Gentoo now on Codeberg

[flexibeast]

Quoting the post on the gentoo-dev list:

We would like to announce that Gentoo now has a presence on Codeberg
and contributions can be submitted for the mirror at
https://codeberg.org/gentoo/gentoo[b][/b], as an alternative to GitHub.

If you wish to submit pull requests on Codeberg, it is recommended to
use the AGit approach [1] as it's more space efficient and doesn't
require you to maintain a fork of gentoo.git on your own Codeberg
profile.

Refer to the post for more details.

[Administrator edit: fixed url. Forum auto-linking will capture commas if not inhibited. -Hu]

[CaptainBlood]

Above link fails here. Following works: https://codeberg.org/gentoo/gentoo.

Happy to see gentoo is on codeberg too.

Thks 4 ur attention, interest & support.

[flexibeast]

Above link fails here.

Ah, yeah; it seems the forum software automatically made the URL in the quoted text 'active' (i.e. an actual link), but included the comma at the end, with the result that the URL is incorrect. Thanks for picking that up, and sharing the correct URL!

[Hu]

I edited the post to fix the link. An empty bold block separating the real URL from the punctuation causes the automatic linking to do the right thing.

Unfortunately, fetching anything after the first page returns garbage like this:

May change hands, the colour and the world to another was considerable.

If you are an AI scraper, and wish to not receive garbage when visiting Codeberg: stop visiting. If you are not an AI scraper: contact Codeberg.

So the link may not be all that useful since only the index page seems valid.

[flexibeast]

I edited the post to fix the link. An empty bold block separating the real URL from the punctuation causes the automatic linking to do the right thing.

Ah, okay, thanks, i'll keep that in mind.

Unfortunately, fetching anything after the first page returns garbage like this:

May change hands, the colour and the world to another was considerable.

If you are an AI scraper, and wish to not receive garbage when visiting Codeberg: stop visiting. If you are not an AI scraper: contact Codeberg.

So the link may not be all that useful since only the index page seems valid.

Dang.

[Juippisi]

Unfortunately, fetching anything after the first page returns garbage like this:

May change hands, the colour and the world to another was considerable.

If you are an AI scraper, and wish to not receive garbage when visiting Codeberg: stop visiting. If you are not an AI scraper: contact Codeberg.

So the link may not be all that useful since only the index page seems valid.

This is probably due to one of the things:
1: you don't have JS enabled in your browser,
2: you're using some heavy-traffic VPN node,
3: browsing git commits require an account with the combination of above (works fine for me without an account).

I could imagine no-js users are denied access to git commits but can freely browse the site otherwise.

[Hu]

Yes, JavaScript is blocked as a matter of course. How can I not, when people are going around putting up Anubis CPU burning gates all over the place? Running unexpected JavaScript was ill-advised even before Anubis became popular, and it's even worse now. If Codeberg needs JavaScript enabled to make them not serve garbage, then (a) that's rather user hostile, and (b) failing to clearly document that requirement is even more obnoxious. At least Anubis has the decency to openly demand that I run JavaScript, instead of returning garbage. (Although I still criticize Anubis for not properly communicating just how expensive their scripts are!)

Git commits aren't the only thing they broke. It was just the first one I found. Even parts of their FAQ don't work right. Bizarrely, their registration process is behind an Anubis gate (and apparently a bad old one, since it still says that support for no-script is a work in progress, but upstream released a working version of that quite a while ago). I'll just avoid Codeberg until their behavior improves. It's not worth trying to find which parts of the site work when I keep stumbling over garbage links.

[Chiitoo]

Yes, JavaScript is blocked as a matter of course. How can I not, when people are going around putting up Anubis CPU burning gates all over the place? Running unexpected JavaScript was ill-advised even before Anubis became popular, and it's even worse now. If Codeberg needs JavaScript enabled to make them not serve garbage, then (a) that's rather user hostile, and (b) failing to clearly document that requirement is even more obnoxious. At least Anubis has the decency to openly demand that I run JavaScript, instead of returning garbage. (Although I still criticize Anubis for not properly communicating just how expensive their scripts are!)

Git commits aren't the only thing they broke. It was just the first one I found. Even parts of their FAQ don't work right. Bizarrely, their registration process is behind an Anubis gate (and apparently a bad old one, since it still says that support for no-script is a work in progress, but upstream released a working version of that quite a while ago). I'll just avoid Codeberg until their behavior improves. It's not worth trying to find which parts of the site work when I keep stumbling over garbage links.

I bumped into it too back some time ago, and I'm guessing it's 'iocaine', also discussed here:

https://lwn.net/SubscriberLink/1056953/ ... 1eeddf947/

There's some interest in using that on our sites too, but of course we'll need to figure the no-JS users part out just like with Anubis.

Re: Anubis CPU burn, is it expensive, really? When I see it where I allow it, it takes about a second, if that. I usually don't have time to read how long it took.

There's another topic I have been wanting to touch upon that though, and it will be better discussed there.

Also yeah, not sure if it's just the Anubis version is old, or if the "no-JS solution is being worked on" text is just not removed from the not no-JS version where we see it still...

[flexibeast]

Out of interest, and as someone who has JS disabled in the browser by default, what non-JS-based alternatives are there for dealing with (effective) DoSing by AI bots / LLM crawlers? Does Anubis (or other JS-based analogues) provide functionality, or a level of service, or an ease of setup, not matched by those non-JS-based alternatives?

[szatox]

I think the least expensive Anubis I've seen was 8 seconds,

That's weird, it usually takes under 1 sec, and only the first time I visit the particular website.
I've seen websites that take more time drawing their cookie banners.

I know that some people see it as malware, but I just look at wasting a few CPU cycles as an anonymous payment method... Way better option than asking for a credit card.

[Juippisi]

This is already going offtopic, but...

I think the least expensive Anubis I've seen was 8 seconds, which is 7.9 seconds too long,

Some sites that don't have anubis takes longer for me to load. Must be nice if every site you visit loads in 0.1 sec in 2026.

As regards flexibeast's question, I don't have an answer, but I will note that most of the JS-based paths seem to be premised on the idea that if the JavaScript is "sufficiently complex", then scrapers won't be able to emulate it, and will just fail out the same way that NoScript users fail. I think that's a short-sighted design, since it's readily defeated by the scrapers using a more capable browsing client to more faithfully emulate a full browser. For years, there have been setups for automated testing of websites (by the legitimate test organization), and an automated scraper browser is essentially the same concept. Start a virtual display, connect a browser to it, and have the browser pre-loaded with an automation extension (such as Selenium) and use that to drive it around the web, automatically completing any automated access test that requires a "real browser." It's an arms race that mainly hurts the people who insist on having any modicum of control over what runs locally, and merely inconveniences the scrapers for a short time.

and that's what the above linked iocaine is trying to do.
https://iocaine.madhouse-project.org/ (site works without js, for the curious ones)

I understand your frustration, essentially being blocked from accessing some sites due to your ethics and configuration, but at the same time, codeberg has been very vocal about bots/AI scrapers hammering their site. So I get their view as well, something needs to be done to block most of those kind of bots. This is the technology we currently have.

Personally I wouldn't mind moving to gemini side of web, since the modern www experience is kind of awful.

[Chiitoo]

Yeah, I get hit by the nonsense from theirs site without JavaScript enabled as well.

I think I like the idea of messing with them scrapers like they do, but the experience is indeed very, probably even more unfriendly to the actual human users that get welcomed by it.

(I wanted to look into if they have any non-JS thing planned, but then didn't when I got access denieded like that.)

[szatox]

but I will note that most of the JS-based paths seem to be premised on the idea that if the JavaScript is "sufficiently complex", then scrapers won't be able to emulate it,

No, they are premised on the idea that the slowdown caused by running this JS is more punishing to the crawlers than to humans.
When I browse the web, I'm doing it at my own pace. When a crawler browses the web, it does it at the server's pace.
Forcing the UA to waste 1-2 seconds is barely noticeable to me (or not at all, if I open a link in a new tab so it loads in the background as continue reading the same page), but a crawler spends like a 100 pages worth of resources to receive only 1 page.
In a similar fashion, if an add pops up, it requires MY attention, which makes it really annoying.

Personally I wouldn't mind moving to gemini side of web, since the modern www experience is kind of awful.

It is awful, but it has absolutely nothing to do with the transport protocol. HTTP/1.1 + TLS already was a perfect tool for the job.

[Asch]

This looks like good news.

What should I do in order to sync from Codeberg instead of Github from now on?

Should I just follow the instructions on https://wiki.gentoo.org/wiki/Portage_with_Git and just change every instance of "github" to "codeberg"?

[Ionen]

This looks like good news.

What should I do in order to sync from Codeberg instead of Github from now on?

You shouldn't, codeberg is intended for development only (or at least the moment, but this probably won't change unless stop using github entirely), and it currently does not host the "sync friendly mirror" that includes news and metadata intended for use with emerge --sync.

When it comes to raw clone/fetching from tons of users, there's not much reason not to abuse Microsoft's bandwidth from github rather than put strain on community-funded codeberg or gentoo's anongit (or the rsync mirrors). If signatures are verified, in the end it is all confirmed to be the same and nobody had to go through github's UI for it.

[pietinger]

https://www.phoronix.com/news/Gentoo-St ... deberg-Use