[sc-525] Float combinators not extracting some terms #245
Conversation
|
This pull request has been linked to Shortcut Story #525: float_combinators not extracting some terms. |
There was a problem hiding this comment.
Can you add a test for recursiveness/combinators in other positions? Like in different application positions, in adt matches and num matches, both in arms and as argument. In lists, tuples, sups, lets, dups, etc.
Also some positions where it should not get extracted, like the body of a lambda (eg foo = @a @b @c (foo a))
Also, update the documentation for this pass to explain the new logic, including the exceptions where we don't float.
| term_info.merge_scope(arg_scope); | ||
|
|
||
| detach | ||
| /// A term is a supercombinator if it is a lambda abstraction without free variables. |
There was a problem hiding this comment.
This is not quite the definition of a supercombinator. This is just a lambda abstraction that is also a combinator, which is all we need here in this case. To be a supercombinator we require that it is of the form
@x1 @x2 ... @xN (Arg1 Arg2 ... ArgM)
Where Arg1-M are also supercombinators, constants or one of the x1-N.
developedby
deleted the
bug/sc-525/float-combinators-not-extracting-some-terms
branch
No description provided.