Unspecified javascript error for an experiment so simple it can't fail

URL of experiment: https://run.pavlovia.org/CorentinGonthier/echo/
URL of public repository: Sign in · GitLab

Hi there,

I encounter unspecified javascript errors on a regular basis, sometimes on scripts whose previous version worked flawlessly. I must be doing something wrong, but I really can’t figure what it is.

This particular case has me quickly losing my sanity. I have removed everything in my experiment except for one routine that includes two image components. The routine is in a loop. The loop has two trials.

Trial 1 : image A.png is displayed in image component 1 and image B.png is displayed in image component 2. This works.
Trial 2 : image B.png is displayed in image component 1 and image A.png is displayed in image component 2. This throws an unspecified javascript error.

This is trial 2 :

The same error appears on Chrome and Edge, on both Windows 7 and Windows 10. This is set to “use PsychoPy version” = [empty]. There is no code component or audio or anything messy in the experiment whatsoever, the images are .png of dimension 189x189 pixels, there are no duplicate or reserved variable names, etc etc.

Any help with this voodoo error is appreciated. Bonus points if you can help me pinpoint why the first screen collecting information has the fields all messed up rather than neatly aligned vertically:

Thanks a lot !

Your ‘messed-up fields’ are due to a missing CSS stylesheet.

See here:

Oh great, thank you so much! So if I understand correctly, this was a missing resource on the server side, right? Nothing to fix on my end?

I was able to trace the origin of the “unspecified javascript error” of my original post. The script specifically crashes when reversing two files for two consecutive image components. In other words:

No idea why this would be the case though. Easy, but cumbersome to fix, by creating multiple copies of the image files so that image A never appears in component 2 (e.g. create A1, B1, A2, B2.png).

Hey!

I tried to find out what’s going wrong with your experiment, but I haven’t figured it out. Asking around a bit.

Best, Thomas

Hi @CGR2, thanks for flagging, sorry, these both look like bugs on our end! I have created a support fork to demonstrate potential fixes, but there should be a new release coming soon to that effect. Please let me know if you need more details, s.

2 Likes

Hi @CGR2, FWIW PsychoPy latest now includes the relevant fix and you should no longer need to rely on the custom lib patch. Please let me know if you hit any hurdles post upgrade, x

Great, thanks !