>>34,36,38The Algol and PL/I goto and Lisp's tagbody/go unwind the stack to the correct point.
These were universally considered to be gotos several years before C or setjmp/longjmp existed.
Landin's paper was about interpreting the goto in Algol 60 in lambda calculus, including non-local gotos, using continuations.
"setjmp/longjmp" had to be invented because C doesn't have nested procedures.
We get to that same thing about limitations in C confusing people again, but this time it's goto.