Tasker 1.1.1u2 crashes at startup

[hackworks]

I am using Tasker on CM7 - O2X (P990). With the recent kernel update, tasker crashes when I open it with the following trace in logcat. Please let me know if there is any other details I can provide to troubleshoot this. I have also submitted a report from the android crash handler.

Kernel code: https://github.com/CyanogenMod/lge-kernel-star

Logcat:

[snip]

W/PackageManager( 1133): Not granting permission android.permission.MODIFY_PHONE_STATE to package net.dinglisch.android.taskerm (protectionLevel=3 flags=0x8be44)

I/ActivityManager( 1133): Start proc net.dinglisch.android.taskerm for broadcast net.dinglisch.android.taskerm/.ReceiverStaticAlwaysOn: pid=1791 uid=10064 gids={1015, 1006, 3001, 3002, 3003}

E/Tasker  ( 1791): 14.11.53#Utils: error: No such device

E/AndroidRuntime( 1791): java.lang.RuntimeException: Unable to start receiver net.dinglisch.android.taskerm.ReceiverStaticAlwaysOn: java.lang.NullPointerException

E/AndroidRuntime( 1791): at net.dinglisch.android.taskerm.ah.h(Unknown Source)

E/AndroidRuntime( 1791): at net.dinglisch.android.taskerm.ah.a(Unknown Source)

E/AndroidRuntime( 1791): at net.dinglisch.android.taskerm.o.a(Unknown Source)

E/AndroidRuntime( 1791): at net.dinglisch.android.taskerm.cg.a(Unknown Source)

E/AndroidRuntime( 1791): at net.dinglisch.android.taskerm.ReceiverStaticAlwaysOn.onReceive(Unknown Source)

I/ActivityManager( 1133): No longer want net.dinglisch.android.taskerm (pid 1791): hidden #16

[Pent]

> I am using Tasker on CM7 - O2X (P990). With the recent kernel update,
> tasker crashes when I open it with the following trace in logcat. Please
> let me know if there is any other details I can provide to troubleshoot
> this. I have also submitted a report from the android crash handler.

This is almost certainly a ROM problem.

Does it still happen if you uninstall and reinstall Tasker ?

Which Tasker version do you have ? 1.1.1u2 from Market ? I need to
know that so I can decode the log.

Pent

[hackworks]

Thank you for quick response. I am using Tasker from Market. I have uninstalled and reinstalled, tried removing all my profiles to ensure it is not due to some specific profile. It is surly a kernel triggered since it worked with the previous version. I am debugging the kernel by bisecting the changes to find which commit is causing this problem. However, it might still be good for the app to handle the error.

[Pent]

> the changes to find which commit is causing this problem. However, it might
> still be good for the app to handle the error.

Thing is, there have to be things I can rely on such as standard
Android features being available.

For example, maybe the developer will mess up the function that gets
the time. I can't check every single call I make to ask Android what
the time is.

Anyway, I noted the log and will have a look see if I can find where
the error is originating when I get a chance.

Pent

[hackworks]

If you can tell what you are trying to do that is resulting in the error, I can bisect faster. There are just too many changes to complete in a long sitting. The error "#Utils No such device" seems to be the key.

[Pent]

> If you can tell what you are trying to do that is resulting in the error, I
> can bisect faster. There are just too many changes to complete in a long
> sitting. The error "#Utils No such device" seems to be the key.

Maybe something to do with the Market app not being installed or wrong
version ?

Or /sys ?

Not clear.

Pent

[hackworks]

I looked at the /sys folder and saw a bunch of block and char devices and could not find anything wrong. When I run tasker repeatedly after a crash, it fails with different device numbers. Hence, I presume that tasker is trying to create a device. Could it be a tty device it is trying to access?

Ex:

E/Tasker  ( 7433): 19.23.19#Utils: error: No such device

[snip]

E/Tasker  ( 7433): 19.24.19#Utils: error: No such device

[Pent]

> I looked at the /sys folder and saw a bunch of block and char devices and
> could not find anything wrong. When I run tasker repeatedly after a crash,
> it fails with different device numbers. Hence, I presume that tasker is
> trying to create a device.

Nope.

Have you got Market up-to-date on the device ? I think it's most
likely the validation library (built-in to Tasker).

> Could it be a tty device it is trying to access?

No, don't think so.

Pent

[hackworks]

I have market 3.3.11 and this is latest. It cannot be market as it works with different kernel. If you can provide a build with extra logging, I can test that. Can you make Tasker run with root privilege for testing purpose?

[Rich]

Chiming in with a crash report here, running CM7 on a G2x (P999) so
it's a very similar device to OP's and the crash also started
happening after the kernel bump. Same specs: latest Tasker from newest
Market, un/reinstalled, no profiles.

I see it's already been escalated so I'll stand by for further
updates.

[Pent]

>  I have market 3.3.11 and this is latest.
>It cannot be market as it works with different kernel. If you can provide a >build with extra logging,  I can test that.

Unfortunately I'm totally snowed under ATM. The unfortunate thing is
that you
can't turn on the debugging without access to the app...

Try deleting the Market version and installing the trial version:

http://tasker.dinglisch.net/download.html

That doesn't have the Market validation library in, so it's useful for
excluding
that culprit at least.

> Can you make Tasker run with root privilege for testing purpose?

No, apps are not started in root processes by Android.

Pent

[hackworks]

I get the same error with the trial version too. Is there an 'strace' equivalent for android apps? I do not know anything about android apps and dalvik. If there is a way to trace, I can do that and provide you the logs.

$ adb logcat

[snip]

E/Tasker  ( 8248): 16.32.25#Utils: error: No such device

D/AndroidRuntime( 8248): Shutting down VM

W/dalvikvm( 8248): threadid=1: thread exiting with uncaught exception (group=0x40015560)

E/AndroidRuntime( 8248): FATAL EXCEPTION: main

E/AndroidRuntime( 8248): java.lang.RuntimeException: Unable to start activity ComponentInfo{net.dinglisch.android.tasker/net.dinglisch.android.tasker.Tasker}: java.lang.NullPointerException

E/AndroidRuntime( 8248): at android.app.ActivityThread.performLaunchActivity(ActivityThread.java:1768)

E/AndroidRuntime( 8248): at android.app.ActivityThread.handleLaunchActivity(ActivityThread.java:1784)

E/AndroidRuntime( 8248): at android.app.ActivityThread.access$1500(ActivityThread.java:123)

[Pent]

> I get the same error with the trial version too. Is there an 'strace'
> equivalent for android apps? I do not know anything about android apps and
> dalvik.

Me also not.

Pent

[hackworks]

Sure you are joking :-) Tasker is one heck of an app and hence I am doing all I can to get it up and working! 

[Pent]

> Sure you are joking :-)

Not at all.

>Tasker is one heck of an app and hence I am doing
> all I can to get it up and working!

I wish I could help more. In a week or two the beta will be out and I
can get more active.

Pent

[hackworks]

I will wait patiently for the best, looking forward to it

[hackworks]

On Wednesday, 30 November 2011 22:43:32 UTC+5:30, hackworks wrote:

I will wait patiently for the best, looking forward to it

I mean beta (which will be the best) 

[Pent]

> > I will wait patiently for the best, looking forward to it
>
> I mean beta (which will be the best)

Grin.

Pent

[Pent]

I think I found this bug (someone else send in a log too).

I'm fairly confident it's when Tasker can't find any CPU governors for
a CPU during initialization of 'capabilities'.

Pent

[Rich]

Nice, hopefully you can get it worked out.

[Tony Sumrall the Younger]

I'd be happy to beta test the change if/when you have one, Pent.
--
Tony

On Dec 3, 8:43 am, Pent <supp...@apps.dinglisch.net> wrote:

[hackworks]

Cool, Tegra does not support CPU governors. Earlier, we had dummy governor which was a placebo. I guess the dev (rmcc) removed it completely to avoid any confusion. I request you to make a beta or an update available. I feel crippled without Tasker as my phone is heavily tweaked with Tasker.

[Pent]

> Cool, Tegra does not support CPU governors. Earlier, we had dummy >governor

Ah, that would explain it.

> which was a placebo. I guess the dev (rmcc) removed it completely to avoid
> any confusion. I request you to make a beta or an update available. I feel
> crippled without Tasker as my phone is heavily tweaked with Tasker.

Around two weeks, I can't do sooner.

Pent

[Pent]

> I am looking at changes if there is anything governor specific. Can you let
> me know which file you look for to read governor details? That will help me
> identify the change.

/sys/devices/system/cpu/cpuX/scaling_available_governors

and

scaling_governor

Pent

[Tony Sumrall the Younger]

Actually, on a functional system with Tasker (CM7 nightly build 201),
I have /sys/devices/system/cpu/cpuX/cpufreq/
scaling_available_governors (note that /cpufreq/ was inserted after
cpuX),
--
Tony

[hackworks]

On Monday, 5 December 2011 23:42:21 UTC+5:30, Tony Sumrall the Younger wrote:

Actually, on a functional system with Tasker (CM7 nightly build 201),
I have /sys/devices/system/cpu/cpuX/cpufreq/
scaling_available_governors (note that /cpufreq/ was inserted after
cpuX),

On the latest kernel where tasker does not work, I have the same path as you have with tasker working. So, the problem is elsewhere I guess. 

[ricardo....@gmail.com]

It appears go be blowing up due to the lack of the scaling_cur_freq node.

[Tony Sumrall the Younger]

No, I found it wouldn't work with the newer commits so I restored
from a nandroid and looked for the files. Looks like Ricardo
(@cm_arcee), the owner of CM7 on the P999, has narrowed it down,
though.

Thanks, Ricardo. -- Tony

[hackworks]

Take this guy seriously, he knows the guts of the kernel when he speaks :-) Thanks rmcc

[Tony Sumrall the Younger]

Has anyone tried to create that particular file just to see if it
resolves the crash? Of course it wouldn't be hooked up to the kernel
but the file could still exist and just contain the text that would
normally be returned. In my version of CM7 nightly, I get "performance
powersave". I may give that a try if I can find the time but if
someone else has already tried, I'd be interested in your results.
It'd have to be done on each boot but I think it could be incorporated
in to one of the various rc scripts.
--
Tony

[Tony Sumrall the Younger]

Don't know what I was thinking -- it's a special filesystem so, no,
you can't just go create a file there. But, since I had a newer build,
I compared the /sys/devices/system/cpu/cpu0/cpufreq on the old build
(CM 201) and a newer one and found that there's one file missing in
the newer builds -- scaling_available_frequencies. So, I'm guessing
that's the file that's causing the failure in Tasker.

Just FYI.
--
Tony

[hackworks]

Hello Pent,

Using workarounds and still waiting for a working Tasker. This is a bug and you have a fix, please make a bug fix release. I feel crippled without Tasker. My phone features are controlled by Tasker.

[Pent]

> Using workarounds and still waiting for a working Tasker. This is a bug and
> you have a fix, please make a bug fix release. I feel crippled without
> Tasker. My phone features are controlled by Tasker.

Beta by Friday evening most likely.

Pent

[Tony Sumrall the Younger]

Yippee! I've been watching CM nightlies go by for over a week because
I can't do without Tasker but, just today I've started looking at
things that might get me by until a patch or beta or fix comes out.
--
Tony

[hackworks]

With the beta out and confirm it works. This thread is closed now.