Quickfix for inferred use of callout-to-field not offered

[stephan]

It seems that changes re #193 broke the quickfix to convert an inferred use of a callout-to-field to a method call.

Specifically, r19586 migrated internal AST translation to using synthetic bindings instead. Thus, our QuickFixProcessor can no longer pick up existing AST modifications, but must generate them independently.

Apparently we are also lacking tests here.

[stephan]

Initially fixed in r19690 (test and implementation).

r19691 provides more variants of this quickfix:

• replace field access and materialize the inferred callout-to-field (see #131)
• replace field access to explicit and inferred callout-to-field alike.

[stephan]

Correction re comment:1 initial fix was r19689 not r19690.