History dependent semi_lex_compare/3

[ridgeworks]

@j4n_bur53

A solution that doesn’t use rbtrees is this here,
replicating the basic idea of the term_factorized/3 algorithm

No objection to opening a second ticket for term_factorized; I also wasn’t sure whether the problem was due to rbtree’s use of the broken compare/3 or whether it was some other issue with cyclic terms.

And even if term_factorize was fixed, is there any reason (other than performance tradeoffs) for preferring it rather than your “custom” factorize/3 for comparing cyclic terms?

[j4n_bur53]

@ridgeworks

preferring it rather than your “custom” factorize/3 for comparing cyclic terms?

factorize/3 is already yesterday’s snow, check out naish/2.
But anything should work that doesn’t give a stack overflow.

[ridgeworks]

@j4n_bur53

factorize/3 is already yesterday’s snow, check out naish/2.

No use confusing things with variations; let’s focus on naish/2 as the prototypical predicate going forward.