Out of memory

[Kieran Fleming]

Submitted by Chris to the bug tracking service from build 133:

java.lang.OutOfMemoryError: bitmap size exceeds VM budget
at android.graphics.BitmapFactory.nativeDecodeByteArray(Native Method)
at android.graphics.BitmapFactory.decodeByteArray(BitmapFactory.java:384)
at android.graphics.BitmapFactory.decodeByteArray(BitmapFactory.java:397)
at org.andnav.osm.views.util.OpenStreetMapTileFilesystemProvider$1.run(OpenStreetMapTileFilesystemProvider.java:184)
at java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExecutor.java:648)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:673)
at java.lang.Thread.run(Thread.java:1060)

[chris_debian (cjhandrew)]

As this mentions memory, I thought I should mention that the Memory Slider in Preferences, was set to 1 gb.

Thanks,

Chris.

[Kieran Fleming (kieran-fleming)]

How long (approximately) were you recording the trace for before you went to upload it?

[chris_debian (cjhandrew)]

kizza wrote:
> How long (approximately) were you recording the trace for before you went to upload it?

Kieran,

5 minutes maximum.

Thanks,

Chris

[chris_debian (cjhandrew)]

Has anybody seen this bug in builds > 133?

Chris.

[Guillaume Rosaire (zerog)]

Same stacktrace experienced 7 times since rev 125 (30/01)
4 times by Stan
1 by Chris
1 by another user (on Android 2.0.1)

Revs 125, 133, 139

There is some kind of memory leak in the GPS trace code.

[chris_debian (cjhandrew)]

ZeroG wrote:
> Same stacktrace experienced 7 times since rev 125 (30/01)
> 4 times by Stan
> 1 by Chris
> 1 by another user (on Android 2.0.1)
>
> Revs 125, 133, 139
>
> There is some kind of memory leak in the GPS trace code.

Is this something we can track back to code in release 125? Will it be difficult to diagnose? Would anyone like to lead on this?

Chris.

[Kieran Fleming (kieran-fleming)]

This is a really nasty bug. We're still getting automatic bug reports about this, so it's definitely not fixed. I haven't been able to replicate this at all, so it's going to be hard for me to fix! Can anyone help me out with this, or help me replicate it somehow?

[Neil Boyd (neilboyd)]

I spent some time making some fixes to osmdroid for this issue, and although I wouldn't go so far as to say it's fixed, I haven't had this error for quite a long time. Maybe that doesn't directly help you, but I'm just hinting that if you look at the code then there's a good chance you're on a good track. You could also just try using the new jar version of osmdroid directly.

This comment relates to the stack trace in the description. I don't know anything about the things mentioned in the comments.

[Guillaume Rosaire (zerog)]

I just had this issue while trying to replicate it (I had never really used the record track function option)

To replicate :
Start OSN
Start a trace
Change screen orientation a few times (I did approximately 20 to 30 screen orientation changes) and you'll see that OSN is less and less responsive, I guess it's GCing a lot before eventually dying with an OOM error.

[evolvedlight (steve-evolvedlight)]

Yes, that works.

[evolvedlight (steve-evolvedlight)]

And also works if not recording a trace

[droidguy (stan-berka)]

Good job, guys! I was going to tell you that it is definitely related to
orientation changes, but you have already figured it out. It killed my gpx
tracks so often that I had to switch to My Tracks to record them: look at
OSM in OSN, and switch to My Tracks to add a waypoint. How is doing the
development for a reliable track management?
droidguy

On Jul 6, 2010 10:05 AM, "evolvedlight" <email address hidden> wrote:

And also works if not recording a trace

--
Out of memory
https://bugs.launchpad.net/bugs/579466
You received this bug notification because...

[evolvedlight (steve-evolvedlight)]

I'm going to be working on it the second half of this week. I've been
working on turn-by-turn nav waypoint detection this week - its pretty much
working but probably needs improvement.

On 6 July 2010 21:53, droidguy <email address hidden> wrote:

> Good job, guys! I was going to tell you that it is definitely related to
> orientation changes, but you have already figured it out. It killed my gpx
> tracks so often that I had to switch to My Tracks to record them: look at
> OSM in OSN, and switch to My Tracks to add a waypoint. How is doing the
> development for a reliable track management?
> droidguy
>
> On Jul 6, 2010 10:05 AM, "evolvedlight" <email address hidden>
> wrote:
>
> And also works if not recording a trace
>
>
> --
> Out of memory
> https://bugs.launchpad.net/bugs/579466
> You received this bug notification because...
>
> --
> Out of memory
> https://bugs.launchpad.net/bugs/579466
> You received this bug notification because you are subscribed to
> OpenSatNav.
>

[droidguy (stan-berka)]

Great. I'm looking forward to have it.

On Jul 6, 2010 2:20 PM, "evolvedlight" <email address hidden> wrote:

I'm going to be working on it the second half of this week. I've been
working on turn-by-turn nav waypoint detection this week - its pretty much
working but probably needs improvement.

On 6 July 2010 21:53, droidguy <email address hidden> wrote:

> Good job, guys! I was going to tell...
> --

> Out of memory
> https://bugs.launchpad.net/bugs/579466
> You received this bug notification because you are subscribed to
> OpenSatNav.

>

--
Out of memory
https://bugs.launchpad.net/bugs/579466

You received this bug notification because you are a member of
OpenSatNav Admins, which is the regis...

[Guillaume Rosaire (zerog)]

May be related to upstream osmdroid as Neil implied. Time to include directly the library ?

http://code.google.com/p/osmdroid/issues/detail?id=54