[open] Random freezing on successive run-throughs
Hi! I'm an undergrad using OpenSesame to run an experiment for my lab. The program I've designed starts out with a loop with several sketchpad items, and then is followed by a loop with several (24) videos. I'm using the legacy back-end for all components. Occasionally the program runs all the way through successfully, but then on the next run through it will freeze halfway through. With each progressive attempt to run the program it freezes earlier (first it's the videos, then it starts freezing in the sketchpad loop). I am closing and re-opening the program every time I run it, since I've read that re-running without re-opening can cause errors, but it makes no difference. Since the problem gets progressively worse every time I try to run it, I'm wondering if there is a buffer or a cache somewhere that is getting filled up and never getting cleared, and this is what causes the freezes. Is there such a thing, and if so, how can I make sure it is regularly cleared? I'm at a loss here.
Other info: I don't get a debug trace or any error messages when it freezes, the program just stops responding and I have to force quit. I've tried switching back-ends, and xpyriment is more stable for canvas items but won't run videos. I tried switching just the canvas items to xpyriment and keeping legacy elsewhere, but my experiment requires the program to run on two monitors, and the xpyriment pieces would not run on the second monitor. I switched back to all-legacy and it worked for maybe about 4 or 5 run throughs, and then it started freezing every time I run it. I've made no changes to the program itself. Please let me know if you have any suggestions! This is also rather time critical, as we will be running experiments all week, so the sooner I can resolve it, the better. Thank you so much for developing OpenSesame and for maintaining this forum!
Best,
Katie
Comments
Hi Katie,
This is very strange, especially that the crashes accumulate even though you close OpenSesame in between successive runs. I suspect, like you, that there is a buffering issue at play, but I couldn't say exactly what or how. I forwarded the problem to @dschreij, who is the developer of the plug-in. Perhaps he has an idea.
Do you run the experiment through the OpenSesame GUI or through opensesamerun? If you are not doing so already, it might be worth a shot to try opensesamerun. This will avoid loading a bunch of libraries that are required for the GUI.
Cheers,
Sebastiaan
Check out SigmundAI.eu for our OpenSesame AI assistant!
Hi Katie,
Communication of OpenSesame with the vlc libraries is very rudamentary and doesn't offer me a lot of control or insight on how the video is presented on screen or if a buffer is overflowing. I'll still try to see if I can make the vlc plugin use less memory and clean up better after itself, which hopefully will fix this problem.
I also haven't been able to reproduce the problem, so it's hard for me to pinpoint the source of your troubles. There has been a report of this problem on the forums before, but it appears to only occur incidentally. Have you tried running your experiment on a different pc? How much memory and (free!) hard disk space does your current test machine have?
Otherwise @sebastiaan, would be useful for Katie to give the new prerelease in combination with multiprocessing a spin? The piece of information "Occasionally the program runs all the way through successfully, but then on the next run through it will freeze halfway through." gives me the impression it might indeed be a garbage collection problem (although Katie also reported she has tried to close and reopen OS, which didn't help and thus undermines this assumption a bit).
Katie,
I have taken better care that the plugin cleans up after itself, hopefully causing it to be more economic with available memory. Could you download the latest version of the plugin from https://github.com/dschreij/media_player_vlc/releases/tag/1.0 and tell us if it works better for you?