I also saved this one: sorcery, written in Rust from another lobsters thread a few weeks ago
I feel like while the concept is not new, its been picking up some steam recently. Seems like a nice minimalist solution, although I worry about people taking down / not wanting to maintain their hosts long-term causing some software to be lost to time, compared to something like Codeberg that will likely stay around for a while
Isn't it orthogonal? It's good to have cheap-to-serve static options for self-hosting. It is also good if large-ish hosters have efficient options to switch to static-hosted public view, if they decide to.
In some cases it could also make it friendlier to Internet Archive crawling, which could easily flip the preservation analysis…
I was just testing out sorcery on my own repos as a potential successor to my little-used Gitea instance. I chose it over stagit and git-arr but hadn't found depp or gituwa, I'll have to evaluate those too.
although I worry about people taking down / not wanting to maintain their hosts long-term causing some software to be lost to time
I think Radicle is trying to solve this problem by creating a network of mirrors, allowing people to contribute their own resources to keeping source code online. I'm not sure it's ready for prime time yet, but I hope they get there or a successor project picks up their torch.
This post and this comment inspired me to add an option to reject requests for those expensive operations (blame, arbitrary diffs/rev trees, grep etc) in my personal cgit fork. I haven't been hit by nasty crawlers particularly hard, but checking my nginx access log consistently shows several requests per second and most of them are to such URLs, so this should help reduce load on my cheap VPS. (disclaimer: i mostly vibecoded it)
telemachus | a day ago
For people who are curious to see more of this kind of tool, there's reporat, which is another implementation in Ruby.
And from reporat's homepage, some others.
jss | a day ago
I also saved this one: sorcery, written in Rust from another lobsters thread a few weeks ago
I feel like while the concept is not new, its been picking up some steam recently. Seems like a nice minimalist solution, although I worry about people taking down / not wanting to maintain their hosts long-term causing some software to be lost to time, compared to something like Codeberg that will likely stay around for a while
lknuth | a day ago
"How can we bare to throw anything away? "
Maybe that's okay. I like that its in the hands or the authors more.
k749gtnc9l3w | a day ago
Isn't it orthogonal? It's good to have cheap-to-serve static options for self-hosting. It is also good if large-ish hosters have efficient options to switch to static-hosted public view, if they decide to.
In some cases it could also make it friendlier to Internet Archive crawling, which could easily flip the preservation analysis…
jaculabilis | 23 hours ago
I was just testing out sorcery on my own repos as a potential successor to my little-used Gitea instance. I chose it over stagit and git-arr but hadn't found depp or gituwa, I'll have to evaluate those too.
I think Radicle is trying to solve this problem by creating a network of mirrors, allowing people to contribute their own resources to keeping source code online. I'm not sure it's ready for prime time yet, but I hope they get there or a successor project picks up their torch.
tonyarkles | 21 hours ago
In case you missed it, Radicle is definitely not ready for prime time atm: https://radicle.dev/2026/09/23/disclosure-of-vulnerability-in-network-protocol
jaculabilis | 19 hours ago
Yes, that was what I had in mind.
antonmedv | a day ago
Another repo generator with github like look: https://github.com/antonmedv/gitmal
qznc | a day ago
I assume this the next escalation level to defend against bots. To be used when I want to publish a repo but not DDOSed?
Doesn't Forgejo provide an options to restrict the public web interface to not show individual commits and stuff just like reporat, grubby, etc?
[OP] bt | a day ago
That’s certainly part of it. Reducing resource usage, especially on low end servers is another plus!
lemon | 6 hours ago
This post and this comment inspired me to add an option to reject requests for those expensive operations (blame, arbitrary diffs/rev trees, grep etc) in my personal cgit fork. I haven't been hit by nasty crawlers particularly hard, but checking my nginx access log consistently shows several requests per second and most of them are to such URLs, so this should help reduce load on my cheap VPS. (disclaimer: i mostly vibecoded it)