Resume command throws RR to 100% - All fixed

**obrienj** · 08-22-2008, 11:41 AM

In my use of RR, I have an external process that sends PAUSE and RESUME commands to RR via a Windows WM_COPYDATA message. The external process is monitoring when IG4 starts and stops speaking. Note I am NOT running IG4 native, just letting it be embedded like any other external GPS package.

This all works just fine for quite a while with the PAUSE and RESUME working just fine, and all of a sudden, right after a RESUME command, RR goes to 100% and the RR debug trace stops right after the RESUME in question.

During the test, I have the Task Manager running and at the time of the failure as soon as the music resumes, RR jumps to the top of the Process List (sorted decscending CPU) and oscillates between 90% and 100%. No skin keys work, but winamp continues, IG4 continues, as do all other processes abiet somewhat slower. If I end the RR Process from the Task Manager, the system runs as it should, of course, without RR and the RR debug trace shows no entries after the resume in question.

I have the 12/13/2007 version of RR installed.

Regards,
Jim

**obrienj** · 08-24-2008, 12:17 AM

This problem is getting more and more curious.

The Resume command seems only to be circumstantial.

I have had it happen every time I run RR, iGuidance V4, winamp, GPSGate, and my external support program.

It happens whether I have iGuidance running native or not.

It happens about the same time in the "demo" routing or real routing whether traces are on or not.

All that is executing is RR (about 2%), iGuidance (about 5%), winamp (about 1%), GPSGate (about 2%) and all of a sudden RR goes to 90-100% and nothing is going on but winamp playing music and iGuidance routing and redrawing the map.

The last thing in the trace output has never been the same. It can be Resume or SetVar or most anything else.

I have no idea where to go with this one but I do notice several simular posts in the forums in the sense that iGuidance is a component of the problem but that could be circumstantial as well.

I have GDIPlussWrapper.dll and dx7vb.dll registered. Are there any other that need to be registered?

Is Guino still working on RR?

Regards,
Jim

**obrienj** · 08-25-2008, 12:17 AM

I have removed all of my external processes and am working simply with RR, iG, Winamp, and iGNative setup. The results are not encouraging:

IGNative works sometimes. On some startups, it does nothing when I run a route through the IG "demo". Other times it does work but not perfectly.

BUT, the issue of RR going to 90-100% CPU is still there.

No real pattern that i can detect, just suddenly at 100%. Curiously, there is NO interaction between IG and RR at the time of failure, IG is just tracking along the route on the map. During the entire test the embedded IG screen is displaying.

**obrienj** · 08-25-2008, 08:37 PM

All fixed.

After starting windbg on the looping Roadrunner.exe and tracing the loop, I found it had something to do with ZipLocator (found some identifiable text in the windbg trace).

I first tried to delete zips.dat and then ziplocator.dll, but that only accomplished "hanging" Roadrunner.exe. I looked in modRoadRunner.bas in thefuncton GetCurrentLocation (I think) and found a coding bug where the code goes right ahread and reads zips.dat after it has determined that it doesn't exist.

By tracing the callers of that function, I found the only caller depended on the setting of the rr.ini variable "localdatacaptureinterval". Since I didn't need any of the functions dependent on it, I simply set that variable to zero and the looping RoadRunner problem went away.

Regards,
Jim