I want to set the visibility of some of my datasets (e.g. 42584) to "public", but I get an "Error 131" when I try this through the "edit" menu.
Steps/Code to Reproduce
Go to a private dataset owned by the logged in user
Click 'Edit'
Change "Who can view this data" to "Everyone" (or to be more precise, leave it at that value, because the page loads with this field set to "Everyone" already)
Try out different combinations of changing other fields (increasing version, uploading data again, providing url to ARFF file in 'URL' field, filling in / leaving blank the 'State what changed' field)
Click SUBMIT
Expected Results
Visibility is updated: file is publically accessible.
Actual Results
Other notes
Possibly related to https://github.com/openml/OpenML/issues/1048 (and like said there, I am submitting this issue to the OpenML instead of the OpenML.org repo), but that one is trying to add ignored features and therefore changing feature metadata, while I am trying to make the dataset public. One could argue that changing feature metadata means it is in fact a new version of the dataset and should have a new version number, ID, re-uploaded data etc. However, I think there is really no point to the visibility = "private" feature if one does not have the option of changing the visibility to public later without changing the ID, since as far as I can see one can not do anything with a private dataset in that case?
Description
I want to set the visibility of some of my datasets (e.g. 42584) to "public", but I get an "Error 131" when I try this through the "edit" menu.
Steps/Code to Reproduce
Expected Results
Visibility is updated: file is publically accessible.
Actual Results
Other notes
Possibly related to https://github.com/openml/OpenML/issues/1048 (and like said there, I am submitting this issue to the OpenML instead of the OpenML.org repo), but that one is trying to add ignored features and therefore changing feature metadata, while I am trying to make the dataset public. One could argue that changing feature metadata means it is in fact a new version of the dataset and should have a new version number, ID, re-uploaded data etc. However, I think there is really no point to the visibility = "private" feature if one does not have the option of changing the visibility to public later without changing the ID, since as far as I can see one can not do anything with a private dataset in that case?