| Reference | Downloads | Github

Pavlovia error: We could not find a record for experiment designer: undefined

URL of experiment:

Description of the problem: For starters this is the exact same experiment I have already uploaded titled ( which runs perfectly. I have not changed anything as I fully copied the original version. Here is my problem:

When I try to switch the experiment from ‘Inactive’ to ‘Piloting’ on Pavlovia I get the following error.

With the full description of the error being:
Full description:

(First part removed as it had personal info)
“name”: “CPB2”,
“nbForks”: 0,
“nbStars”: 0,
“pathWithNamespace”: “Mi_Lab/cpb2”,
“pavloviaScore”: 0,
“platform”: “PSYCHOJS”,
“recruitment”: {
“policy”: {
“type”: “URL”,
“singleRun”: false
“runMode”: “LICENSE”,
“saveFormat”: “CSV”,
“status”: “INACTIVE”,
“status2”: “PILOTING”,
“updateDate”: “2020-05-13 23:15:34”,
“userIds”: [
“visibility”: “private”
“error”: “unable to create link to PsychoJS library folder: [Errno 17] File exists: ‘/var/www/’ → ‘/var/www/’”,
“errorCode”: null,
“httpStatusCode”: 500

I then get this message when I try to pilot the study:

Absolutely any advice is welcome! Thank you.

1 Like

I have the same issue at the moment. :confused:

unable to create link to PsychoJS library folder: [Errno 17] File exists: '/var/www/' -> '/var/www/'

Did you manage to resolve it?

For me, the problem was the html/lib folder which I had to delete. I think the folder got created when I tried to debug locally with a custom .gitignore file.
See this post: Unable to create link to PsychoJS library folder

1 Like

Okay, right. Our options to prevent recurrence would be to implement one of:

  • raise psychopy.alerts.alert if local debug gets used and html/lib is not in .gitignore (this would be relatively easy to write)
  • change the operation of local debug so that it occurs in a separate location entirely, such as having a psychopy/virtualServer folder (then copy the project files to that folder when needed)
1 Like

Would option 2 have the benefits of a) reusing the libraries for multiple experiments and b) failing due to missing resources in the same way that the experiment would fail online?

Yes, I think both those points would be true, but it would also mean the experiment files would have to be duplicated, increasing disk usage (at least on windows which doesn’t have “symbolic link” files) and possibly increasing work (detecting if files need updating with each run?)