MWorks: difficult possible bug on variable change

Support
1

Chris,

I’m going to give you another bug report, but I think it will be hard to come up with a test case for it.

There are times when I change a variable in the Variables window, and the event makes it to the Matlab bridge, but the variable change is not seen by the XML experiment. This has happened to me a few times now, but it’s very sporadic and I can’t figure out when it is likely to happen. I have about 120 variables in the experiment; it’s happened to both integer and float variable types. Anything I can do here to try to help diagnose it?

Thank you,
Mark

2

Hi Mark,

I just spent some time playing with the variables window, and I wasn’t able to reproduce this problem.

Are you certain the variable doesn’t change on the server side? If that’s the case, then what you should see is that as soon as your hit return in the variables window, the variable reverts to its previous value and stays there, as if your input had just been discarded. Is that what you observe?

If yes, then can you provide more details about how the variables in question are used in your experiment? For example, are they used as parameters to I/O devices or stimuli, or do they hold experiment statistics or other data? Basically, I’m wondering where else these variables might be getting updated.

It’s also possible that the client is dropping events, although it will be difficult to verify that without a reproducible test case.

Thanks,
Chris

3

Chris -

Thanks for your time on this. I’m remembering one case where a variable change made it to the Matlab window but not in the client. This sort of thing is very rare. The variable is a persistent float, but it goes through several “if” clauses based on other variable values to assign a final floating-point parameter to a stimulus for display. All of my variables have python bridge callbacks bound to them. Other than that they’re not used as parameters to IO devices or stimuli or other plugins.

Are you certain the variable doesn’t change on the server side? If that’s the case, then what you should see is that as soon as your hit return in the variables

Hmm. I will watch for this. When would this happen?

I appreciate you looking into this but I agree that it will be very hard for you to do anything without a testcase. I wouldn’t spend any more time on it. I’ll keep my eyes open and if it happens again I’ll try to trace exactly which variables and which values are affected.

Best
Mark