efforts to make a working Nexus 10 image, i wanted to have a looksee

The abovementioned site describes a way to use a custom image server (in this case Ubports' server) but i can't get it to work.
This is my preferred way, because it preserves my current Android installation.

so my question is:

Can anyone help me with the right way to point to Ubports' server, or tell me about better ways to dualboot (and preferrably keep my Android setup intact)?

This is what i did up to now:
I took the server location from Ubports' install script:
sudo ubuntu-device-flash --server=http://system-image.ubports.com touch --channel=ubuntu-touch/stable --device=manta --bootstrap

and pasted it into the command found on the wiki:
adb shell "su sh 'cd /data/data/com.canonical.ubuntu.installer/files/ && echo $CUSTOM_SERVER_URL > custom_server && chmod 777 custom_server'"

creating:
adb shell "su sh 'cd /data/data/com.canonical.ubuntu.installer/files/ && echo $http://system-image.ubports.com > custom_server && chmod 777 custom_server'"

the custom_server file is created, but the installer doesn't give me any new channels to install.
Am I just missing some things, or am I just embarrassing myself here?

@flohack

Here's the log for dev:
pebbledumpdev-summary.pdf
pebbledumpdev.pdf
pebbledumpdev.log
bluetoothctl terminal output

and here's the log for stable:
pebbledumpstable-summary.pdf
pebpledumpstable.pdf
bluetoothctl terminal output

stable connected fabulously , from what i could see, rc and dev failed in the exact same way...

@moem
from what little i understand, i think a specific filesystem overlay would do what we want, provide a cusom bluez stack on one device while leaving the image for that port intact.

someone would need to create that overlay and maintain it, for every affected port, test it against new OTA's and adapt it accordingly, and the user would need to implement the overlay in their device after every OTA manually.

oof

@flohack

that is sad news for the few UT+pebble users. but understandibly if it's great news for a lot of new devices then i see where the priorities are, and rightfully so.

at least we learned something. thanks for your time

@moem @Flohack

sure, i'll stay on rc or maybe move to dev to help test or dig up logs that may be of interest, as long as my phone keeps working. i just got a 'project' of sorts today that needs me to be on call for the coming week (given your stated schedule that shouldn't be problem )

just lemme know if i can help

@flohack

Thanks for your update.
Good news for Pebblers I guess, but yikes thats a setback... sorry to hear that.

I just installed my Pixel 3a (stable) OTA-24 and the BT connection with my Pebble survived. Is that because of the delay, or was OTA-24 safe anyway?

is it usefull to get a new log on OTA-24?

@flohack

Here's the log for dev:
pebbledumpdev-summary.pdf
pebbledumpdev.pdf
pebbledumpdev.log
bluetoothctl terminal output

and here's the log for stable:
pebbledumpstable-summary.pdf
pebpledumpstable.pdf
bluetoothctl terminal output

stable connected fabulously , from what i could see, rc and dev failed in the exact same way...

@flohack i already did the btmon+wireshark thing for RC channel. the output is gibberish to me so it would be great if someone with more skills could look at it.
https://forums.ubports.com/post/67758

My latest question was if the same btmon+wireshark output for dev and/or stable would be useful to generate. if so i'll happily do that.

@klh sorry, life happened and i lost track of this thread. my pixel 3a is my main device so while i can switch channels between stable/rc and dev, 20.04 may be a bridge too far

@Flohack i did switch it to dev now. is that of any use for testing and logging bluetootctl?

i also have a bq aquarius M10 HD but its dev channel is apparently stuck on 30.06.2022 and i dont suppose it has a 20.04 channel as it is a legacy device?
i have a One Plus One with a broken screen, can't be operated...
then i have a Fairphone 2. is that of any help at all?

(what a sad bunch if i put it like that, huh?)

@moem
from what little i understand, i think a specific filesystem overlay would do what we want, provide a cusom bluez stack on one device while leaving the image for that port intact.

someone would need to create that overlay and maintain it, for every affected port, test it against new OTA's and adapt it accordingly, and the user would need to implement the overlay in their device after every OTA manually.

oof

@flohack

that is sad news for the few UT+pebble users. but understandibly if it's great news for a lot of new devices then i see where the priorities are, and rightfully so.

at least we learned something. thanks for your time

@flohack

Did you have a chance to look at it?
I've gone and tried wireshark.
i have no idea what i'm doing lol, but i managed to print out a packet list (both summary and all written out, one packet per page) of what i think are two complete cycles of finding, trusting, pairing, connecting and sadly disconnecting the Pebble.

i hope it helps

pebbledump summary.pdf
pebbledump.pdf

@flohack
pebbledump.log

pastebin of terminal output(as far as i could scroll back and copy)
i hope you can make sense of this?
the device is 'Pebble Time 3247, B0:B4:48:A3:32:47'. there's also a 'Pebble Time LE 3247' but that's a (useless) limited mode that should be ignored.

@moem @Flohack

sure, i'll stay on rc or maybe move to dev to help test or dig up logs that may be of interest, as long as my phone keeps working. i just got a 'project' of sorts today that needs me to be on call for the coming week (given your stated schedule that shouldn't be problem )

just lemme know if i can help

I'm gonna +1 this

device: Pixel 3a
OS: Ubuntu 16.04 (2022-W39)
Channel: RC

connecting my Pebble Time Steel always was a bit iffy through System Settings, but using bluetoothctl in terminal gave me more control but still fails at connecting:

$ bluetoothctl
$ discoverable on
$ scan on
$ agent off
$ agent NoInputNoOutput (this to make sure no verification code is asked. the watch doesn't expect one)

--wait until you see your device pop up in the scan list --

$ trust [device mac]
$ pair [device mac]

--pairing should pop up a confirmation on the pebble. press top right button to confirm. terminal reports success--
so far so good, but then the last command gives trouble:

$ connect [device mac]

--it connects... for about three seconds and then drops off.--

before the update, the pebble, once paired, had no trouble connecting and reconnecting. i could walk out of range of my phone and back in range and it would just pick up again.

Other devices (BT mouse, car's integrated hands-free set, windows laptop) seem to have no trouble.

/edit:

I realize all the devices i tested have one thing in common: they are all much newer than the Pebble smartwatch.
Could it be that an old protocol or some such was removed from the bluetooth stack, causing older devices to fail?