People freaking out that is considering to delete inactive free-tier repositories [1]... I mean what did you expect? Someone to host all your stuff for free forever?

Github will do exactly the same if that's at any point convenient to them. You are responsible for keeping your projects online/available/backed-up, if you don't have a contract with someone to do it for you.

[1] (so far that seems to be a still just a rumor/leak? No official confirmation afaik.)

If that's a smart move for gitlab or not is a different question and I don't know the answer to that.

But giving multi-month notice on something that you are providing for free sounds pretty fair to me.

I am monthly donating to #Codeberg and I hope they collect enough money to keep their gead above water, but after reading your toot and the linked article, I went and checked #Gitlab prices and I'm shocked with the tiers they have. $228 for annual subscription! 🤯 I certainly would rather pay Microsoft $44/year or $180 to Bitbucket than to pay $228 to Gitlab!

They could at least create a new budget-friendly tier!! I'll move out my repos first thing tomorrow!

@Mehrad Agreed to that. But the good thing is that self-hosting gitlab is really easy and transferring projects between instances works reasonably well.

But yeah, I wish gitlab would have lower-priced offers as well, also for small-businesses which can't really justify the price over a self-hosted foss-only version 🤷.

@Mehrad they had a $48/year tier until last year, but got rid of that

@HeNeArXn Seems they are really going to the "right" direction!

The only good things came out of this is that now I know I have to:
1. donate **more** to @codeberg [Done]

2. add one more layer of redundancy [this weekend]

3. Continue my work on (Free and Libre OpenSource Software Archiving and Monitoring) which is a project I started with @chriscz, currently living on Codeberg in a private organization until it is mature enough to be publicly released. [in progress]

@Bubu I didn't see any link to an official statement either.
I guess they are looking at the public response to the rumor, before publishing something.

@Bubu An update on GitLab's plan to start deleting "inactive" repos:

TL;DR they responded to community pressure following the critical coverage in the tech press and have cancelled the plan.



@strypey @Bubu thanks for the link, but as the article said, I now have even more questions as this has just added an extra layer of vagueness an uncertainty. I at least would create a mirror on Codeberg or spin a Gitea on a small VPS until I decide on what to do with this mess.

I understand the extra cost for their business, but the way forward is to create a tier between Free and Premium that is <$20/year and give users nothing extra other than 50 inactive repo

Given that Git was designed to decentralize development, it's intriguing that repo hubs like GH and GL have been all to induce such dependence. I recommend keeping local backups of all software on public repos, your own and anything it depends on as a minimum. Having a couple of official public homes for each repo is probably a good idea too.


#git #SoftwareDevelopment #FreeCode

@Bubu Wow, I just realized that some of the code for my scientific articles are living on #Gitlab and (of course) have not seen an update for ~3 years! Deletion of repos can hurt science as well! I have no way to update the URL of these repos in those published articles!

@Mehrad thanks for reminding us. I remembered my paper was at GitLab also. I should update it now. Thanks again!


@Mehrad @Bubu that was exactly my point here My Snort machine learning preprocessor hasn't seen an update in almost 12 years, yet I still receive emails every now and then from researchers who are using it. My C library for self-organizing maps hasn't been updated in a decade either, yet it gets regularly starred/forked.

Projects that don't receive updates in months/years and yet are still actively used by people years down the line are the norm in academia. By opting (or even considering) aggressive retention policies on open-source projects created by free accounts #Gitlab is basically favouring profitability over their duty of preserving scientific and technological knowledge.

The only good thing here is that, now that we know their intentions, we have plenty of time to move our content somewhere else. Platforms that don't value the content that people create on them deserve to be barren lands.


Just one quick comment though regarding this part:

> Gitlab is basically favouring profitability over their duty of preserving scientific and technological knowledge.

Gitlab has zero duties regarding "preserving scientific and technological knowledge". The are a for[e]profit company. This is one of the reasons that @chriscz and I founded FLOSSAM. because we believe: "Free and Libre OpenSource Software (FLOSS) should be considered as world-wide heritage." Wanna join the cause?

@blacklight @Mehrad @Bubu

What duty? There's no duty here. They're a company, not a university or museum.

@fossil @Mehrad @Bubu when you are a company that hosts open-source code that other companies and researchers rely on, you have a duty of preserving that content. If you don't, if you decide to arbitrarily remove projects based on your business priorities, then you don't deserve to be in that job.

@blacklight @Mehrad @Bubu

Bullshit. If people have code they want to preserve, and choose to host it on the free tier of a for-profit company, they deserve everything they get.

Let's keep it civil ! Calm down. We are on mastodon to discuss, not to fight. Explain your logic and the other party has the freedom to read it, digest it, and accept it if they agree with your logic.

@blacklight @Bubu

@fossil @Mehrad @Bubu do you have any idea of how many npm, pip or Go dependencies in widely used projects point to repositories that haven't been updated in more than a year, and are often maintained by a single developer? Can you imagine the cost that taking down those repos overnight would have on a lot of production software?

Do you know how many academic notebooks and libraries are hosted in repos that haven't been updated in a long time? And yet a few years later a Ph.D or a graduate student finds them, uses them and maybe writes a couple of papers on the findings? Can you imagine the negative impact on scientific research that taking down those repos just because they haven't been updated would have?

These are all things that you must take into account if you're a company that hosts open-source code. If you don't take those things into account, and you pick profitability over responsibility for the availability of the content that you host, then you're in the wrong job, period.

And fair enough, those who are in free tiers shouldn't expect to have everything for free. But they shouldn't trust a company that claims the right to delete their hard work on an arbitrary retention basis either. That's why now it's more important than ever to #GiveUpGitlab.

@blacklight @Mehrad @Bubu

"These are all things that you must take into account if you're a company that hosts open-source code."

No. No. A thousand times no.

Those are the things that you must take into account before choosing a company to host your code.

A for-profit company has absolutely zero obligation to people that are not paying it anything.

If you want your code to survive forever, pay for hosting.


@fossil @Mehrad @Bubu you can also self-host, like I've been doing for years, not necessarily pay for hosting. But this is not the point.

The point is there are *already* thousands of useful projects on and Github that haven't been updated in a year or more, and yet it's fundamental to preserve them.

Some are used in goddamn deep npm dependency graphs that power big production software, without developers on either side even being aware of that, and shutting down the repos means to break production stuff.

Some of them contain code used in research papers. Once the paper is out they are unlikely to receive many updates. Still, it's fundamental to preserve them, because maybe 10 years down the line the conclusions of that paper become suddenly relevant for a bunch of reasons, and it's crucial to be able to replicate the results - and you can't do it if Gitlab or Github took down the repo because it was inactive.

Or course we can tell people "if you want your code to be preserved forever, then pay for your hosting" (and that would already be unfair, because then the only code that can be preserved is that of people who can either pay for hosting or self-host). But that doesn't answer the question: what are we going to do if Gitlab takes down for inactivity a repo that, although inactive, is crucial either to our infrastructure, to other businesses, or to scientific research? Gitlab can't just shrug it out and say "it's profit baby, I don't care". They have a freaking moral responsibility to ensure that other production code won't break, and that their retention policies won't prevent the next big scientific discovery.

@Mehrad @Bubu What can we realistically do about this? Pipe all URLs through an easily and cheaply self-hostable URL shortener to make repository migrations possible afterwards? Aggressively archive all links, maybe through

@piegames as it's only said about repos not updated for a long time, it might be a "rescue task" to place a canary into some subdirectory (e.g. .gitlab/) which just contains a last-updated-date and is pushed weekly (with a fresh date). Keeps the repo alive – though it might cause confusion concerning when the *real* content was updated last… So maybe a wiki page, or an issue (which is weekly opened and closed)? @Mehrad @Bubu

@IzzyOnDroid @Mehrad @Bubu Yeah that's maybe not the best solution. If we could get GitLab to auto-archive repositories somewhere instead (under the assumption that a static HTML file host is a lot cheaper than a full Git forge), that would already be something.

@IzzyOnDroid @piegames @Bubu

Gitlab has an API system with which one can do many things to keep a repo "active", but I personally can no longer trust Gitlab as the API can also change at any time.

@piegames @Bubu there are quite a few things we can do, one of which is some sort or publicly accessible archiving with DOI binding to specific commits. I perhaps should add some more modification to the RFC and implementation of #FLOSSAM initiative I'm working on.

Ultimately we should have a permanent, reliable, citable storage for #academic software (and even perhaps preprints. GitLab is making a few official changes to its free tier that are going to be applied starting in september and that this article takes as granted:
Deleting repositories is not part of that announcement

@Bubu GitHub keeps your project to train Copilot and will dump all that useless data :blobcatlul:

Also, git is decentralized. People who centralize at one hosting service are responsible if things get deleted! Just push to multiple remotes, maybe even host one yourself!

Sign in to participate in the conversation – a Fediverse instance for & by the Chaos community