Pavlovia storing all key_resp_stimuli.corr as wrong

[GPesi]

URL of experiment: https://gitlab.pavlovia.org/GabPes/perspectivetaking

Description of the problem:
The experiment is working well offline and correct answers are correctly stored. I managed to push the experiment on pavlovia however I have just noticed looking at the ouptut file that all the column key_resp_stimuli.corr gives back a value “0” even when the answer given by the participant is correct and therefore it should be “1”

Thanks a lot in advance for your support.

[jonathan.kominsky]

Please make your experiment public and I’ll have a look, but my blind guess is that it’s going to be an issue with how keys are identified in JS versus Python.

[GPesi]

Hi Jonathan,
Thanks for your answer. You are correct. I had to amended the correct answer box and push everything again on Gitlab and Pavlovia.
I did this intially but the pavlovia was still coding them wrong.
After having looked into this, i found out that despite the new files were all uploaded on Gitlab, pavlovia was still running the old files with the wrong key coding.

I don’t know why this was happening, however to sort this out I deleted the experiment and created a new one with a different name.
I tried initially to create a new one with the same name but Pavlovia was still looking for the old wrong files.

[jonathan.kominsky]

I’m worried that Pavlovia might be a touch overburdened right now, there seem to be several issues along these lines on the forum.

@apitiot could load cause the git and sync issues I’m seeing in this thread and others? See also:

Local changes succesfully pushed to gitlab, but pavlovia reports merging conflict

I thought about it. But assuming that this error happens again, I would need to upload the whole experiment again, creating repository after repository. I am convinced that there is something I don’t see and someone else will.

Unspecified JavaScript error

I tracked down the repository, and something weird is going on here. You actually have multiple copies of the experiment code. One of them is in the HTML folder, where it should be, while the other is just in the main folder, where it should not be. When did you add a gender field to your participant information? That will help me figure out which one is the more recent version.

which also are possibly (but not definitely) the result of weird sync behavior.

[jon]

I don’t think this is caused by the server being overloaded - git is pretty good at handling large numbers of calls simultaneously - but the truth is we’re puzzled about how such merge conflicts occasionally arise. It’s a nut for us to crack when we aren’t quite so buried in a world that just dramatically changed its work patterns!