Fix docs about initializer method interceptors (spockframework#1666)

docs/extensions.adoc

@@ -990,13 +990,13 @@ An ellipsis in the figure means that the block before it can be repeated an arbi
 The `... method interceptors` are of course only run if there are actual methods of this type to be executed (the white
 boxes) and those can <<Injecting Method Parameters, inject parameters>> to be given to the method that will be run.
 
-The difference between shared initializer interceptor and shared initializer method interceptor and between initializer
-interceptor and initializer method interceptor - as there can be at most one of those methods each - is, that there are
-only the two methods if there are `@Shared`, respectively non-`@Shared`, fields that get values assigned at declaration
-time. The compiler will put those initializations in a generated method and call it at the proper place in the
-lifecycle. So if there are no such initializations, no method is generated and thus the method interceptor is never
+The shared initializer method interceptor and initializer method interceptor are called around two methods that are
+generated by the compiler if there are `@Shared`, respectively non-`@Shared`, fields that get values assigned at
+declaration time. The compiler will put those initializations in a generated method and call it at the proper place in
+the lifecycle. So if there are no such initializations, no method is generated and thus the method interceptor is never
 called. The non-method interceptors are always called at the proper place in the lifecycle to do work that has to be
-done at that time.
+done at that time. There can also be multiple such methods, if you for example have a super specification that itself
+has fields with initialization expressions.
 
 To create an interceptor to be attached to an interception point, you need to create a class that implements the
 interface `IMethodInterceptor`. This interface has the sole method `intercept(IMethodInvocation invocation)`. The
@@ -1009,9 +1009,10 @@ interception points, there is also the convenience class `AbstractMethodIntercep
 provides various methods for overriding that are called for the various interception points. Most of these methods have
 a double meaning, like `interceptSetupMethod` which is called for the `setup interceptor` and the `setup method
 interceptor`. If you attach your interceptor to both of them and need a differentiation, you can check for
-`invocation.method.reflection`, which will be set in the method interceptor case and `null` otherwise. Alternatively you
-can of course build two different interceptors or add a parameter to your interceptor and create two instances, telling
-each at addition time whether it is attached to the method interceptor or the other one.
+`invocation.method.reflection`, which will be set in the method interceptor case and `null` otherwise, or you can check
+`invocation.method.name` which behaves the same, or you can check for `invocation.target == invocation.instance`.
+Alternatively, you can of course build two different interceptors or add a parameter to your interceptor and create
+two instances, telling each at addition time whether it is attached to the method interceptor or the other one.
 
 [source,groovy,indent=0]
 ----