|
|
2011-07-17
, 15:15
|
|
Posts: 38 |
Thanked: 10 times |
Joined on Jan 2010
@ GTA Ontario
|
#232
|
|
|
2011-07-17
, 15:19
|
|
Posts: 38 |
Thanked: 10 times |
Joined on Jan 2010
@ GTA Ontario
|
#233
|
|
|
2011-07-21
, 18:52
|
|
Posts: 35 |
Thanked: 27 times |
Joined on Jun 2008
@ Finland
|
#234
|
|
|
2011-07-25
, 18:22
|
|
Posts: 35 |
Thanked: 27 times |
Joined on Jun 2008
@ Finland
|
#235
|
| The Following 2 Users Say Thank You to Kegetys For This Useful Post: | ||
|
|
2011-08-01
, 08:17
|
|
Posts: 42 |
Thanked: 8 times |
Joined on Dec 2009
|
#236
|
|
|
2011-08-02
, 20:36
|
|
|
Posts: 4,086 |
Thanked: 8,796 times |
Joined on Aug 2010
@ Ruhrgebiet, Germany
|
#237
|
Thanks again for a great app for Maemo!! I don't seem to get hints downloaded (I'm running 0.8.0.7) so should there be a parser update? I can of course check the link but it is not as handy as checking it directly from the AGTL cache description.

The high cpu use started to annoy me so I took a look at it myself. It would seem a lot of the cpu use is caused by the drawing in gtkmap.py __expose_event. This function is called once per second when the gps fix updates and does some drawing regardless if the screen is off, the windows is on the background, etc... I know nothing about gtk/hildon so I "fixed" it for myself by changing the GPS update rate from 1 second to 5 seconds in gpsreader.py. This dropped the cpu use considerably.
Another constant cpu hog appeared to be in threadpool.py. I dont know why but the thread run loop there causes a constant cpu use, I made a crappy fix by calling time.sleep(1) if the _request_queue is empty. I didnt notice this causing noticeable changes in the app usability. I also changed CONCURRENT_THREADS from 10 to 5.
With these changes, the cpu use dropped from >30% to ~3%. These are pretty ugly "fixes", hopefully the author could implement something similar but properly made
). I tried to check Your changes and found, that CONCURRENT_THREADS in osm.py did not change my CPU usage, neither did the sleep in the threadpool.py (and there is no CPU hog in the loop, as there is timeout after poll_interval (5sec). But at first I also thought so.).
|
|
2011-08-03
, 13:54
|
|
Posts: 28 |
Thanked: 3 times |
Joined on Jun 2010
|
#238
|
|
|
2011-08-16
, 08:10
|
|
Posts: 294 |
Thanked: 509 times |
Joined on Aug 2010
@ The Netherlands
|
#239
|
BUT the GPS update interval is the one (so call of gtkmap), which makes the CPU busy. I changed the interval to 2s and I am under 6-10%) and happy
|
|
2011-08-17
, 07:14
|
|
|
Posts: 4,086 |
Thanked: 8,796 times |
Joined on Aug 2010
@ Ruhrgebiet, Germany
|
#240
|
| The Following User Says Thank You to peterleinchen For This Useful Post: | ||
But just for clarification, as it may could be read from Your last post:
all credits go to webhamster, who delivered the most used and loved app on my N900.
I hope, he is keeping his habit providing us with updates for geocaching.com's (bad) attitude to change their website from time to time
THX to webhamster