busybox authoritative git repository (was: AI recommended changing from using git port to https?)

[Michael D. Setzer II]

On 11 Apr 2026 at 21:27, Laurent Bercot via busybox wrote:

To:             	busybox at busybox.net
Subject:        	busybox authoritative git repository (was: AI 
recommended changing
	from using git port to https?)
Date sent:      	Sat, 11 Apr 2026 21:27:38 +0000
From:           	Laurent Bercot via busybox 
<busybox at busybox.net>
Send reply to:  	Laurent Bercot <ska-dietlibc at skarnet.org>

> 
>   So https://busybox.net/ is responsive now, but any URL under
> https://git.busybox.net/ returns a 404.
> 
>   My interpretation is that:
>   - busybox's cgit was getting *hammered* by LLM crawlers to the point
> of unusability
>   - The server hosting the cgit was also hosting the website, so overload
> on cgit impacted the website
>   - Denys, or whoever is in charge of busybox's hosting, chose to take
> the cgit down so the website could function again.
> 
>   (I suspect it's what happened to busybox because it happened to me as
> well. I eventually defeated the LLM crawlers by exhaustively blocking
> all their IPs from accessing cgit, but this is not doable with Apache
> as it is commonly used.)
> 
>   We still need an official main git repository for busybox. We cannot
> make do with copies patched out the wazoo and not maintainer-reviewed,
> or the GitHub mirror that is 2+ years out-of-date.
> 
>   So I would like to know what the plan is moving forward, and only Denys
> (assuming he's in charge of the busybox web site / primary git
> repository) can tell us what his intentions are.
>   I see 3 possible solutions:
> 
>   1. Installing Anubis on git.busybox.net. This is the "standard" 
> solution
> to protect a site from being DoSsed by LLM crawlers.
>   2. Protecting the git.busybox.net cgit the way I've protected the
> skarnet.org one. That requires some sysadmin work on the web server,
> but if there's interest, I'm willing to help.
>   3. Moving the git repository to another place, one that has enough
> resources to withstand the LLM crawler plague. e.g. Codeberg.
> 
>   I'd be okay with any of these 3 solutions. What I'm not okay with is
> the uncertainty and the lack of an authoritative place to get (and
> sometimes contribute to) busybox.
> 

I've seen no issue using this that uses https versus the git protocal.
git remote set-url origin https://git.busybox.net/busybox
git pull

Not sure if that links to one of your 3 options.

Using my regular ~/Downloads/busybox directory
./gitpull2.ch 
Nothing changed (Already up to date).

separate clone in /tmp/busybox
cd /tmp/busybox/
# git pull
fatal: read error: Connection reset by peer

Sometime it works many times fatal??

script.
cat gitpull2.ch 
cd /home/msetzerii/Downloads/busybox
git remote set-url origin https://git.busybox.net/busybox
# Save the current state
PRE_PULL=$(git rev-parse HEAD)

# Pull (silencing the progress noise if you want)
git pull > /tmp/gitpull_out 2> /tmp/gitpull_err
EXIT_CODE=$?

# Check the new state
POST_PULL=$(git rev-parse HEAD)

if [ $EXIT_CODE -ne 0 ]; then
    echo "Pull failed with code $EXIT_CODE. Check /tmp/gitpull_err"
elif [ "$PRE_PULL" == "$POST_PULL" ]; then
    echo "Nothing changed (Already up to date)."
else
    echo "Success: New changes were applied."
fi


> --
>   Laurent
> 
> _______________________________________________
> busybox mailing list
> busybox at busybox.net
> https://lists.busybox.net/mailman/listinfo/busybox


+------------------------------------------------------------+
 Michael D. Setzer II - Computer Science Instructor (Retired)     
 mailto:mikes at guam.net                            
 mailto:msetzerii at gmail.com
 mailto:msetzerii at gmx.com
 Guam - Where America's Day Begins                        
 G4L Disk Imaging Project maintainer 
 http://sourceforge.net/projects/g4l/
+------------------------------------------------------------+