Notices


Reply
Thread Tools
Posts: 169 | Thanked: 149 times | Joined on Mar 2010
#31
I started a blog series about upcoming features in AGTL and Python/PyGTL development on Maemo: http://www.danielfett.de/privat,blog,tag?tag=maemo
 
Posts: 169 | Thanked: 149 times | Joined on Mar 2010
#32
Version 0.6.0 hast just been uploaded to extras-devel. See the full changelog here.

New features include:
* Sun Indicator. Because the N900 doesn't have a built in magnet compass, the display for the direction to the geocache is not always accurate, especially when one is not moving. Therefore, there's a small sun icon in the compass view. If you align the "real" sun with the icon, the direction to the geocache is correct, even when standing still.
* Better Coordinate Input. I played around with various coordinate input alternatives and found that the best one is the one from the legacy AGTL.
* Variable map size. As requested by a lot of users, you can now have the map displayed in double size to increase readability. It looks ugly, though.

And many more.
 

The Following 2 Users Say Thank You to webhamster For This Useful Post:
Posts: 7 | Thanked: 0 times | Joined on Feb 2010
#33
I installed on N900, created an account, created a shortcut on desktop.

When I click on the icon, the applications starts "loading", then quits. Even after a few times, and after switching off and restarting the N900.

Any suggestions on how to deal with this?

Thanks.
 
Posts: 19 | Thanked: 3 times | Joined on May 2010
#34
Thanks for the update webhamster. I just installed it.... but...

Seems that I have the same problems as elinap.

I tried running it on the command line (python core.py --hildon) to see what it was doing, but all it says is

'Segmentation Fault'

Tried restarting the phone.. turned off overclocking etc... no change.
 
Posts: 19 | Thanked: 3 times | Joined on May 2010
#35
Sixth time lucky.... I didn't change anything, but AGTL just started up OK for me. No segfault this time. How odd.

Icons for zoom in and zoom out seem to be corrupted/missing now.. but the buttons still work ok.
 
Posts: 169 | Thanked: 149 times | Joined on Mar 2010
#36
Hm... this bug is hard to debug, but did you try the following?

Open xterm
root
rm /opt/advancedcaching/*.pyo
 
Posts: 19 | Thanked: 3 times | Joined on May 2010
#37
Just tried removing all the .pyo files.... Now I'm back to segfault instead of working app.

I tried starting 10 times in a row... segfault every time. ... didn't seem to be compiling new .pyo files...

rebooted

Tried starting from console 10 times in a row... segfault every time, but noticed that it now seemed to be creating new pyo files.

Deleted all .pyo files (again)

Tried running from console... This time it gave me an error about sqlite3 unable to open database file.

Tried starting it from the desktop again.... Hey presto.. it worked (zoom button images still broken though).

Very very odd behaviour... I don't undertsand what's going on here.
 
Posts: 169 | Thanked: 149 times | Joined on Mar 2010
#38
Very odd, indeed. Had the same problem here and now I'm trying to figure out what goes wrong.
 
Posts: 5 | Thanked: 5 times | Joined on Jan 2010 @ Athens, Greece / Edinburgh, Scotland
#39
Just to add some information so it's in-band (original conversation on Twitter, as @vincula):

I've had the same issue, but in my case, the segmentation faults are 100% deterministic.

I tried from the console, running with:

Code:
python2.5 /opt/advancedcaching/core.py --hildon
This coincides with how the .desktop file (/usr/share/applications/hildon/advancedcaching.desktop) executes the program.

I removed the .pyo files, with no change. I tried (half-heartedly) to cause a core dump with no luck, and also realised there's no strace/ltrace package. Around that time I gave up as it had been a very long day already.

I'm available if you want to have me try any experiments. I'm not a Python virgin, and, intuitively, this seems like a Python (or related bindings/library) issue, not a AGTL issue. You shouldn't be able to cause a segmentation fault of the Python interpreter, no matter how bad the code.
 
Posts: 169 | Thanked: 149 times | Joined on Mar 2010
#40
Originally Posted by alexios View Post
I'm available if you want to have me try any experiments. I'm not a Python virgin, and, intuitively, this seems like a Python (or related bindings/library) issue, not a AGTL issue. You shouldn't be able to cause a segmentation fault of the Python interpreter, no matter how bad the code.
You are absolutely right. I think I found the bug, but did not analyze what was the real issue. I used a mixin at some point in the code and this caused the segmentation fault. To get rid of the bug, I switched to a static class now.

Preparing a 0.6.1 release, will upload it any moment now.
 
Reply

Thread Tools

 
Forum Jump


All times are GMT. The time now is 10:43.