Update instructions to restrict/discourage students from creating new repos

[jcohenadad]

what has happened, is that students have created new repos based on this documentation.

however, the problem is that the new repos are created and pushed without being verified by the admin team.

this is a problem, because some repos were created and they should not have been created.

example: datasets/data_gre-epi

two things to do:

• update the instructions and/or permissions to discourage/restrict students from creating new repos-- maybe just the doc would be sufficient--
• remove this repos: datasets/data_gre-epi

[mguaypaq]

Looking at the configuration of data.neuro.polymtl.ca, I see that in general the master branch is protected for all repositories under datasets/ (that is, basically all repositories), meaning that only @admins users can update the master branch. But currently every user is able to create a new repository, which involves creating the master branch with some existing commits, bypassing (just once) the master branch protection. So if we want full protection for the master branch, I can change the configuration so that only @admins users can create new repositories (this is simple to do).

I've already removed the datasets/data_gre-epi repository. Now I'll change the configuration so that only @admins users can create new repositories under datasets/, and update the wiki documentation to reflect this.

[jcohenadad]

So if we want full protection for the master branch, I can change the configuration so that only @admins users can create new repositories (this is simple to do).

I would be in favor of this solution for now (maybe in the future we will revisit)

I've already removed the datasets/data_gre-epi repository

thank you

[mguaypaq]

I've changed the gitolite configuration, now I just need to update the wiki documentation.

[kousu]

I'm confused about what the policy around datasets should be. Isn't the point of the server to share data? Discouraging seems antithetical.

@mguaypaq explained the problem here was that that datasets didn't pass bids-validator. But the solution to that is to work with the creator to make it pass. And if we decide a repo is useless afterall, we can delete it in one step, like we would with any repo on GitHub.