Users circumventing new repo creation limitation #465

Open
opened 2026-03-23 08:45:16 +00:00 by mattia · 10 comments

Summary

Users can circumvent new repo creation block by cloning existing repos

Background

It has been reported on matrix chat that a user successfully managed to create a repository for their personal use:
https://forge.fedoraproject.org/mabbs/blog
That repo seems to have been forked by this very one, probably the user managed to erase the entire existing git history and put their stuff in it.

Details

Since it is supposed that the new Fedora Forge has to be used only for Fedora related work, this is something that needs to be fixed.

Summary

Prevent users to create repos for their personal stuff

### Summary Users can circumvent new repo creation block by cloning existing repos ### Background It has been reported on matrix chat that a user successfully managed to create a repository for their personal use: https://forge.fedoraproject.org/mabbs/blog That repo seems to have been forked by this very one, probably the user managed to erase the entire existing git history and put their stuff in it. ### Details Since it is supposed that the new Fedora Forge has to be used only for Fedora related work, this is something that needs to be fixed. ### Summary Prevent users to create repos for their personal stuff
Owner

Yeah, this was a known "workaround" to the setup.

Now it has happened We will need to come up with a way to manage this with moderation.

Yeah, this was a known "workaround" to the setup. Now it has happened We will need to come up with a way to manage this with moderation.
Member

@ryanlerch, can / should we restrict forking to folks only belonging to the namespace that they are trying to fork the repository from? Like, if I am not a part of the science namespace, that would restrict me from being able to fork science/stem unless I end up being a part of that group, which has to happen manually since the sponsors can only add me. Assuming that what I stated is indeed possible, that is.

@ryanlerch, can / should we restrict forking to folks only belonging to the namespace that they are trying to fork the repository from? Like, if I am not a part of the `science` namespace, that would restrict me from being able to fork `science/stem` unless I end up being a part of that group, which has to happen manually since the sponsors can only add me. Assuming that what I stated is indeed possible, that is.
Owner

Sadly, That will be too restrictive, as a contributor that wants to contribute to a project would have to become a member first.

Maybe we can restrict forking to CLA+1 somehow?

Forking a repo and replacing the content with your own because you cant make a fresh repo is a little shady IMHO.

Sadly, That will be too restrictive, as a contributor that wants to contribute to a project would have to become a member first. Maybe we can restrict forking to CLA+1 somehow? Forking a repo and replacing the content with your own because you cant make a fresh repo is a little shady IMHO.
Member

CLA+1 acceptance-based restriction feels right. Not too restrictive, not too allowing.

I do suggest nuking the repository, it is a repository of an already existing repository https://github.com/Mabbs/mabbs.github.io, and on a superficial skim, has nothing to do with the Fedora Project at all.

CLA+1 acceptance-based restriction feels right. Not too restrictive, not too allowing. I do suggest nuking the repository, it is a repository of an already existing repository https://github.com/Mabbs/mabbs.github.io, and on a superficial skim, has nothing to do with the Fedora Project at all.
ryanlerch added this to the Sprint 17 project 2026-03-23 10:29:07 +00:00

I'd suggest talking to the user, perhaps they misunderstood things?

But in any case, it might be worth waiting to see how big a problem this is before investing in a lot of work to avoid it, or restricting users?

I'd suggest talking to the user, perhaps they misunderstood things? But in any case, it might be worth waiting to see how big a problem this is before investing in a lot of work to avoid it, or restricting users?

@t0xic0der wrote in #465 (comment):

@ryanlerch, can / should we restrict forking to folks only belonging to the namespace that they are trying to fork the repository from? Like, if I am not a part of the science namespace, that would restrict me from being able to fork science/stem unless I end up being a part of that group, which has to happen manually since the sponsors can only add me. Assuming that what I stated is indeed possible, that is.

For me this sounds a bit too restrictive..

What if one wants to fix some documentation, let's use this as example: fesco/docs#122

I'm not part of fesco so instead of just creating PR I would need to either:

  • become a member, or
  • create an issue and wait for someone from the group to pick it up and fix it

where first one is probably not going to happen

@t0xic0der wrote in https://forge.fedoraproject.org/forge/forge/issues/465#issuecomment-581249: > @ryanlerch, can / should we restrict forking to folks only belonging to the namespace that they are trying to fork the repository from? Like, if I am not a part of the `science` namespace, that would restrict me from being able to fork `science/stem` unless I end up being a part of that group, which has to happen manually since the sponsors can only add me. Assuming that what I stated is indeed possible, that is. For me this sounds a bit too restrictive.. What if one wants to fix some documentation, let's use this as example: https://forge.fedoraproject.org/fesco/docs/pulls/122 I'm not part of `fesco` so instead of just creating PR I would need to either: - become a member, or - create an issue and wait for someone from the group to pick it up and fix it where first one is probably not going to happen
Author

I think the requirement on CLA+1 signing is acceptable. After all, the reason behind forking a project in the new forge should be to contribute to Fedora in some way.

I think the requirement on CLA+1 signing is acceptable. After all, the reason behind forking a project in the new forge should be to contribute to Fedora in some way.

Please don't require group memberships and create extra barriers to contribution for people getting started in Fedora.

Please don't require group memberships and create extra barriers to contribution for people getting started in Fedora.
Owner

i'm leaning towards even CLA+1 being too restrictive still.

First change, going to add a note to the fork page stating that this behaviour is not permitted. -- #466

Then going to write a script or come up with a way to identify forks that are misusing this.

Also, I will reach out to the individual user here and inform them that what they have done is not permitted on Fedora Forge.

i'm leaning towards even CLA+1 being too restrictive still. First change, going to add a note to the fork page stating that this behaviour is not permitted. -- https://forge.fedoraproject.org/forge/forge/issues/466 Then going to write a script or come up with a way to identify forks that are misusing this. Also, I will reach out to the individual user here and inform them that what they have done is not permitted on Fedora Forge.
Contributor

I agree with @kevin, @gotmax23, and @ryanlerch. "FPCA+1" as a requirement makes sense for dist-git, but not Fedora Forge. I am thinking of the wave of Outreachy intern applicants we have currently surging into our infrastructure and tools. These folks are brand new and would immediately hit this issue where they wouldn't be able to fork an important repo, and the mentoring team wouldn't be able to support the applicant easily. The workflow of trying to teach a new open source contributor how to do a Pull Request across two different git forges is a time investment I would prefer not to have to make. Fedora Mentored Projects would suffer if forking is disabled and limited to this special group only.

@kevin wrote…
I'd suggest talking to the user, perhaps they misunderstood things?

@ryanlerch wrote…
Also, I will reach out to the individual user here and inform them that what they have done is not permitted on Fedora Forge.

I think @kevin was on the right track. There are clear usage terms about using Fedora Forge. There are appropriate uses and inappropriate usages. The correct action is to have a conversation with the user, make sure they are aware that this is an intentional limitation, and then take action based on the outcome of the conversation.

@ryanlerch wrote…
First change, going to add a note to the fork page stating that this behaviour is not permitted. -- #466

I like this idea a lot. I have a hunch that most of the people attempting to get around this requirement likely are unaware that this is a limitation put in place on purpose.

@ryanlerch wrote…
Then going to write a script or come up with a way to identify forks that are misusing this.

👍

I agree with @kevin, @gotmax23, and @ryanlerch. "FPCA+1" as a requirement makes sense for dist-git, but not Fedora Forge. I am thinking of the wave of Outreachy intern applicants we have currently surging into our infrastructure and tools. These folks are brand new and would immediately hit this issue where they wouldn't be able to fork an important repo, and the mentoring team wouldn't be able to support the applicant easily. The workflow of trying to teach a new open source contributor how to do a Pull Request across two different git forges is a time investment I would prefer not to have to make. Fedora Mentored Projects would suffer if forking is disabled and limited to this special group only. > _**@kevin wrote…**_ > I'd suggest talking to the user, perhaps they misunderstood things? > _**@ryanlerch wrote…**_ > Also, I will reach out to the individual user here and inform them that what they have done is not permitted on Fedora Forge. I think @kevin was on the right track. There are clear usage terms about using Fedora Forge. There are appropriate uses and inappropriate usages. The correct action is to have a conversation with the user, make sure they are aware that this is an intentional limitation, and then take action based on the outcome of the conversation. > _**@ryanlerch wrote…**_ > First change, going to add a note to the fork page stating that this behaviour is not permitted. -- #466 I like this idea a lot. I have a hunch that most of the people attempting to get around this requirement likely are unaware that this is a limitation put in place on purpose. > _**@ryanlerch wrote…**_ > Then going to write a script or come up with a way to identify forks that are misusing this. 👍
Sign in to join this conversation.
No milestone
No project
No assignees
7 participants
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
forge/forge#465
No description provided.