macroscope.js

macro $testadd {
  case { $$mac $expr } => {
    var ident = makeIdent('test', #{ $$mac });
    return withSyntax ($ident = [ident]) {
      return #{
        return $expr + $ident;
      }
    }
  }
}

macro $wrap {
  case { $$mac $expr } => {
    var ident = makeIdent('test', #{ $$mac });
    return withSyntax ($ident = [ident]) {
      return #{
        (function($ident) {
          $testadd 12
        })($expr);
      }
    }
  }
}

$wrap 42 

Result:

(function (test$242) {
    return 12 + test;
}(42));

Do you need to use makeIdent in $wrap? If you just put test directly in as the parameter everything works out:

macro $wrap {
    case { _ $expr } => {
        return #{
            (function(test) {
                $testadd 12
            })($expr);
        }
    }
}

What about using just rule? The following fails to resolve to the same identifier also (just like above):

macro $testadd {
  rule { $expr } => {
    return $expr + test;
  }
}

macro $wrap {
  rule { $expr } => {
    (function(test) {
      $testadd 12
    })($expr);
  }
}

$wrap 42 

Is it expected that it should resolve? It seems like it should, but that could be debatable.

Yep, it makes perfect sense for rule to behave this way, it's just preserving hygiene. Just think of hygiene as preserving lexical scope. To find the binding of test inside $testadd all you should do is look at the scope at the definition of $testadd (not at the use of $testadd). Since there is no binding of test at the definition then we assume it's either unbound or bound as a global (and so don't perform any renaming).

If we were to bind the test introduced by expanding $testadd 12 to the binding in the IIFE this would be breaking hygiene.