ancient chess problem

I once read a story (that i don't know if it's true) about a chess player (i believe it was Fischer, but i could be wrong) who proposed a problem to another group of chess players (i think it were russians), in which he made a bet with them, that they could not solve the problem. After a while trying to solve it, the group gave up and the other chess player showed the solution. The solution was based on old chess rules regarding castling.
Does anyone know the exact entire story and has a source / link to it?
Thanks.

The old hole in the castling rules is not actually "ancient" and I am not aware of any stories surrounding it involving Fischer, but what you are looking for is very likely what is called "Pam-Krabbé-Rochade" in the German Wikipedia: https://de.wikipedia.org/wiki/Pam-Krabb%C3%A9-Rochade

See here for information in English about it:
https://en.wikipedia.org/wiki/Castling#cite_note-5
https://www.reddit.com/r/chess/comments/gy2uwo/before_1972_the_castling_rule_had_a_hole_in_it/

Both the German Wikipedia link and the Reddit link contain (different) interesting chess puzzles involving the hypothetical "very long" castling 0-0-0-0.

This?

https://lichess.org/analysis/standard/6B1/4P3/8/8/8/2pk4/2N5/4K2R_w_K_-_0_1

This famous chess puzzle was made by Danish chess player and writer C. Staugaard. His idea was so creative, that in 1972 FIDE had to change the rules of chess to accommodate it (although they changed the rules in response to a separate puzzle by Tim Krabbé which uses the same concept)

Until 1972, the FIDE rules allowed castling as long as the following 5 conditions were true:

1. The King has not moved.

2. The Rook has not moved.

3. There are no pieces between the King and Rook.

4. The King cannot pass through or land on a square controlled by an enemy piece.

5. The King cannot be in check.

According to these rules, you could technically promote a pawn to a rook and castle across the board "vertically". In response to puzzles like this one, FIDE changed the rules to specify the king and rook must be "on the same rank"

Solution:

1. e7–e8R! Kd3xc2
2. 0–0–0–0#
   https://lichess.org/editor?fen=6B1%2F8%2F8%2F8%2F8%2F2p1K3%2F2k1R3%2F7R+w+-+-+0+1

There's also another type of puzzle like this by Tim Krabbé:
https://www.reddit.com/r/chess/comments/q2fka0/tim_krabb%C3%A9_invented_this_puzzle_in_1972_which_was/

Via Reddit

Thank you both for your replies, but i remember a specific story, involving a bet.
And in that "problem", the castling was in the line (horizontally).
And as mentioned before, it was due to a "hole" in the rules.
Note: if i'm not mistaken, it was a actual problem position (fixed), not a gameplay position.