rsync was basically done until the maintainer discovered vibecoding

@hailey

Fork that.

rsync-nai

Fork of https://github.com/RsyncProject/rsync at commit 236417cf354220669014317b1ba818b9d931afbb, just before the first obviously LLM-influenced commit

Codeberg.org
openrsync(1) - OpenBSD manual pages

@viq @fogti @zl2tod @hailey OpenRsync will be really nice when it actually gets developed. Last time I checked it hadn't been touched in forever.
@jackemled @viq @fogti @zl2tod @hailey is that actually a concern here? I don't use rsync much but it seems like the sort of thing that could happily exist in a maintenance-only state almost indefinitely. (but I also haven't needed anything exotic out of it, so idk)
@gsuberland @jackemled @viq @zl2tod @hailey It is a concern as afaik openrsync lacks a lot —in terms of features— behind samba's rsync.
@fogti @gsuberland @jackemled @viq @zl2tod @hailey Most alternative rsync implementations are stuck on the simpler protocol version 27, which is before many improvements such as the generation of the file list in parallel: https://github.com/gokrazy/rsync#existing-rsync-implementation-survey
Instead of trying to reimplement the undocumented rsync protocol, I would rather wish for an active community around a modern alternative file transfer tool based on content-defined chunking (as used by restic and borg backup). Google had a proof-of-concept with benchmarks for Stadia, but abandoned it: https://github.com/google/cdc-file-transfer
Following the ideas of FastCDC, resync looks promising on redefining the protocol, while keeping the command line interface familiar to users of rsync: https://github.com/vedLinuxian/resync-rs
GitHub - gokrazy/rsync: rsync in Go! implements client and server, which can send or receive files (upload, download, all directions supported)

rsync in Go! implements client and server, which can send or receive files (upload, download, all directions supported) - gokrazy/rsync

GitHub
@raimue @fogti @gsuberland @[email protected] @viq @zl2tod @hailey any more info on the project or author?, because popping up with a +9k lines initial commit in 2026 is very sus, particularly coming from a recommendation by someone with a slop PFP
@Ember Take the linked projects as examples of what will be possible if we leave the rsync protocol behind. I am not affiliated with any of these projects.