I think this essentially happened with XEmacs. It's been a while (since the 90's), but I recall that they had a different architecture and more advanced font support. I believe the latter was folded back into Emacs eventually.
io.js (node.js fork) seems like a parallel. It split off to get away from Joyents advisory board that was slowing down development and the trademark issues around Node. It was more performant in the short term with faster development cycles. Once the issues with Joyent were worked out it was merged back into mainline node.
Different political issues here so whether it gets ported back into Rust or becomes what Rust was (or just fades away) is still TBD but who knows?
Well, Joyent spun off Node as a separate org, and io.js's developers pretty much ran the dev side of the new Node.js org. So it really didn't disappear so much as replaced the upstream for the most part.
I think a lot of Rust's governance and leadership has left some poor impressions in a lot of ways and the Trademark licensing is just the step too far for a lot of people. I don't think the goal is to necessarily destroy rust, but definitely change the direction somewhat dramatically.
So more like if C++ PLUS g++ where forked. Then C++ already had many compiler implementations before I was even born, and Rust has one serious one (and several other with explicitly different goals).
This assumes it doesn't slow down velocity/introduce time wasted with bureaucracy/trying to keep 2 separate codebases in sync/timelines/initiatives, etc.
