Best posts made by vandys

Is there any way to block a caller (better yet, an entire phone number pattern)? Either in the system or via an app?

If not, consider this a feature request... guess why I woke up at 2AM today.

Update...

During my trip, I had to switch back to my old phone due to a bug in the Python3 notification2 library. _closed_callback() was using an "nid" which wasn't in the notifications_registry any more. I coded some defensive logic and logging, and will see what's going on.

But with that workaround, I have been able to daily drive my Nexus 5 for the second day now. I'm able to keep real time notifications with one 150 byte packet sent and received every 2.5 minutes. It lives fine behind NAT, and also seems fine with the transition from Wifi to mobile data and back.

I just today finished coding up an alternative notification system; since I run my own message and notification server, I can do things like this. It uses a UDP-based protocol I cooked up, and I'll see how it holds up over a trip I'm taking for the next week. I have pop-up display, sound, and blinking LED for notifications. Reading the actual notifications is via a web interface.

Anyway, anybody who's interested:
https://github.com/vandys/webXMPP

The files udp.py and notified.py are the main parts of this work. At its core the server handles being an XMPP client as a proxy on your behalf, it also talks SMS via a VoIP provider for SMS messaging.

Andy

@PINE64 For shipping of batteries, I suggest that you fab a flashlight which uses the same battery. One switch, one LED, minimal plastic case. Then you can ship a battery within a device, national post office regs are satisfied, and the recipient can perhaps crack open the flashlight and use the battery elsewhere.

Re: Caller Blocking?

So I couldn't stand it any more, and remembered an old Nokia N900 hack which sorta did the job. I have a daemon running now which watches for calls on the dbus, and hangs them up if they're in my blacklist config file:

http://sources.vsta.org:7100/webXMPP/dir?ci=tip

You want the tools/blacklist.py file in that project, if you want to check it out.
No, it's not a GUI or a package; if somebody wants to work with me to wrap this
in something prettier, I'd be happy to help.

I couldn't find any way to get word back to my carrier that they're not even worth
a shot in voicemail. But at least they can't ring my phone....

Ok, I got impatient and dug through the UT Tweak tool. Here's a workaround if you're comfortable in a terminal (or, in my case, adb shell):

gsettings get com.canonical.qtmir lifecycle-exempt-appids

For me, this had a value:

['com.ubuntu.music', 'com.ubuntu.terminal']

Then I updated this value:

gsettings set com.canonical.qtmir lifecycle-exempt-appids "['com.ubuntu.music', 'com.ubuntu.terminal', 'morph-browser']"

Voila! My browser is happily running in the background.

@NeoTheThird said in Pointer for those debugging web apps:

@lduboeuf said in Pointer for those debugging web apps:

write an up to date turorial for ubports community then

That would be very much appreciated! Feel free to post something on wiki.ubports.com

Have a look at:
https://wiki.ubports.com/wiki/Web-Browser-app-debugging
and wordsmith it if you can think of nice changes. Thanks!

(BTW, not clear how to link it into the rest of the document hierarchy.)

Is there any chance we could get Ubuntu Touch engineers from Canonical to attend an AMA? There's a ton of mechanism in Touch, and much of the internal design is not well documented. It'd be a great way to give the UBports effort a boost in the right direction if we could pull together a bunch of key questions and have them describe which parts in which packages make the magic happen.

@advocatux, and indeed, no lack of gratitude here for the epic task of catching up with Xenial! Once that one browser glitch goes away, I'm up for trying it as a daily driver.

Ok, dug through the source to obports-installer, looks like most of the goodies hang off of:

https://system-image.ubports.com

In particular:

https://system-image.ubports.com/ubports-touch/16.04/devel/hammerhead/index.json

which tells you where to get bits for various versions. The latest, currently 72, says that there's bits at :

/pool/device-e34756d822387fa9371568fa42292f8baf76d79ec7f2cee0f5c002b909dc08f5.delta-device-cc610cfc131187e182588a0f9c7c5c23ba769679708a4ba08b16878f4e93c101.tar.xz

which, once you pull it and "xzcat device-e3* | tar -xvf -" will give you boot, recovery, and system img files. I booted just to the bootloader and fastboot flash'ed those partitions, and my phone still boots OK. So I guess I'm at least sorta on the right track....

@Flohack FYI, probably just got attacked at the local Cafe. Turned on BT for my external keyboard, but keyboard typing was locked up, then an authentication dialog box popped up. I shut the device down ASAP, will do a clean install since God knows what got scribbled.

No BT in public for me until this is fixed!

I of course mostly develop on my desktop Chromium browser, but every now and then I have a problem which only happens on my UBports browser. I found the file:

.cache/upstart/application-legacy-webbrowser-app-.log

which is the only place I've been able to find Javascript errors.

Also note that the browser is very, very clingy to JS source files, a reload of the overall app will often not update the contents of the JS source referenced from the app.

@UKPhil said in What controls the LEDs?:

If I am reading this correctly, does that mean that the ago old issue of the notification led not working for things like Deko and such like may be resolved soon?

There are two things going on here. One is that the basic push notification service is going away, due to Ubuntu's phasing out of Touch as one of their projects. That's the one I'm working on. I have notifications working well (timely, efficient in both OTA data and battery), but it is an extra task to figure out which apps using old API's might be accomodated.

And then there's the simple issue that it appears the /sys/fs LED files were too protected, so that existing code which knew how to light up LED's couldn't. I understand that will be fixed soon.

I had a hard time finding this! When you open a terminal (or get an adb shell), you're the user "phablet". In the past the default password was "phablet" or "ubuntu". But with the latest UBports it'll be your screen unlock PIN (or, apparently, password, but I used the PIN option).

I'm posting this here in the hopes that somebody looking for this answer will find it quickly!

FWIW, I used pypi.org/project/notify2/ which uses the dbus, worked for my own notification app.

I have a bug unique to the Morph browser. I knew how to invoke the debugger under the old browser--what's the wisdom on this new one? TIA!

@libremax said in Making Ubuntu Touch an easy choice:

The development of UT has to be embedded in a long time, several decades in the same way that GNU/Linux software emerged in the 70s and 90s and only became known to tens of millions of users in the 2000s.

No reason to be impatient only to show consistency and determination.

Well, I don't think Linus was doing his thing in the 70's, but otherwise I agree with your sentiment.

Many kept waiting for the thing or decision or Big Business which would make Linux win. But it really just kept being there, attracting enthusiasm and moving forward. It was just suddenly obvious and inevitable.

So I think the project is doing all the right things. Catch up with a shared SW base. Fix the abandoned browser. Over time adopt a new flagship and, yes, set some old supported devices aside (I'm already a little worried about the flawed binary blobs in my Nexus 5). Keep chipping away at what's missing, what's broken--software and docs. If you keep at it, odds are you'll run across things which will make UT stand out. Especially since the big players can make game changing mistakes (I'm watching Google's tensions with the EU and their increasing control of Google Apps with interest).

Ha, I had the opposite last night. I was running a (music streaming) web app, closed the tab--and the web app kept running, now with no way to stop the music. Had to close the browser app and restart.

The Morph browser is looking really promising for streaming media--it has decent HTML6/ES6 support and its <audio> element handles ogg, mp3, and flac.

But if you change to a different tab, background the browser, or turn off the screen--the whole browser is suspended, making it useless as a media player.

So close, and yet so far.

In the past, the tweak tool could set the browser to not be suspended. As noted on another Morph Browser thread, the tweak tool doesn't "see" this browser.

So... can a UBports wizard suggest how to let the tweak tool see the app? Or name some other way (no matter how grotty) to tell the environment to not SIGSTOP the browser when it's not foreground?

TIA!

But... UI Tweak doesn't appear to know about it! So I can't let it run in BG?