Notices

hehe ok.
Has been working fine with the older version btw.
Gotta wait then.

//edit: just tested, it's working now. Thx

Hi

big thanks for that great app!

I have found two bugs.

to reproduce the first one:

1. start fahrplan
2. From: Station, search via gps. Recive the message that the Station can not found.
3. dont close that dialoge and type your station in the textbox
4. click search and choose your station from the list
5. to: Station, type the seccond station in the text box and klick search.
Now it shows me the list of station one. And not of station two.

The second bug is that the aplication close hisself after 1-2 minutes.
And the realy strange thing is, if i start the app with the terminal to see why it close. It dont close himself.


Mfg Michael whos waiting for a comforable GPS usability ;-)

I can't reproduce this, because for me it finds stations when using the GPS-button.
The problem with GPS is, that (at least) the first "fix" is a pseudo-fix (the last known coordinate, or sometimes the center of the current country). Therefor, its likely not to find stations for the first fix.
This should be filtered in the software. I didn't have a look at how this is implemented here, perhaps I can look at it tomorrow…
Of course, what you mention here is a different bug, but I can only look at it if I find a way how to get the "no stations found" message

Ok, I can reproduce this… I don't know where to start looking for that problem. Perhaps its even a feature of Maemo, since it doesn't happen when starting from commandline…

@leachim: what's the exact error message you are getting when clicking the gps-button? Can you reproduce the error even after you have used the gps-receiver (e.g. with the Maps program or some other GPS-program)?

This information would help me, because I just had a look at the code and it seems reasonable…

Only thing I would change, are METHOD and INTERVAL for the GPS device, as the default interval could be very long to obtain a fix, and method could be set to require a data-link, as a data-connection is required anyway for fahrplan, and the fix is way faster with such a method…

Program is closing after 1-2 mins too
Würde es gerne länger geöffnet lassen um schnell Mal zwischendurch draufzuschauen

It's missing an osso_initialize (), hence the crashes after 2 mins.

Do something like:
#include <config.h>
#include <libosso.h>

osso_context_t *osso_ctx = osso_initialize (PACKAGE, VERSION, FALSE, NULL);
g_assert (osso_ctx != NULL);

*after the program has ran*

osso_deinitialize (osso_ctx);

thanks qwerty12, that fixed the 2 minute crash

@smurfy: I committed the change to the SVN. Could you please pay attention not to commit code which is not compilable? The global variables used for the departure window are not yet in the header files… Had to comment out a lot of stuff to compile it
SVN code may contain bugs, but should always compile. That way, people can play around with it without having to fix compile-errors first

@hcm: The exact error message is "Station not found!", i cant reproduce it after a GPS fix. becouse it find everytime a station via GPS search. So probably the implementation of an working GPS search will fix this too.

When will be a fixed version available ?

ok, I also get this error just after restarting the device without using the GPS first. But I still can't reproduce the "wrong result list" error you described in your previous post. Does this error happen for any combination of search terms?

So about the "first GPS fix" issue: apparently, the "status" variable of the GPS receiver doesn't work as expected, status ==LOCATION_GPS_DEVICE_STATUS_FIX can't be used to tell if its a real fix…
So the workaround would be to check the horizontal accuracy to be less than a certain number. So what should that number be?

I would prefer to use just cell based positioning because its much faster and doesn't drain the battery (and you need a network connection anyway to use fahrplan).
An accuracy of 5km is usually available in a relatively short time. In my experience, the position is still much better than 5 km, when I checked it, it was usually a few hundred meters up to about 1 km… In urban neighborhoods, this could of course find way too many stations…

Better positioning can be achieved by using "real" GPS, but it will take longer to get a (better) fix and drains the battery (not much of course, since we only need one fix).

I would rather NOT switch the GPS receiver on when starting the application, since it is not always needed and it would drain the battery if you leave the application open to have a look at it from time to time…

So in the attachment there is a version which uses cell based positioning and checks for an accuracy better than 5km. Please check if is fast enough and finds your station in your neighborhood… there is no need to replace the existing fahrplan app, just gunzip it somewhere and run it from console… It will crash if you try to switch to the departure window, since this is not implemented, yet