Download Results returning empty 160B zip archive

OS: macOS

PsychoPy version: 2026.1.3

Standard Standalone Installation: y

URL of experiment:

Do you want it to also run locally? n

What are you trying to achieve? Downloading participant result CSV files from two completed experiments via the “Download Results” button on the Pavlovia dashboard.

What did you try to make it work?

  • Tried downloading results from both projects separately

  • Confirmed both experiments are in RUNNING mode with credits properly consumed (2 credits each)

  • The first participant’s data was successfully downloaded immediately after their session completed

  • Checked the .gitignore file — data/ is excluded to prevent local test runs from being committed. However, this is not the cause: Pavlovia’s “Download Results” pulls from the Pavlovia server directly, not from the GitLab repository. Furthermore, this same .gitignore configuration was in place when previous downloads from SAPC worked correctly.

Link to the most relevant existing thread: Issue with Download Results returning an empty archive despite 36 submissions

What specifically went wrong? Since approximately 11:00 UTC on May 31, 2026, clicking “Download Results” on both projects returns an empty zip archive of only ~160 bytes. Both projects are affected simultaneously, suggesting this is not a project-specific issue. No error message is shown — the zip simply downloads empty.

Has the experiment been set to inactive since data collection?

Is there anything in the commit history on GitLab.pavlovia.org?

Yes, When experiments finished, I have changed the mode.

I add a capture of commit history.

I can’t check any commits excepting mine.

Changing the experiment to inactive deletes the data from the run server.

Did you add data to the .gitignore file manually? This might prevent data being added to GitLab.

My assessment is therefore:

Data wasn’t added to Gitlab as per the edit to .gitignore
Data was deleted from the run server by setting the experiment to inactive.

Does this sound correct?

Thank you so much for your detailed explanation, wakecarter.

Your assessment sounds correct. I now understand that setting the experiment to inactive deleted the data from the run server, and since data/ was excluded in .gitignore, the data was never transferred to GitLab either.

I’ll make sure to always download participant data before changing the experiment status in the future. This was a costly lesson learned.

Thank you again for taking the time to help.