PriiDash(TM): Enhanced Instrument Panel / Data Logger - Now Open Sourced!

Re: 2012-04-19 new release

Same thing with this version, reinstalled XP on the little netbook.. reinstalled cygwin as in the instructions.. ran the binary "bad address" just like before.. until I can get the binary running not sure its worth even trying to recompile it since that had so many issues.

Cygwin installed, Com7 = USB based ELM, F: drive setup.... no idea any more hints?

 

Re: 2012-04-19 new release

$ ./priidash.exe
-bash: ./priidash.exe: Bad Address


That's it unfortunately, no more info. Is it possible there is something else in cygwin that needs to be installed that isn't in the Readme?

 

I have been able to run Priidash on a laptop using XP and 2 different netbooks ( a Dell and a Asus eePC running Windows 7) and the exe runs fine in Cygwin-X.
You may want to try it on a different Laptop or even a Desktop just to see if it will load for S&G.
I sometimes compile on a Dell desktop for the higher speed of compliation, load it there just to check before moving it to the Asus.

 

For the sake of completeness I will just list what I have done, and what I have encountered. In software engineering it is often the case that there are missing steps or “things you just know” often referred to as tribal-knowledge which aren’t reflected in the setup steps. When I walk SQA or Human Safety through the sign off on a project at work for Verification and Validation it is not uncommon for things to be omitted because as the creator (or the group of creators in the case of a large team) there are things you just know to do because you either designed the project or you have been working with it for months or years so you don’t think about it. I am under the impression that this is the case here, I have lead several projects where something like this has happened in the past so it’s no big deal it just makes it hard to duplicate later.

First, let me say that I have tried this on 4 different physical machines and 3 virtual boxes none of which I could get to work.

1) [ATTEMPT#1] I tried my main personal dev computer first (an m17x R3), getting it setup per the instructions and attempted to recompile the source from scratch in a win7 environment using Cygwin. I ran into some issues with includes, and const/non-const pointers. I was able to work around those issues, however hit a wall at an object reference (FL if I remember correction) which wasn’t allowing it to compile even though there was no reference to an FL object directly visible in the code.

2) [ATTEMPT#2] I then tried to get it setup on my main development Ubuntu box (this was an existing box as well I use for lots of other linux based development). I ran into the exact same compile time issues, and I wasn’t able to resolve it. This included the creation of some defines like “Ubuntu” to properly pull in some linux headers, among other things.

3) [ATTEMPT#3] After this point I decided I would try and get your binary to work first, and since that only existed for the Cygwin based system I decided to put the native linux setup on the back burner. However, this is the eventual place I will take this development effort as its much more flexible and powerful for these types of low level serial interfaces.

4) [ATTEMPT#4] Since I wanted to ensure I did EXACTLY the setup you had cygwin wise, I uninstalled it and only listed the items which you call out in your readme. I had a preexisting cygwin install (I use it for scripting etc in windows) so I wanted to take that variable out of the equation. I setup the ELM device to be COM7, setup the F drive, along with the paths we needed to do have done etc. I was presented with the bash error of “Bad Address”. Unfortunately this is a common error and not very descriptive to help debug the problem.

5) [ATTEMPT#5] So now I was just trying to control the variables I could, perhaps it was my m17x and there was something about it specifically that was causing issues. So I dug out an old Compaq/HP laptop from a few years ago which had XP on it. Set it up from scratch using the instructions only putting on there what was requested (cygwin, F drive, COM setup etc) and I was greeted with the same error

6) [ATTEMPT#6] Since you were using a “netbook” I had an AcerOne which I got for my wife when she went to Germany she wasn’t using anymore so I borrowed that and ghost imaged it (so I could restore it when I was done). I then put a fresh XP install on it, along with a fresh setup of the readme directed text. Again the Same result.
7) [ATTEMPT#8] I tried a virtualbox setup of XP running ON that XP machine. Same result.

8) [ATTEMPT#9] One final try, I said ok at this point I am at least 70% sure there is some step I am missing so let me try the “compile from scratch” approach one more time. I blew away that Acer One XP install and put Ubuntu on it got it all setup and then tried to recompile I ran into the exact same errors as #2.

So I can provide whatever logs etc you need to try and diagnose this but I am at a loss at this point. I do NOT doubt this thing work and I really want to use it and expand upon it with some logging/analysis layers as well. I am all ears for what you need to figure this out. Perhaps its worth one of you guys who has this all working to start a virtual box from scratch or something and JUST do the steps that are outlined in the readme to see what’s missing… what is the key piece that wasn’t captured. I think its unlikely that all these different hosts both virtual and physical hardware could be defective.

How about this, could one of you also export your /etc/setup/installed.db file so I can compare it with mine and see if there are any packages you guys have which I don't. My guess is there will be but maybe one of them is the key. There could also be additional packages for many other reasons of course.

 

Ok so.. first things first. I don't see anything blatently wrong with the *.db file. Now for the "test"


$ ./welcome_test.exe
Welcome ...

So seems to work.

Now for the one you sent me

$ ./priidash20120419_download.exe
-bash: ./priidash20120419_download.exe: Bad address


MD5s match what I have too, on the welcome and the above binary (compared to the one I downloaded I mean)

 

Ok since the .db didn't have anyhign that really stood out I did some perl processing on it. Then ran that through beyond compare to look for differences (we have lots of different things but that's not the point.. something related to flk or boost was what I was looking for on a hunch)...

So I ended up with these two reports:
My FLK entries:
libfltk1.1-gdi libfltk1.1-gdi-1.1.10-1.tar.bz2 0
libfltk1.3 libfltk1.3-1.3.1.9285-1.tar.bz2 0
libfltk-devel libfltk-devel-1.1.10-1.tar.bz2 0
libfltk-doc libfltk-doc-1.1.10-1.tar.bz2 0

Yours:
libfltk1.1-gdi libfltk1.1-gdi-1.1.10-1.tar.bz2 0
libfltk1.1 libfltk1.1-1.1.10-1.tar.bz2 0
libfltk-devel libfltk-devel-1.1.10-1.tar.bz2 0
libfltk-doc libfltk-doc-1.1.10-1.tar.bz

Notice your lib is 1.1 and mine is 1.3? Perhaps this is one of the issues 1.3 fltk might have issues with your stuff? (I then installed YOUR 1.1 version)

Ok now for boost:
Mine:
libboost libboost-1.33.1-4.tar.bz2 0
libboost_python1.48 libboost_python1.48-1.48.0-1.tar.bz2 0
libboost1.48 libboost1.48-1.48.0-1.tar.bz2 0
libboost-devel libboost-devel-1.48.0-1.tar.bz2 0

Yours:
libboost libboost-1.33.1-4.tar.bz2 0
libboost_python1.43 libboost_python1.43-1.43.0-1.tar.bz2 0
libboost1.43 libboost1.43-1.43.0-1.tar.bz2 0
libboost-devel libboost-devel-1.43.0-1.tar.bz2 0

Ok again notice you have 1.43 and I have 1.48. So I installed the 1.43 verison..

AND TADA!
Everything works now! No idea why those two newer versions where the issue howver you might add to your Readme.txt to use those two versions for now until you figure out whats up with the newer ones. Obviously since I was installing from scratch and following the instructions, cygwin just grabed the "latest and greatest". So yay it works now! I figured it was something simple, just a matter of finding it.

 

Ok so after getting the GUI to open up I can't seem to get any data from the ELM device. I can click "go" and the times start ticking away in the gauges but no data moves on anything. doing a simple cat on the /dev/ttyS6 shows no data flowing (though I have had hit and miss with just dumping serial stuff that way before)

Any chance there is another silly little thing missing? Also not seeing any of the ELM lights flashing/moving etc.

 

Have a look at the log file which should indicate if the initalization commands got to the obd2 interpreter.
FAQ outlines how to set the UART speed using stnterm.
STN based parts 2M.
True ELM limited to 500K.
Clones??

 

Hi.
I would like to run Priidash on an android tablet. Any way?