Page 1 of 2

Recognising An OLV Crash - Please Report These

Posted: Thu Feb 09, 2012 12:34 am
by Fred
If something goes wrong with OLV it will output a stack trace or some other message in the console. If you don't start it from the console you don't have access to that information, so for the time being, please launch it from a CMD or Bash window of some sort.

If something goes wrong with a decoder you'll get this:

Image

Or something similar. If you see this, please report it, preferably with the log file, a set of actions leading to it, and the text from the console.

If you don't report bugs, we can't fix them! :-)

Fred.

Re: Recognising An OLV Crash - Please Report These

Posted: Tue Mar 06, 2012 1:51 am
by Peter
Fred wrote:please launch it from a CMD or Bash window of some sort
Sorry, I'm a lazy double click windows user.

I opened the OLV by double clicking it on the desktop, File->Open->JSeriesFeb24rdTuning.bin waited a little bit for it to load and then when it showed me the Graphing Option Pane I pulled EGO down to be graphed. Then it kinda froze, and ended up displaying this.

Re: Recognising An OLV Crash - Please Report These

Posted: Tue Mar 06, 2012 2:00 am
by Fred
Fair enough, I might add a force console option to the jar, if I can do it in a platform independent way.

Re your "crash" what were the symptoms? No response from anything? Did you try closing the "graphing option pane" and reopening it or could you not even do that? There wasn't a "do you want to grant this app permission to do something" window in the background was there? Which version was this exactly? The last release or current from source? If it was the last release, please get the latest. If it was from source, please provide a hash to go with this report. Also, if you don't launch with a console, any stack traces are not visible and there is absolutely no way to diagnose what happened.

Hope that helps a little :-)

BTW, it's easy to get JDK + Maven3 on a win box so you can run the latest by typing "git pull;mvn"

Re: Recognising An OLV Crash - Please Report These

Posted: Tue Mar 06, 2012 2:16 am
by Peter
I didn't try closing the graphing option pane. It still let me drag the boxes around in it so I assume it would have opened and closed. I have the security thing that asks you if you trust this program turned off on this computer. It is version 0.0.2, the latest release. I can get it to do it consistently if the graphing pane is already open from a previous log and I don't let it finish loading a new log before I drag one of the boxes down to be graphed, but that's not what I did the first time. Pretty similar error though.

Re: Recognising An OLV Crash - Please Report These

Posted: Tue Mar 06, 2012 2:29 am
by Fred
If you can install msysgit and maven and run the latest from source and duplicate the error, that would be useful. It's probably/possibly still happening in the latest and it sounds like you know how to generate the conditions. One thing, though, there are known issues with loading second files in terms of memory usage, so it's unlikely to work well for you. Having said that, I do it a lot and suffer no ill effects, so mileage may vary, and may be good.

Re: Recognising An OLV Crash - Please Report These

Posted: Fri May 11, 2012 9:05 am
by Kennybug
OLV log files not displayed correctly in Win32.

My setup in office computer:
WinXP 32 bit with JDK1.7 & maven3

OLV plugin is latest from Ben's GIT with POM fixed 11 days ago

Installation and compile of OLV is smooth without error.
Only one issue is I use mirror on repo1 server (it seems repo server is dead)

All bin, csv files can be loaded correctly. The log is able to run and stop.

But there is no log graphics on display. (or the graph displayed in black color only)
I have changed the color pane and setup. There is still no trace graphic at all.
:indiff:

Then i had tried Fred's GIT version, but it got the same issue.

Re: Recognising An OLV Crash - Please Report These

Posted: Fri May 11, 2012 12:44 pm
by Fred
Hey Kenny,

That's probably not a bug, your data is probably corrupt. Take the data file and run it through the freeemsParser.c file (after compiling it) and compare the output, it will match OLV almost exactly.

How have you recorded this data? You need to use a reliable program with the correct settings or you'll just record garbage.

Fred.

Re: Recognising An OLV Crash - Please Report These

Posted: Fri May 11, 2012 7:25 pm
by Fred
BTW, I've got a man writing a tiny utility so that you can verify that OLV is working properly. It'll be something like this:

Double click application
Type in file name of log
Find output.txt file
Open output.txt file
Copy contents
Paste into forum
Be happy :-)

Fred.

Re: Recognising An OLV Crash - Please Report These

Posted: Sat May 12, 2012 4:39 am
by Kennybug
fred, thanks. will try to run the C program.
I still not get any running engine
My log is from pete:
viewtopic.php?f=62&t=1533&hilit=maven&start=20

I have try another csv file from hot cat, it was a clean csv table. it also not working.
viewtopic.php?f=32&t=1604&p=24413

I think it may due to the latest jdk 1.7.0_u04

will get OLV run on debian & OS X tonite for double check :P

Re: Recognising An OLV Crash - Please Report These

Posted: Sat May 12, 2012 5:27 am
by Fred
Try this one: http://stuff.fredcooke.com/forMal.bin

The firmware and OLV versions have to match or it will just reject them. Look at your console for info. The above linked file should work with the latest OLV from all three repos. The hash is: c43d3f8e796e8a4e80c42668ce51c380eac473fd

Fred.