GPS tracking in Background - Edge Hack (Not Devel, Stable, RC)

[Pulsar33]

Thank you all for this subject !
Is there a description (or source code) of the test_gps command somewhere, please ?
BR
Pulsar33

[doniks]

@Pulsar33 https://github.com/libhybris/libhybris/blob/master/hybris/tests/test_gps.c

[Pulsar33]

Excellent ! Thank you very much @doniks
Best regards
Pulsar33

[ernest]

Latest version has been uploaded to openstore.

Please read prior use : https://github.com/ernesst/Utrack/blob/master/README.md

[aimar]

Hi all,
I think this app is very useful.
However, I can’t make it work.
I use an Aquaris E4.5 with ubuntu 16.04 (OTA11).
I used the previous script and now the recent updated script with no luck.
I have no issues with the installation, but when I launch the script, it gets stuck with “Waiting for satellites ...”.
When I close the app (Ctrl+C), it creates an empty gpx file.
The rest of apps (ActivityTracker, OSMScout) fetch the data and show my position, so I am sure that GPS is working properly.
What could be the cause?
Thank you for your suggestions.
aimar

[ernest]

@aimar
i noticed that the script is neither working with OP3 device. I had to modify it to make it works.
Are you on telegram, if so please contact me @ern_st

[hsce1]

Hi there,
Thank you @ernest for your work. Background GPS tracking is one of the features that I am missing most on Ubuntu Touch!

I tried to make it work on my XperiaX and my Nexus 4 but unfortunately without success. After a short while the following error is returned:

Waiting for satellites

Traceback (most recent call last):
  File "./Utrack.py", line 317, in <module>
    read_gps(p)
  File "./Utrack.py", line 219, in read_gps
    if 70 > ACCURACY > 0 and len(longitude) > 0 and len(latitude) > 0:
TypeError: unorderable types: int() > list()

Utrack_log.txt is empty but in the Utrack_tmp.txt I can see that the satellite data seemes to be logged correctly. The location is correct. I X'd out the exact position:

*** location callback
flags:  55
latitude:       50.XXXXX
longtide:       4.XXXXXX
accuracy:       23.000000
utc:    -758278816

The error appears strange to me since none of the inputs is a list. Do you have a clue what to do?

[ernest]

Yes likely, it has register a list and now it tried to compare it with a int.

I've another version in WIP, that should fix that ...but has other bug. I'll try to find the click.

Edit : Try this one : https://drive.google.com/file/d/1I-_nGDVoFy5hdqiR_ApuA4hKCk8qmqSo/view?usp=sharing

[hsce1]

Hi @ernest!

Thank you for the quick response. Unfortunately I've another error now:

Traceback (most recent call last):
  File "./Utrack.py", line 350, in <module>
    read_gps(p)
  File "./Utrack.py", line 177, in read_gps
    event = event.decode("utf-8")
AttributeError: 'TailEvent' object has no attribute 'decode'

[ernest]

@hsce1 yep i've this thing once every 20 usage...I'll troubleshoot this summer.

[hsce1]

Thanks a lot for your work!!!

[herr-b]

How did you get GPS tracking work? I got the program running in the terminal, but it seems to have no access to the GPS. The symbol in the indicator bar remains black.

In the description there are requirements mentioned (which I can not fullfill):

Depending of the Ubuntu Touch version and AppArmor security, you might need to run this app into an iperspace. See https://open-store.io/app/iperspace.emanuelesorce Currently RC and Stable requires this tool.

What do I have to do with this link/Information? I get a page: The app you are looking for has been removed or does not exist ...

[ernest]

@herr-b You have to allow the terminal to run in background.
This app use test_gps which bypass Lomiri thus the icon color.

You can ran manuall in the terminal test_gps

However this app will likely not work for pinephone and Halium9.

[herr-b]

@ernest Yes I already allowed the terminal to run in the background.
test_gps shows 'init gps interface' and the program shows 'Waiting for satellites' but both get no position, while Pure Maps has GPS reception and shows the position on the map. Foreground or background doesnt matter. It looks like the terminal does not get access to the GPS?
I switched from stable to RC without any change. I use it on a BQ Aquaris E4.5 Ubuntu Edition.

[ernest]

Try to let the phone outside for 5-10 min.
If test_GPS doesn't work for you, then my app will not work.

[herr-b]

GPS reception is not the problem, the app Activity Tracker is working and recording. I switched now the Activity Tracker to run in the background and it seems to keep running in the background. So I can navigate with Pure Maps in the foreground and record the track in the background. But as soon as the screen gets locked all the apps are stopped? Not sure, I will keep testing ... but if I play music in the Web Browser and the screen locks, also the music stops.
It looks that in the case the UT phone should be my only phone I will have to reactivate good old Garmin Foretrex 401 for track recording.

[ernest]

Yes app not in foreground doesn't get GPS signal, see my request there : https://github.com/ubports/ubuntu-touch/issues/1067

[herr-b]

I can't confirm this. I allowed Activity Tracker to run in the background in UT Tweak Tool and started track recording in Activity Tracker. Then I started Pure Maps to navigate and switched apps from time to time. Most of the time I had Pure Maps in the foreground. At the end I had a complete track in Activity Tracker.

So this means, that Activity Tracker has GPS reception and works in the background (if allowed in UT Tweak Tool) in my case.

[herr-b]

Another test showed again the problem with the locked screen. In my test Activity Tracker did not continue recording, once the GPS reception was lost shortly in a building, also GPS reception should have been good again after the interruption. Only after unlocking the screen it continued track recording.

[hsce1]

Hi @ernest,

I just had a look at the Utrack.py and made the following change (line 183), which made the code run on my phone:

if re.search("^longtide", line): #longtide bug in test_gps
    try:
        long = line.split()
-       if re.match(".\d+.\d+", long[1]):
+       if re.match("(?:\d+(?:\.\d*)?|\.\d+)", long[1]):
            longitude = long[1]
        else:
            subprocess.Popen(["truncate", "-s", "0", temp_file], shell=False)

You are trying to match a floating point number here, right? Following this post it is apparently better to use "(?:\d+(?:\.\d*)?|\.\d+)". I've made a pull request.

I hope this does not break other things. On the my phone it works..

Here is the click if someone wants to test:
https://drive.google.com/file/d/1AADkOrVwV-YvTo00XtHnC997e3gRNIi0/view?usp=sharing

Best,
Hendrik