15.8.3.1 Single internal body

Let’s create a single lexical contour with the syntax internal-body and compare a top–level binding with a local binding:

(define A 1)
(internal-body
  (define B 2)
  (begin-for-syntax
    (pretty-print (syntax A))
    (pretty-print (syntax B))
    (pretty-print (eq-ribs? (syntax A) (syntax B))))
  (void))
-| #<syntactic-identifier expr=A mark*=(src)>
-| #<syntactic-identifier expr=B mark*=(src)>
-| #t

we see that the syntactic identifiers created by syntax in the begin-for-syntax have the same list of marks and the same list of rib objects.

Let’s see the rib objects:

(define A 1)
(internal-body
  (define B 2)
  (begin-for-syntax
    (let ((ribs (xp::stx-rib* #'A)))
      (pretty-print (length ribs))
      (pretty-print ribs)))
  (void))
-| 2
-| (#<rib name*=(B)     mark**=((src))     label*=(lab.B)>
    #<rib name*=(A ...) mark**=((src) ...) label*=(lab.A ...)>)

the list of rib objects in the identifier ‘(syntax A)’ has two items: the first is the rib associated to internal-body; the second is the top rib. We see the internal rib holds a single tuple describing the syntactic binding for ‘B’; the syntactic binding for ‘A’ is in the top rib.

With this layout we can understand what the function id->label does in the call:

(define A 1)
(internal-body
  (begin-for-syntax
    (pretty-print (id->label #'A)))
  (void))
-| lab.A

it extracts from its argument the symbol source–name ‘A’, the list of marks ‘(src)’ and the list of rib objects, then it scans the list of rib objects left–to–right looking for a tuple whose source–name is eq? to ‘A’ and whose list of marks is equal? to ‘(src)’; it finds one in the top rib.

This action of looking for a rib’s tuple with the same source–name and the same marks is the syntactic identifier resolution, and we say that the tuple captures the identifier.

This is not the whole story: to handle macro expansion we need more than this; but for lexical contours: building a list of rib objects and using a single mark is enough.