CVI 8.0 Linux runtime crashing in PlotStripChart

[nickb]

Just to follow up on this issue, we've investigated the crash, and determined that the UIR file was corrupt.  There was an invalid value for the Line Style of one of your traces that was causing problems in Linux that were being ignored in the Windows implementation.  To resolve the issue with your UIR, you should just need to modify the linestyles for all your traces in the UI Editor.  Change them to something different than they are currently, and then back to what you intend them to be.  That should resolve the issue for you.

NickB

National Instruments

[wally_666]

Thanks for confirming what I had speculated in the reply I marked as "solution".  Since most of the default attributes for this control had been tweaked, I cut and pasted a version of it from a UIR file in a different project that used a version of it.

--wally.

[gdargaud]

Sorry, but this bug is biting me really hard right now. I've just wasted 2 full days tracking it down before I saw it here.

And the solution provided doesn't work (I saved as TUI and reloaded, I copy-pasted a Strip control from another working project...).

Isn't there a fix by now ? I'm using CVI 10 for Linux.

It's infuriating because it shows in only a very limited set of circumstances: 2 traces (but not 4) and negative values in the data and auto-scale axis and... And it doesn't crash the program, just make it hang in some inner loop of the RunUserInterface library so no single line appears to be the cause.

I you really need I can simply my project to make a sample case, but it's gonna take a while.

[Kevin_B.]

My guess is that you're experiencing a different issue even if it's somewhat similar. We weren't entirely sure what caused the corrupted UIR that wally had, but on the off chance it was a CVI had corrupted the UIR somehow we eliminated that particular issue from occuring again (his UIR had a bogus value for line-style which is now being coerced so it won't crash in 2010).

If you could cobble something together a project for us to take a look at, it'd help us a great deal with figuring out what's going wrong for you.

[gdargaud]

Hello Kevin,

I spent some time simplifying my project to isolate the problem, but it went away...

So here's the complete project (actually only one .c is relevant). Make sure to read the README for the compilation command and the key to press to activate the problem.

Let me know what you find because I'm stuck for now.

Thanks

[gdargaud]

Just a little extra info, I've changed the file /usr/local/natinst/cvi2010/bin/compile.ecc as such:

< C99_FLAGS = -std=gnu99
> C99_FLAGS = -std=c99
< Debug Opt = -ggdb
> Debug Opt = -g

[Kevin_B.]

I'll have somebody here give it a look. Were you ever able to test this issue on Windows or was it strictly happening on Linux?

[Guillaume Dargaud]

> I'll have somebody here give it a look. Were you ever able to test this
> issue on Windows or was it strictly happening on Linux?

Thanks.
Works fine on Windows.
--
Guillaume Dargaud
http://www.gdargaud.net/

[gdargaud]

Anybody had the time to look it up ?!?

[Kevin_B.]

Hey Guillaume,

I had somebody with access to a Linux machine take a look at the code, and he initially had some issues getting it to plot anything. I looked at it on Windows and had to make some modifications to the source to get it to work there (the event logging was messing things up, at least on Windows, because the STDIO window was stealing the focus on every click/keypress on the panel). Once that was done it worked correctly, just like you said. I gave him some suggestions on what to try for Linux, but I think the e-mail might've gotten missed over the Holiday break. I've pinged him again, and hopefully we'll have some answers for you in a day or two.

I apologize for the delay.