OpenSatNav crashes on startup

[Paspartout]

OpenSatNav (rev. 133) crashes on startup on my HTC Tattoo running android 1.6 (Donut).

But how to debug this?

Laias

[chris_debian (cjhandrew)]

Hi, Laias!

Can you tell us what OSN was doing when it crashed? Are you able to replicate the error? Were you prompted to submit any bug information?

Thanks,

Chris.

[Paspartout (pv-ndsmail)]

chris_debian wrote:
> Hi, Laias!
>
> Can you tell us what OSN was doing when it crashed? Are you able to replicate the error? Were you prompted to submit any bug information?
>
> Thanks,
>
> Chris.

Yes, it just says:
Sorry!
The application OpenSatNav (process org.opensatnav) has stopped unexpectedly. Please try again.

Tried restarting the phone, removing some apps and delete the OSN-Folder in the root-dir.

PS: I will add this to the Description

Laias

[Paspartout (pv-ndsmail)]

Only this Message is showing on startup:
Sorry!
The application OpenSatNav (process org.opensatnav) has stopped unexpectedly. Please try again.

I tried restarting the phone, removing some apps and delete the OSN-Folder in the root-dir.

[Paspartout (pv-ndsmail)]

Hey!

I reinstalled rev. 128 which works fine!
And on firts starup i got some ErrorMessages to report!

I posted it as txt file!

Hope this can help

Laias

[chris_debian (cjhandrew)]

Thanks, Laias; hopefully the Devs will be able to interpret this.

Going by what you say, the problem may be in the code added since Rev. 128.

Not sure why the error was not automatically turned into a bug report.

Thanks, again.

Chris.

[chris_debian (cjhandrew)]

Looks like it mentions Tripstats.

[droidguy (stan-berka)]

I have looked at it. And, yes, it definitely has a problem starting the tripstats service. I'm not sure what the stack trace really means. It just says the argument can't be null when trying to start LocationHandler. I'm not getting this error on my G1, although, in the emulator there is an error:
01-29 06:00:47.359: DEBUG/OpenSatNav(3968): Ignoring: java.lang.IllegalArgumentException: listener==null
No more info. The TraceRecorderService had a setForeground(true); before starting the LocationHandler, so I added this to tripstats service and the listener==null error seems to be gone in my emulator. Do you know what is the meaning of setForeground(true); in service onStart()?

Simple patch attached. Laias, could you test the development build tonight?

Has anyone more insight into the LocationHandler stuff?

droidguy

[Paspartout (pv-ndsmail)]

Yes i will test it.

But i also have the Android SDK here... I will try to apply the patch and rebuild from svn.

Laias

[Paspartout (pv-ndsmail)]

I applied the patch but nothing changes. I got the same error report :-(.

[droidguy (stan-berka)]

Laias,
Do you see this error in the emulator or only on your HTC Tatoo?
droidguy

[Paspartout (pv-ndsmail)]

Its not working on my htc tattoo... Tested in the emulator(andorid 1.6) and it works...

[Paspartout (pv-ndsmail)]

Heey!
I tested it again in emulator with QVGA (240*320) Resolutin and got the same error!
So the problem have to be the resolution...

Laias

[droidguy (stan-berka)]

Laias,
Great job diagnosing. How do you switch emulator to QVGA?

[Paspartout (pv-ndsmail)]

I just created a new AVD with Android 1.6 and QVGA.
Do you use the original sdk?

[droidguy (stan-berka)]

Laias,
An older development build works fine on your Tatoo, rifgt? I believe you mentioned #128. Can you try, what happens when you start the version that works and do Collaborate - Start Recording Trace?
droidguy

[Paspartout (pv-ndsmail)]

Yeah,
i think i found the problem... You're right OSN crashes too, when i try to start recording...(Error Message attached)
But the problem is, that everything works (rev. 133) when my gps is enabled and only crashes if its disabled.
But what is when i just wanna browse the map without gps enabled?
Hope that could fixed later.

[droidguy (stan-berka)]

Thanks, Laias. I hope we may be able to fix it now. Let's see what other devs think.

Guys, to summarize what we know about this problem so far:
* Applies to HTC Tatoo, or just any android device using QVGA screen size.
* Occurs only if GPS is off while starting OSN.
* Problem applies to revisions 128 or earlier ones.
* What happens is: calling mLocationHandler.start() causes OSN to crash.
* Two similar error messages attached.

[Christian-uni-paderborn (christian-uni-paderborn)]

I have the same problem on my Milestone (now on Android 2.0.1), with the latest version r136.

I also tried to debug the app, but the Android launcher crashes with a NPE (which it does not for my own "play around with Android" application)...

[droidguy (stan-berka)]

Workaround: turn GPS on before startin OSN.

[droidguy (stan-berka)]

After working with Laias on his problem, I thought this issue is related to the QVGA format of the display (Laias, you have reproduced this problem in emulator with Android 1.6 with QVGA, right?).
Now, that csoltenborn has reported the same problem, I'm not sure, since the Milestone definitely doesn't use the QVGA display format. Is this related to the version of Android OS. csoltenborn, does the above workaround work? Any ideas?

[Paspartout (pv-ndsmail)]

Hi!
No i think its not the resolutin. Its working again in the emulator... I think it was my fault... I tried compiling OSN for Android 2.0 and 2.1 myself and got this error. So its not the reolutin. Sorry.
Laias

[Kieran Fleming (kieran-fleming)]

Can you paste in here the error that occurs? If you can get it to happen in the emulator the error will be in the 'LogCat' dialog.

[Kieran Fleming (kieran-fleming)]

Sorry, didn't see you already had one :)

[droidguy (stan-berka)]

So, Laias, what version of Android OS do you have on your Tatoo device where you originally observed the crash?
droidguy

[chris_debian (cjhandrew)]

With regard to the last question 'What version of Android OS do you have on your Tatoo device where you originally observed the crash?', does this get submitted as part of the automated bug reporting? If not, is it something we can implement?

Chris.

[Guillaume Rosaire (zerog)]

chris_debian wrote:
> With regard to the last question 'What version of Android OS do you have on your Tatoo device where you originally observed the crash?', does this get submitted as part of the automated bug reporting? If not, is it something we can implement?
>
> Chris.

It is, see below for an example of bug report subject :
> OpenSatNav 0.9 [SVN Rev 137] Error Report (*Android: 1.6* - model: T-Mobile G1)

[chris_debian (cjhandrew)]

Guys,

Anybody know whether this bug is going anywhere? Do we need to update it? Do we have enough information to do any work on it?

Thanks,

Chris.

[droidguy (stan-berka)]

The first thing with this issue, is to reproduce it, hopefully in the emulator. I'm not sure anymore whether it's related to the resolution (QVGA) or the version of the OS, or possibly smth else.
droidguy

[Kieran Fleming (kieran-fleming)]

I've tried the latest code from this branch:
https://code.launchpad.net/~kieran-fleming/opensatnav/osn-small-tweaks
in the emulator running a QVGA android 1.6 image and there's no crash. Turning location services off doesn't change anything for me. I'm marking this as Incomplete, meaning that unless we get a way to reproduce this issue reliably it will have to be closed.

[droidguy (stan-berka)]

Chris has got this error today. Chris, what is your Android version and phones screen resolution. Anyone has an idea how to pin this one down?

[Guillaume Rosaire (zerog)]

Kieran doesn't reproduce the issue, but Chris with the automatic build does. Couldn't it be possible that it comes from my android build ?
Kieran, could you download and test with the uploaded apk for rev 150 ?

[Kieran Fleming (kieran-fleming)]

OK, rev 150 does FC if I turn off both location providers. This is on my G1 running android 2.1.

"Guillaume Rosaire" <email address hidden> wrote:

>
>Kieran doesn't reproduce the issue, but Chris with the automatic build does. Couldn't it be possible that it comes from my android build ?
>Kieran, could you download and test with the uploaded apk for rev 150 ?
>
>--
>OpenSatNav crashes on startup
>https://bugs.launchpad.net/bugs/579464
>You received this bug notification because you are a member of
>OpenSatNav Admins, which is the registrant for OpenSatNav.
>
>Status in OpenSatNav: Incomplete
>
>Bug description:
>OpenSatNav (rev. 133) crashes on startup on my HTC Tattoo running android 1.6 (Donut).
>
>But how to debug this?
>
>Laias
>
>

[chris_debian (cjhandrew)]

I had both location providers turned off, too. I'm using Cyanogenmod (Stable) and don't know what the default resolution is on my G1, but that is what I am using. The Cyanogenmod build is listed in my APK150-chris_debian attachment in the bug I filed, but as I'm sending this from my G1, it's difficult to copy the information. I think Kieran may be on the same build, though.

Cheers,

Chris.

[Kieran Fleming (kieran-fleming)]

Fix committed:
http://bazaar.launchpad.net/~opensatnav-admins/opensatnav/release-1.0/revision/147
Please test this when a new APK comes up but it fixes the problem for me.

[Paspartout (pv-ndsmail)]

Hi!
Build OSN from Source and now its just working fine!
Can Confirm... Fix works! Thanks!

[Alastor (alastor-ru)]

Crashes on Android 2.2 (Nexus One phone)

[Murphy (murphy2712+launchpad)]

Here's the error (Nexus One - Froyo):

E/AndroidRuntime( 960): FATAL EXCEPTION: main
E/AndroidRuntime( 960): java.lang.RuntimeException: Unable to start service org.opensatnav.TripStatisticsService@449cf3d0 with Intent { cmp=org.opensatnav/.TripStatisticsService }: java.lang.IllegalArgumentException: syntax error: U_UNQUOTED_SPECIAL: ###.##
E/AndroidRuntime( 960): at android.app.ActivityThread.handleServiceArgs(ActivityThread.java:3063)
E/AndroidRuntime( 960): at android.app.ActivityThread.access$3600(ActivityThread.java:125)
E/AndroidRuntime( 960): at android.app.ActivityThread$H.handleMessage(ActivityThread.java:2096)
E/AndroidRuntime( 960): at android.os.Handler.dispatchMessage(Handler.java:99)
E/AndroidRuntime( 960): at android.os.Looper.loop(Looper.java:123)
E/AndroidRuntime( 960): at android.app.ActivityThread.main(ActivityThread.java:4627)
E/AndroidRuntime( 960): at java.lang.reflect.Method.invokeNative(Native Method)
E/AndroidRuntime( 960): at java.lang.reflect.Method.invoke(Method.java:521)
E/AndroidRuntime( 960): at com.android.internal.os.ZygoteInit$MethodAndArgsCaller.run(ZygoteInit.java:868)
E/AndroidRuntime( 960): at com.android.internal.os.ZygoteInit.main(ZygoteInit.java:626)
E/AndroidRuntime( 960): at dalvik.system.NativeStart.main(Native Method)
E/AndroidRuntime( 960): Caused by: java.lang.IllegalArgumentException: syntax error: U_UNQUOTED_SPECIAL: ###.##
E/AndroidRuntime( 960): at com.ibm.icu4jni.text.NativeDecimalFormat.applyPattern(NativeDecimalFormat.java:569)
E/AndroidRuntime( 960): at com.ibm.icu4jni.text.NativeDecimalFormat.applyLocalizedPattern(NativeDecimalFormat.java:261)
E/AndroidRuntime( 960): at java.text.DecimalFormat.applyLocalizedPattern(DecimalFormat.java:633)
E/AndroidRuntime( 960): at org.opensatnav.services.TripStatistics.getSpeedString(TripStatistics.java:146)
E/AndroidRuntime( 960): at org.opensatnav.services.TripStatistics.getAverageTripSpeedString(TripStatistics.java:137)
E/AndroidRuntime( 960): at org.opensatnav.TripStatisticsController.tripStatisticsChanged(TripStatisticsController.java:69)
E/AndroidRuntime( 960): at org.opensatnav.services.TripStatistics.addTripStatsListener(TripStatistics.java:56)
E/AndroidRuntime( 960): at org.opensatnav.TripStatisticsService.onStart(TripStatisticsService.java:81)
E/AndroidRuntime( 960): at android.app.Service.onStartCommand(Service.java:420)
E/AndroidRuntime( 960): at android.app.ActivityThread.handleServiceArgs(ActivityThread.java:3053)
E/AndroidRuntime( 960): ... 10 more

It looks like the error is in org.opensatnav.services.TripStatistics.getSpeedString(TripStatistics.java:146):
formatter.applyLocalizedPattern("###.##");

I just looked at it in a test project using a try/catch around the applyLocalizedPattern:
DecimalFormat formatter = new DecimalFormat();
try {
    formatter.applyLocalizedPattern("###.##");
} catch (Exception e) {
    Log.e(TAG, e.getMessage());
}
It's quite strange since it does not bug in the 2.2 emulator but not in our "preview" of Froyo.
For a test value of 10, and with the "###.##" pattern,
Log.i(TAG, formatter.format(10 * 3600f / 1000f) + " km/h");
Emulator 2.2 prints: "36 km/h"
Nexus 2.2 p...

[Murphy (murphy2712+launchpad)]

This kind of error appears again for me when I switch language to french.
I added some try/catch in TripStatistics to avoid the force closes:
...
try {
 distanceFormat.applyLocalizedPattern("###,###.#");
} catch (Exception e) {
 e.printStackTrace();
 // TODO: handle exception
}
...
try {
 speedFormat.applyLocalizedPattern("###.##");
} catch (Exception e) {
 e.printStackTrace();
 // TODO: handle exception
}

But it still don't start: black screen... then force close. Not a clue in the logs.
Using Cyanogen 6.0 with Nexus One.

[Murphy (murphy2712+launchpad)]

Even in english it doesn't start.
When I compile it in debug mode it starts... strange!

[Kieran Fleming (kieran-fleming)]

I can't replicate this anymore in English or French, can you try again with the latest code?

[Murphy (murphy2712+launchpad)]

I found how to replicate it (no matter which language is selected):
1. record a trace (with points, not a empty one, so you must have a fix),
2. quit osn,
3. start osn => black screen.

[Kieran Fleming (kieran-fleming)]

Yep, I see that too. I managed to get it to work again by commenting out
some stuff in SatNavActivity but I haven't got a cause yet.

On Mon, 2010-10-18 at 01:15 +0000, Murphy wrote:
> I found how to replicate it (no matter which language is selected):
> 1. record a trace (with points, not a empty one, so you must have a fix),
> 2. quit osn,
> 3. start osn => black screen.
>

[Kieran Fleming (kieran-fleming)]

I've fixed this in rev 174. Turns out there was a division by 0 error that led to an infinite loop when trying to draw the track.

[droidguy (stan-berka)]

Kieran,
Thanks. It first showed up after I added trip stats and I thought it was
somehow related to it. Now, the trip stats seems to be partly/completely
disabled, doesn't it?

droidguy

On Oct 26, 2010 5:06 AM, "Kieran Fleming" <email address hidden> wrote:

I've fixed this in rev 174. Turns out there was a division by 0 error
that led to an infinite loop when trying to draw the track.

** Changed in: opensatnav
      Status: In Progress => Fix Committed

--
OpenSatNav crashes on startup
https://bugs.launchpad.net/bugs/579464
You received this bug noti...
Status in OpenSatNav: Fix Committed

Bug description:
OpenSatNav (rev. 133) crashes on startup on my HTC Tattoo running android
1.6 (Don...